Applications of CCTO supercapacitor in

energy storage and electronics

Cite as: AIP Advances 3, 062126 (2013); https://doi.org/10.1063/1.4812709
Submitted: 24 May 2013 . Accepted: 18 June 2013 . Published Online: 25 June 2013

R. K. Pandey, W. A. Stapleton, J. Tate, A. K. Bandyopadhyay, I. Sutanto, S. Sprissler, and S. Lin

ARTICLES YOU MAY BE INTERESTED IN

One-step internal barrier layer capacitor

Applied Physics Letters 80, 2153 (2002); https://doi.org/10.1063/1.1463211

Influence of Mn doping on the semiconducting properties of ceramics

Applied Physics Letters 88, 232903 (2006); https://doi.org/10.1063/1.2200732

Microstructure and electrical properties of ceramics

Journal of Applied Physics 99, 084106 (2006); https://doi.org/10.1063/1.2191447

AIP Advances 3, 062126 (2013); https://doi.org/10.1063/1.4812709 3, 062126

© 2013 Author(s).
 AIP ADVANCES 3, 062126 (2013)

Applications of CCTO supercapacitor in energy storage

and electronics
R. K. Pandey,1 W. A. Stapleton,1 J. Tate,1 A. K. Bandyopadhyay,2
I. Sutanto,1 S. Sprissler,1 and S. Lin3
1
Ingram School of Engineering, Texas State University, San Marcos, TX 78666, USA
2
Department of Physics, Texas State University, San Marcos, TX 78666, USA
3
Dan F. Smith Department of Chemical Engineering, Lamar University,
Beaumont, Texas 77710, USA
(Received 24 May 2013; accepted 18 June 2013; published online 25 June 2013)

Since the discovery of colossal dielectric constant in CCTO supercapacitor in 2000,

development of its practical application to energy storage has been of great interest.
In spite of intensive efforts, there has been thus far, no report of proven application.
The object of this research is to understand the reason for this lack of success and to
find ways to overcome this limitation. Reported herein is the synthesis of our research
in ceramic processing of this material and its characterization, particularly with the
objective of identifying potential applications. Experimental results have shown that
CCTO’s permittivity and loss tangent, the two most essential dielectric parameters
of fundamental importance for the efficiency of a capacitor device, are intrinsically
coupled. They increase or decrease in tandem. Therefore, efforts to simultaneously
retain the high permittivity while minimizing the loss tangent of CCTO might not
succeed unless an entirely non-typical approach is taken for processing this material.
Based on the experimental results and their analysis, it has been identified that it is
possible to produce CCTO bulk ceramics with conventional processes having proper-
ties that can be exploited for fabricating an efficient energy storage device (EDS). We
have additionally identified that CCTO can be used for the development of efficient
solid state capacitors of Class II type comparable to the widely used barium titanate
(BT) capacitors. Based on high temperature studies of the resistivity and the Seebeck
coefficient it is found that CCTO is a wide bandgap n-type semiconductor material
which could be used for high temperature electronics. The temperature dependence of
the linear thermal expansion of CCTO shows the presence of possible phase changes
at 220 and 770 ◦ C the origin of which remains unexplained.  C 2013 Author(s). All

article content, except where otherwise noted, is licensed under a Creative Commons
Attribution 3.0 Unported License. [http://dx.doi.org/10.1063/1.4812709]

I. INTRODUCTION
In the recent past, many efforts have resulted in devices with high energy-density storage
capacities which have been classified as “supercapacitors”. A supercapacitor is differentiated from
other capacitors by its orders-of-magnitude advantage in energy density. Energy storage is achieved
by means of static charge rather than by an electrochemical process inherent in a battery. The
supercapacitor concept has been around for number of years but it has experienced resurgence in
recent years because of the advancement in newer designs, different dielectric materials such as
the “electrolytic” capacitors, and most importantly in the discovery of new ceramics exhibiting
colossal relative dielectric constant (CDC), exceeding values of 50,000 at room temperature. One
such material is Calcium-Copper-Titanate (CCTO) which was first reported to have the CDC in
early 2000.1 Since then CCTO has been researched worldwide in the search for miniaturized
capacitor devices for applications ranging from large scale energy storage systems to advanced
microelectronics.

2158-3226/2013/3(6)/062126/13 3, 062126-1 
C Author(s) 2013
062126-2 Pandey et al. AIP Advances 3, 062126 (2013)

It is very important to realize that only superior quality materials with excellent physical
properties can produce reliable working devices. However, it is easily forgotten that engineering
considerations must also be addressed if a device is to be integrated in a system that would make it
commercially viable.
While for a supercapacitor device the colossal dielectric constant and its temperature and fre-
quency dependence are of paramount importance, equally important are its energy storage param-
eters, charge-and-discharge cycles and life span of operation. Also, energy density and breakdown
voltage are important parameters for a capacitor device. The working voltage needs to be below the
breakdown voltage of the dielectric material otherwise the device is not commercially viable. When
making these considerations, the material properties of CCTO prove favorable.
CCTO crystallizes in cubic phase of the perovskite structure with lattice constant,
ao = 0.739 nm and the symmetry group of Im3.1, 2 The dielectric constant of CCT0 ceramic has been
reported to have as high a value as 104 for polycrystalline ceramic and 105 for single crystal.1–4 With
optimized processing parameters it is possible, as we will see later, that bulk ceramic samples may
be produced with dielectric constant approaching the value of single crystal. The origin of the CDC
effect in CCTO still remains unresolved. Two theories advanced to explain this phenomenon are
classified as: (a) internal barrier layer capacitances5–7 and (b) surface barrier layer capacitances.8–12
Environmentally, CCTO is stable and is constituted in the ratio of 1 mole of CaTiO3 to 3 moles
of CuTiO3 as represented by the following chemical equation:

CaT i O3 + 3CuT i O3 = CaCu 3 T i 4 O12 (1)

Both Ca-titanate and Cu-titanate are members of the perovskite group with Ca-titanate being the
classic example of the perovskite structure. Their mixture assumes a complex perovskite phase when
processed as a pseudo-binary solid solution to yield CaCu3 Ti4 O12 (CCTO). Tiny single crystals of
CCTO were grown using the so-called flux method where copper oxide itself acts as the flux.2 It is
to be noted that the permittivity of single crystal may reach a value of 105 which is at least a factor
of 101 to 102 greater for the best reported values for ceramics and films. The nonlinear current-
voltage characteristic typically found for varistor devices is reported also for CCTO ceramic.13, 14
The discovery of the varistor effect in CCTO is technically important because it might lead to its
applications in electronics. Also, its very high dielectric constant and cubic unit cell structure make
it an excellent gate oxide material which can be seamlessly integrated with oxide semiconductors.15
On the one hand, CCTO shows the “colossal dielectric constant (CDC)” effect which is tempera-
ture independent over a wide range of temperature; on the other hand, it is also a very lossy dielectric
material which is a serious drawback so far as its applications are concerned. Its loss tangent (or,
loss factor), tan δ, remains high and as we will see later that the dielectric constant and tan δ are
intrinsically coupled together making CCTO a less attractive material for many applications espe-
cially as a capacitor material. Serious efforts have been made in the past decade to retain the CDC
in CCTO ceramics while drastically minimizing its loss factor (tan δ). Unfortunately this problem
still remains unsolved and as a consequence CCTO has found almost no commercial viability as a
capacitor material or as an energy storage medium.

II. CERAMIC PROCESSING AND STRUCTURAL CHARACTERIZATION

The standard processing steps were followed to produce CCTO ceramic samples as cylindrical
pellets of approximately 13 mm in diameter and 3-4 mm high using a stainless steel die. A mixture
consisting of high purity grade powders of CaTiO3 , CuO and TiO2 in appropriate weight ratio was
ball milled to get particle sizes varying between 100 to 150 nm using a high speed vibratory bill
miller ((MTI Corporation’s SFM-3Desk). Small particle size promotes the grain growth and inter-
connectivity between the grains during the process of high temperature sintering and annealing steps.
The annealed mixture was ground to fine powder which was used for pressing of ceramic pellets
after confirming it to be CCTO by x-ray diffraction analysis. Multiple samples were made by varying
processing parameters such as maximum pressure applied, die temperature, pressure holding time,
etc. to determine the optimum conditions for reproducibly producing high quality ceramic.
062126-3 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 1. Comparative representation of ceramic processing parameters Legend :1. Ratio CCTO: PVA by weight; 2. Die
Temperature in ◦ C; 3. Force in pounds; 4. Time of pressing in hours; 5. Annealing temperature in ◦ C; and 6. Annealing time
in hours.

After the completion of the final pressing step the green pellets were carefully released from
the stainless die and transferred to an annealing furnace in a platinum boat. The annealing was done
at 1050-1150 ◦ C in air for 6-11 hours. This produced shiny black ceramic pellets.
In Figure 1 the ceramic processing parameters are summarized for five of the best pellets
considered suitable for detailed analysis and evaluation for identifying the best candidate for energy
storage and other applications. The selection criteria were the maximum value of the dielectric
constant, the minimum value of the loss tangent and some samples having both the dielectric
constant and loss tangent between these two extremes of the spectrum. The sample labeling was
strictly followed to prevent any error because of the large number of samples processed (25-30);
here B denotes the batch number of the ceramic powder and S the specific pellet made from this
charge. In ceramic science it is desirable to process highly dense and homogenous ceramic samples
for scientific studies and practical applications. The density of each pellet was determined by the
standard Archimedes method after the completion of the annealing process. Highest density of
4.75 g.cm−3 was achieved which is about 94 % of the theoretical density of 5.0 g.cm−3 . For this
experiment one large pellet was pressed for each pressure corresponding to each data point shown
in Figure 2. The die temperature was maintained at 400 ◦ C during pressing. Annealing time and
temperature were kept the same for each pellet. From a total of 9 pellets 2-3 wafers were cut from
each pellet and polished for density measurements. Each data point represents the mean of the
values for density obtained for all samples of the same pellet. The values varied by ±5% between
the samples of the same set. In all about 20-22 samples were used for finding the density as a function
of pressure which is shown in Figure 2. We notice here that between 50 and 230 MPa of pressure the
density rises rather rapidly and then reaches its maximum value gradually as the pressure increases.
This threshold appears to be reproducible.
In Figure 3 the XRD diffractogram is shown for B2S2 and B2S4 samples. Comparing the peaks
of this figure with the published XRD patterns we find that all the major peaks are present in our
samples.13, 16 Besides these peaks the plot also includes two minor peaks belonging to CaTiO3 . Also
the peaks for each of the two samples overlap each other. The XRD pattern for other samples was
identical to the one shown in this figure.
062126-4 Pandey et al. AIP Advances 3, 062126 (2013)

TABLE I. Elemental analysis of CCTO ceramic.

Elements Nominal % EDAX % Deviation

Ca 5 4.21 −0.79
Cu 15 12.12 −2.88
Ti 20 15.15 −4.85
O2 60 68.51 8.51

FIG. 2. Density vs. pressure for CCTO ceramic.

FIG. 3. XRD diffractogram of CCTO ceramic.

Table I gives the compositional analysis done by EDAX. We find that the sample is slightly
deficient in Ca, Cu, and Ti and excess in oxygen. A similar trend was found also for the other
samples.
In Figure 4 we show a typical SEM micrograph of the B2S2 sample. The grains are well
developed and the grain boundaries between them also clearly defined.
062126-5 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 4. SEM micrograph of B2S2 sample.

III. ELECTRICAL AND DIEELECTRIC PROPERTIES

In order to understand the semiconductor nature of CCTO electrical resistivity and Seebeck
coefficient were determined from 150 to 650 ◦ C. The high temperature regime was chosen to get
some information in the intrinsic nature of CCTO semiconductor as well as to determine if CCTO
would be a good material for high temperature electronics. The measurements below 150 ◦ C could not
be taken because the resistivity increased to a level that exceeded the instrumental range for reliable
measurement. Both these parameters were determined using the latest model of the Thermoelectric
Test Set (ULVAC-RICO ZEM-3 M8, Yokohama, Japan).
The electrical resistivity, ρ, was measured by the four-point probe method. The temperature
dependence of ρ showed the typical behavior found for semiconductors, that is, as the temperature
decreases the resistivity increases rapidly such that all semiconductors tend to be insulators close to
0 K.
In intrinsic region (at high temperatures) the temperature dependence of a semiconductor is
given by the well- established exponential law:
 
Eg
σ = T ex p −
3/2
(2)
2k T
where, Eg denotes the band gap, k is the Boltzmann constant and T is the temperature in K. In
equation (2) the exponential part dominates over the T3/2 part.17
The slope of (ln σ ) vs. T−1 plot, shown in Figure 5, gives the value of the (−Eg /2 k) term
enabling us to estimate the band gap of CCTO to be 3.40 eV. This is comparable to the values of Eg
for many oxides.
The Seebeck coefficient, κ, is defined as the thermoelectric potential (dV) developed per unit
temperature difference, or
dV dV
k= = (3)
dT T2 − T1
where, temperature T2 represents the hot junction and T1 the cold junction of the sample.
The Seebeck coefficient can have both negative and positive signs depending upon the nature
of the semiconductor material. For CCTO it is negative for the entire temperature regime indicating
that CCTO is an n-type semiconductor which is in agreement with the reference.13 The bandgap of
3.4 eV and n-type nature makes CCTO also a wide bandgap material. The temperature dependence
of the Seebeck coefficient of CCTO is given in Figure 6. From the slope of this plot we get for CCTO
ceramic the value of the thermoelectric dV ≈ −0.12 mV. In comparison it is +0.643 mV for a 90%
Pt-10% Rh thermocouple which is a widely used thermocouple at high temperatures.17 Also we find
062126-6 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 5. Temperature dependence of conductivity (σ ) for CCTO ceramic.

FIG. 6. Temperature dependence of Seebeck coefficient.

that the data points deviate from the ideal linear curve at higher temperatures. This trend has been
reported also for many other ceramic materials.
The third parameter measured at high temperature was the coefficient of thermal expansion
(CTE). It is an important parameter for understanding the high temperature induced changes for
ceramic processing as well as for determining the temperature range of operation of a ceramic based
device such as a capacitor or a varistor. The equipment used for measuring CTE was a thermo -
mechanical analyzer, TMA (of TA Instruments, model TMA Q400 with vertical furnace). The linear
CTE is ratio of axial thermal strain and change in temperature. Our TMA precisely measures change
in length of a specimen with respect to change in temperature. The coefficient of thermal expansion
is defined as:
   
εther mal 1 L
α= = · (4)
T T L
where, α is the linear coefficient of thermal expansion, εthermal the axial thermal strain, T the
temperature gradient and L the change in original length, L.
For the determination of CTE of CCTO we used the temperature ramp rate of 2 o C/min from
room temperature to 1000 o C. Argon was used as the purge gas with flow rate of 5 mL/min to create
an inert atmosphere in the chamber. The change in dimensions (μm) of the CCTO sample was
measured precisely with the data sampling rate of 60 points/minute. The total numbers of data points
collected were over 29,000. The reason for large number of data points was to capture all important
062126-7 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 7. Temperature dependence of linear dimensional change.

events of phase changes. Figure 7 displays the change in length with respect to temperature. The
value for CTE was determined from the slope between 300 and 500 o C which is found to be 8.92
× 10−6 deg−1 for CCTO. It is comparable to the literature value for Al2 O3 of 8.1 × 10−6 deg−1 .
We see in this figure the evidence of two phase transitions; one at 220 ◦ C and the other at
◦
770 C. The presence of some sort of phase changes at these temperatures is supported by the
visual observation of dimensional changes around these temperatures. There has been no other
report in literature that would support this experimental fact; perhaps because there is almost no
high temperature properties of CCTO reported in open literature. There is a need to investigate this
phenomenon further to ascertain the existence of a phase change and its nature. But this is beyond
the scope of this paper. On the other hand, there have been numerous reports of phase transitions
in CCTO at low temperatures. It was reported first when the material was discovered in 2000.
It was observed both in the temperature dependence of the permittivity and of loss tangent that
abrupt changes take place below 100 K.1 This has been supported by many subsequent publications
including in the low temperature behavior of magnetic susceptibility, ac conductivity, and specific
heat.18, 19
For the determination of the frequency dependence of the permittivity and loss factor samples
from each pellet were made by cutting into square pieces of approximately 3 mm × 3 mm × 1 mm.
The opposite large faces were polished to high shine and visually inspected for any evidence of hair
line cracks or other obvious surface deformity. Only those samples that appeared to be without micro
cracks or other surface defects were selected. The flat opposite faces were gold plated by sputtering.
Permittivity as well as loss tangent were determined at room temperature using a precision LCR meter
(Wayne Kerr 6500B) with frequency range of 20 Hz to 10 MHz. The samples were clamped between
the spring loaded pins of an Agilent 16034 ECV fixture for data collection. This fixture enables one
to determine precisely the values of the permittivity and tanδ without the superimposition of parasitic
effects caused by connecting wires. Figures 8(a) and 8(b) represent the frequency dependence of the
permittivity of the best sample from each pellet. The labeling of individual samples was done by
identifying first the pellet number followed by the sample number; for example, B2S2-1 means the
sample #1 of the pellet B2S2. The sample B2S2-1 exhibits the highest permittivity of approximately
52,500 and practically dwarfs the values for all other samples (Figure 8(b)). Also it goes through a
sharp resonance at about 550 kHz. The next highest permittivity was found for the B2S2-an sample.
This sample was identical to B2S2-1 in every respect except that gold had diffused into it during the
annealing at 1100 ◦ C after the electrodes were sputtered. All other samples were only heat treated at
about 200–300 ◦ C after sputtering of gold electrodes.
In both Figure 8(a) and 8(b) we find that not all samples undergo resonance; most probably
because the condition of resonance is not met in these samples. Both B2S2-an and B3S1-2 show
strong frequency dependence whereas the others practically remain flat for the entire range of
frequency covered. The quasi independence with log (frequency) of the permittivity has also been
reported by other investigators.
062126-8 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 8. (a) Frequency dependence of permittivity for B1S2-1, B2S2-an, B2S3-1,B3S1-2 and B3S2-4. (b) Frequency depen-
dence of permittivity for B1S2-1, B2S2-1 and B2S2-an.

The frequency dependence of the tanδ is given in Figure 9(a) for all the samples except the
sample B2S2-1. Here we see that sample B1S2-1 shows strong resonance around 450 kHz. The sharp
resonances of the B2S2-1 sample, (Figure 9(b)) occur at 550 kHz. For all other sample the curves
remain almost unchanged as log (frequency) increases which is consistent with the conclusion made
previously.
In Table II the frequency dependence of the permittivity and loss tangent for all samples have
been tabulated at the set frequencies of 1 kHz, 10 kHz, 100 kHz and 1 MHz. This gives an insight
into the interdependent nature of the permittivity and loss tangent of CCTO.
The relationship between the loss tangent and the permittivity for CCTO is represented graph-
ically in Figure 10. We can conclude that a strong coupling exists between these two parameters
and that they increase or decrease in tandem with each other. The interdependence of these two
parameters follow the empirical relationship of tanδ = a ln(εr ) – c. Here a ≈ 0.057 and c = 0.36.
This coupling appears to be the intrinsic nature of CCTO and is indicative of the fact that it is almost
impossible to produce CCTO ceramic with giant dielectric constant but very low loss tangent. Unless
062126-9 Pandey et al. AIP Advances 3, 062126 (2013)

TABLE II. Permittivity and loss tangent of selected CCTO samples.

Frequency B1S2-1 B2S2-1 B2S2-an* B2S3-1 B3S1-2 B3S2-4

(Hz) tan δ εr tan δ εr tan δ εr tan δ εr tan δ εr tan δ εr

1 × 103 0.092 3407 0.414 524647 0.192 19144 0.030 1633 0.131 3108 0.133 3282
10 × 103 0.025 3238 0.441 325142 0.310 13551 0.012 1602 0.080 2630 0.076 2780
100 × 103 0.014 3193 0.872 187012 0.426 8641 0.019 1580 0.068 2437 0.042 2580
1 × 106 0.076 3433 1.18 −46099 0.936 5614 0.075 1655 0.303 2386 0.047 2644

FIG. 9. (a) Frequency dependence of loss tangent for of B3S2-4, B3S1-2, B3S1-2, B2S2-an and B1S2-1. (b) Frequency
dependence of loss tangent for B1S2-1, B2S2-1 and B2S2-an.

a dopant can be discovered which can reverse this trend, the loss tangent and permittivity will rise
or fall in tandem for CCTO bulk ceramic.

IV. APPLICATIONS
Irrespective of the fact that the ideal combination of loss factor and giant dielectric constant has
not been found, it is still possible to identify practical applications of CCTO bulk ceramic. Some of
the possible applications are discussed briefly in this section.
Electronics: CCTO being a non-polar cubic perovskite with dielectric with giant dielectric
constant appears to be a good material for integration with microelectronics technology as a high-k
gate oxide.15 Additionally, its nonlinear current-voltage characteristics similar to those found for
good varistors might pave the way for CCTO in general stream of integrated circuit technology
062126-10 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 10. Interdependence of loss tangent and permittivity at 1 kHz for all samples.

as protective devices.13, 14 As a wide bandgap n-type semiconductor with excellent temperature

dependence of resistivity it is poised to be a potentially good material for high temperature electronics.
Energy Storage: Energy storage devices (ESD) are characterized by two interdependent param-
eters: the energy density and the power density when a load is present. There are variety of ESD
devices; some of which are batteries, SEMS (superconducting magnetic energy storage devices),
flywheels, supercapacitors, electrolytic capacitors, and film capacitors. Their energy density and
power density cover a broad range of spectrum which is represented by the so-called Ragone plot.20
According to this plot, in general, a device with high energy density shows low power density.
Batteries and capacitors are two prominent ESDs and widely used in all sorts of applications.
Batteries have the high energy density and low power density. In comparison, capacitors have
high power density but low energy density. There are many factors in the favor of supercapacitors as
an ESD rather than batteries; they are: higher power density and faster charge-discharge time; wide
temperature range of operation: −40 to +300 ◦ C and long life cycle >106 cycles.
The guiding forces behind batteries are embedded in the transport of ions and electrons, acti-
vation barriers, and the impedance of the electrode interface. For a capacitor the guiding force is
embedded in its parallel plate configuration, which is the most used capacitor configuration in real
world applications (as well as for film and electrochemical capacitors). The stored energy (E) for a
capacitor given by one-half of the product between the capacitor C and the square of the potential V
being applied:
 
1
E= CV 2 (5)
2
The energy density ( ) is defined as the energy stored per unit volume. That is for a parallel plate
capacitor device it is:
   
E 1 V εr ε0 V 2
= = Q = (6)
Ad 2 Ad 2 d
where, A is the area of the metal electrode, d the thickness of the dielectric, Q the charge accumulated
on the electrodes, V the voltage applied and ε0 the permittivity of free space. We find from equation (6)
that the energy density is directly proportional to the stored charge in a capacitor. By increasing the
charge one automatically increases the energy density of the capacitor. ED also increases as square
of the applied potential. Since CCTO is expected to have a very high break-down voltage like barium
titanate (BaTiO3 ), a large potential can be applied to its device resulting in very large ED.
We find that B2S2-1 meets the benchmark requirement of a practical ESD if it could be
produced as 0.5 μm tick film with the permittivity equal to that found for B2S2-1 ceramic.
For many applications the requirement for the ED is 5 J.cm−3 . Using the dielectric param-
062126-11 Pandey et al. AIP Advances 3, 062126 (2013)

FIG. 11. Frequency dependence of energy density ( ) of B2S2-1 with different dielectric thickness.

FIG. 12. (a) Comparison between the permittivity of CCTO and BT (BaTiO3). (b) Comparison between loss tangent of
CCTO and BT (BaTiO3).
062126-12 Pandey et al. AIP Advances 3, 062126 (2013)

eters that we measured for the B2S2-1 sample, we have calculated the ED using the thick-
ness d = 1.1 mm (actual sample thickness) and V = 100 V, d (film of B2S2-1) = 1 um for
V = 10 V and d = 0.5 um for V = 10 V. The ED is plotted as frequency dependent parameter in
Figure 11. Clearly the film with 0.5 um meets the benchmark below 100 Hz.
Class II capacitors: Some samples of CCTO also meet the requirements of Class II capacitors
and are in fact quite competitive with barium titanate (BT) capacitor, which is the current state-of-art
standard of solid state capacitors. The materials requirements for a Class II capacitor are: 1 × 103
< εr < 20 × 103 and 0.03 < tanδ < 0.10. Also, they should have very high breakdown voltage. In
Figure 12(a) we present the permittivity of three CCTO samples in comparison with BT; and in
Figure 12(b) we compare the loss factor, tanδ, of these samples with the values found for BT
ceramic. We see in these figures that all the three samples of CCTO have permittivity higher than
BT and their tanδ at 1 MHz compares favorably with BT. Even at lower frequencies their values for
all the three samples lie in the acceptable range for a Class II capacitor.

V. CONCLUSIONS
CCTO’s perovskite cubic structure, high temperature stability, high permittivity, wide bandgap
and n-type semiconductor nature are favorable for applications in many areas of technology as
discussed above. A large number of ceramic samples were made for this research and character-
ized for their permittivity and loss tangent in a wide range of frequency from 20 Hz to 10 MHz.
The maximum value of permittivity at 1 kHz was found to be greater than 50x103 and the low-
est tanδ to be about 0.09. However, these two best values did not exist simultaneously for the
same sample. In fact, we conclude that permittivity and tanδ are strongly coupled; and their rela-
tionship follows an empirical law given by tanδ ≈ a ln(εr ). The high permittivity goes with high
tanδ. That is a serious drawback and has greatly handicapped CCTO for its adoption in practical
applications. It is a challenge for materials scientist to minimize the tanδ while still retaining its
“giant” permittivity. In spite of this serious drawback we have identified its potential as a relevant
material for energy storage and class II capacitor as its appeal for high temperature electronics and
microelectronics.

ACKNOWLEDGMENTS
We acknowledge the support of the U.S. Air Force Office of Scientific Research (AFOSR), STTR
- Contract # FA8650-10-M-2117, and of U.S. Ferroics Company for the subcontract to Texas State
University. We also acknowledge the support of NSF Grant # CBET- 1126745 (Lamar University), the
support of the Office of the Associate Vice President for Research and Ingram School of Engineering
at Texas State University-San Marcos; as well as of Lamar University Research Enhancement Grant
under Grant # 420254 for this research.
1 A. P. Ramirez, M. A. Subramanian, M. Gardel, G. Blumberg, D. Li, T. Vogt, and S. M. Shapiro, Solid St. Commun. 115
(5), 217 (2000).
2 M. A. Subramanian, D. Li, N. Duan, B. A. Reisnet, and A. W. Sleight, J. Solid St. Chem. 151 (2), 323 (2000).
3 C. C. Homes, T. Vogt, S. M. Shapiro, S. Wakimoto, and A. P. Ramirez, Science 27(293), 673 (2001).
4 M. A. Subramanian and A. W. Sleight, Solid St. Sciences 4, 347 (2002).
5 D. C. Sinclair, T. B. Adams, F. D. Morrison, and A. R. West, Appl. Phys. Lett. 80, 2153 (2002).
6 S-Y Chung, I.-D. Kim, and S.-J. L. Kang, Nature Mat. 3, 774 (2004).
7 V. Brice, G. Gruener, J. Wolfman, K. Fatayeyeva, M. Tablellout, M. Gervais, and F. Gervais, Mat. Sci. Eng. B 129, 135

(2006).
8 P. Lukenheimer, R. Ficht, S. G. Ebbinhaus, and A. Loidl, Phys. Rev. B 70, 172102 (2004).
9 P. Lukenheimer, V. Bobnar, A. V. Pronin, A. I. Ritus, A. A. Vlokov, and A. Loidl, Phys. Rev. B 66, 052105 (2002).
10 S. Krohns, P. Lukenheimer, S. G. Ebbinghaus, and A. Loidl, Appl. Phys. Lett. 91, 022910 (2007).
11 S. Krohns, P. Lukenheimer, S. G. Ebbinghaus, and A. Loidl, Appl. Phys. Lett. 91, 09902 (2007).
12 G. Deng, T. Yamada, and P. Murralt, Appl. Phys. Lett. 91, 202903 (2007).
13 Lajun Liu, Liang Fang, Yanmin Huang, Yinhua Li, and Shi Damping, J. Appl. Phys. 110, 094101 (2011).
14 S. Y. Chung, IL-D00 Kim, and S.-J. L. Kang, Nature Mats. 3,774 (2004).
15 R. K. Pandey, William A. Stapleton, Ivan Sutanto, Amanda A. Scantlin, and Sidney Lin, J. Electronic Mats. (in review).
062126-13 Pandey et al. AIP Advances 3, 062126 (2013)

16 A. E. Smith, T. G. Calvarese, A. W. Sleight, and M. A. Subramanian, J. Solid St. Chem. 182, 409 (2009).
17 S. O. Kasap, Principles of Electronic Materials and Devices, Second Edition, (2002).
18 Andrey A. Levchenko, Loic Marchen, Yosuke Moriya, Hitoshi Kawaji, Torru Atake, Sophie Guillemet-Fritsch, Bernard

Durand, and Alexandra Novrotsky, J. Mats. Res., 23(6), 1522, (2008).

19 S. Krohns, J. Lu, P. Luckenheimer, V. Brize, C. Autret-Lambert, M. Grevais, F. Gervais, F. Bouree, F. Porcher, and A.

Loidl, Eur. Phys. J. B 72, 173 (2009).

20 Thomas Christen and Martin W. Carlen, J. Power Sources, 91, 210–216, (2000).