Structure and microwave dielectric characteristics of Ca 1+xNd1-xAl1-xTixO4 ceramics

Article abstract of DOI:10.1111/j.1551-2916.2009.03216.x

CaNdAlO₄ microwave dielectric ceramics were modified by Ca/Ti co-substitution, and their dielectric characteristics were evaluated along with their structure and microstructures. Ca_1+xNd_1-xAl _1-xTi_xO<

Full text of DOI:10.1111/j.1551-2916.2009.03216.x

J. Am. Ceram. Soc., 92 [10] 2286–2290 (2009)
DOI: 10.1111/j.1551-2916.2009.03216.x
r 2009 The American Ceramic Society
ournal
J
Structure and Microwave Dielectric Characteristics of
Ca11xNd1ꢀxAl1ꢀxTixO4 Ceramics
Hong Xia Yuan, Xiang Ming Chen,*,w and Min Min Mao
Laboratory of Dielectric Materials, Department of Materials Science and Engineering, Zhejiang University, Hangzhou
310027, China
CaNdAlO4 microwave dielectric ceramics were modified by
Ca/Ti co-substitution, and their dielectric characteristics were
evaluated along with their structure and microstructures.
Ca11xNd1ꢀxAl1ꢀxTixO4 (x 5 0, 0.025, 0.05, 0.10, 0.15, 0.20)
ceramics with the relative density of over 95% theoretical den-
sity were obtained by sintering at 14001–14501C in air for 3 h,
where the K2NiF4-type solid solution single phase was deter-
mined from the compositions of xo0.20, while a small amount
of CaTiO3 secondary phase was detected for x 5 0.20. With Ca/
Ti co-substitution in CaNdAlO4 ceramics, the dielectric con-
stant (er) increased with increasing x, and the temperature co-
efficient of resonant frequency (sf) was adjusted from negative to
positive, while the Q ꢁ f0 value increased significantly at first
and reached an extreme value at x 5 0.025 and the maximum at
x 5 0.15. The best combination of microwave dielectric charac-
teristics were achieved at x 5 0.15 (er 5 19.5, Q ꢁ f0 5 93 400
GHz, sf 5 ꢀ2 ppm/1C). The improvement of the Q ꢁ f0 value
primarily originated from the reduced interlayer polarization
with Ca/Ti co-substitution, while the decreased tolerance factor,
the subsequent increased interlayer stress, and the appearance of
CaTiO3 secondary phase brought negative effects upon the
Q ꢁ f0 value.
coefficient of resonant frequency tf combined with the improved
Q ꢁ f0 value. With Ca/Ti co-substitution, the temperature coef-
ficient of the resonant frequency tf of (Ca11xSm1ꢀx)(Al1ꢀxTix)O4
ascends with increasing x, and near-zero tf could be achieved.11
CaNdAlO4 ceramics have the same crystal structure as Ca-
SmAlO4; therefore, with Ca/Ti co-substitution, a near-zero tf is
expected to be attained by forming the solid solution of
Ca11xNd1ꢀxAl1ꢀxTixO4. Moreover, the intrinsic dielectric loss
of CaNdAlO4 ceramics is governed by the inner stress, which can
be characterized by the interlayer polarization and interlayer
mismatch.12,13 The reduced interlayer polarization in the solid
solution of Ca11xNd1ꢀxAl1ꢀxTixO4 is expected to improve the
Q ꢁ f0 value. Therefore, the synergistic modification, the near-
zero tf, combined with the improved Q ꢁ f0 value, should be ex-
pected by the co-substitution of Ca/Ti in CaNdAlO4 ceramics.
In the present work, the modification of CaNdAlO4 ceramics
by Ca/Ti co-substitution is investigated. Ca11xNd1ꢀxAl1ꢀx
TixO4 (x 5 0, 0.025, 0.05, 0.10, 0.15, 0.20) ceramics are prepared
by a solid-state reaction approach and their dielectric charac-
teristics are evaluated along with their structure and microstruc-
tures. The roles of interlayer polarization and the tolerance
factor are emphasized to interpret the modification of intrinsic
dielectric loss, while the effects of microstructure upon the ex-
trinsic dielectric loss are also discussed.
I. Introduction
ICROWAVE dielectric ceramics with a suitable dielectric
II. Experimental Procedures
M
constant (er), high-quality factor Q, and near-zero tem-
Dense Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics were prepared by a
routine solid-state reaction process, using high-purity CaCO3
(99.99%), Nd2O3 (99.9%), Al2O3 (99.9%), and TiO2 (99.99%)
powers as the raw materials. The weighed raw materials were
mixed, by ball milling with zirconia media, in ethanol for 24 h,
and the mixtures were calcined at 13001C in air for 3 h after
drying. The calcined powders with 7–8wt% polyvinyl alcohol
(PVA), were pressed into disks of 12 mm in diameter and 2–6
mm in height. The green disks were sintered at 13751–14501C in
air for 3 h. After cooling from the sintering temperature to
11001C at a rate of 21C/min, the ceramics were cooled inside the
furnace naturally. The bulk density of specimens with regular
shapes was measured by the Archimedes method. The crystal
phase constitutions of calcined powders and sintered samples
after crushing and grinding were determined by powder X-ray
diffractometry (XRD), using CuKa radiation (RIGAKU D/
max 2550PC; Rigaku Co., Tokyo, Japan). The lattice parame-
ters a and c along with the unit cell volume Vcell were obtained
by cell refinement. The specimens were thermally etched at
13501C for 30 min. The backscattered electron micrographs of
polished and thermally etched surfaces of sintered ceramics were
observed by scanning electron microscopy (JSM-5610LV,
JEOL, Tokyo, Japan). It should be mentioned here that all
the above structural characterization was conducted for the
samples sintered at 14001C in air for 3 h, which indicated the
highest Q ꢁ f value. The microwave dielectric constant er and the
temperature coefficient of the resonant frequency tf were mea-
sured around 9 GHz by the Hakki–Coleman method,14 using a
network analyzer (E8363B, Agilent Technologies, Palo Alto,
perature coefficient of resonant frequency (tf) have been inten-
sively studied for the last several decades. Because the Q factor
generally varies inversely with the frequency f in the microwave
region, the product of Q ꢁ f0 is usually used to value the dielec-
tric loss instead of Q. Recently, a number of microwave dielec-
tric ceramics have been proposed and investigated in MRAlO4
(M5 Ca and Sr; R 5 Y, La, Nd, and Sm) with a tetragonal
K2NiF4 structure,1–3 for which the single crystals were previ-
ously widely used as substrates for high-Tc superconducting
films.4,5 Considering the developing trend of lower costs in mi-
crowave dielectric ceramics,6,7 MRAlO4 ceramics are expected
to be very competitive ultra-low loss microwave materials with-
out noble elements such as Ta and Nb, compared with the pre-
vious materials such as Ba(Mg1/3Ta2/3)O3,8 Ba(Zn1/3Ta2/3)O3,9
and Ba(Zn1/3Nb2/3)O3.10 Among the CaRAlO4 system, the sta-
ble tetragonal K2NiF4 structure is obtained for R 5 Nd, Sm,
and Y,3 and CaNdAlO4 ceramics have the combined merits of
low cost and good microwave dielectric properties (erB18,
Q ꢁ f0B70 800 GHz, and tfBꢀ13.8 ppm/1C). The key issue
for CaNdAlO4 ceramics is achieving the near-zero temperature
J. C. Nino—contributing editor
Manuscript No. 25669. Received December 22, 2008; approved May 19, 2009.
This work was financially supported by Chinese National Basic Research Program
under grant number 2009CB623302.
*Member, The American Ceramic Society.
wAuthor to whom correspondence should be addressed. e-mail: xmchen59@zju.edu.cn
2286

October 2009
Structure and Microwave Dielectric Characteristics
2287
CA). The temperature coefficient of resonant frequency tf at
microwave frequency was measured in the temperature range
from 201 to 851C. In order to obtain accurate results for high Q
materials, the quality factor Q (the inverse of the dielectric loss,
tan d) of the samples was evaluated by the resonant-cavity
method under microwave frequency.15 The quality factor Q
was evaluated around 7 GHz by the cavity reflection method,
using a silver-coated cavity connected to the analyzer.
III. Results and Discussion
Dense Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics are obtained by sinte-
ring at 13751B14501C in air for 3 h. Figure 1 shows the vari-
ation in relative density of these ceramics with sintering
temperature. The relative density of over 95% theoretical den-
sity (TD) is reached at 14001C, and the highest relative density
(497% TD) is achieved at 14501C in the entire compositional
range.
Figure
2
presents
XRD
patterns
for
Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics sintered at 14001C for 3 h.
For compositions of x 5 0, 0.025, 0.05, 0.10, and 0.15, all re-
flection peaks can be assigned to those for CaNdAlO4 (JCPD
card No.24-0210), or its cell-refined results are marked by the
relevant indices of crystallographic planes, and the single phase
for these compositions is a tetragonal phase with a I4/mmm
space group. At x 5 0.2, the Ca11xNd1ꢀxAl1ꢀxTixO4 solid so-
lution primary phase is identified together with a minor amount
of CaTiO3 secondary phase (JCPD card No.42-0423). There-
fore, the actual terminus for single phase with a K2NiF4 struc-
ture is between x 5 0.15 and 0.2, and it should be close to
x 5 0.20. Some of the reflections shift toward higher angles,
while some of them shift toward lower angles, indicating that the
unit cell dimensions (a and c) have opposite variation trends
with increasing x. As shown in Fig. 3, the variation of a and c
lattice parameters along with the unit cell volume Vcell vary lin-
early with increasing x, and this confirms the formation of a
solid solution of Ca11xNd1ꢀxAl1ꢀxTixO4 in the entire composi-
tional range. Backscattered electron images of the thermally
etched surfaces of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics sintered at
14001C in air for 3 h are used as the complementary means to
determine the phase constitution (see Fig. 4). With increasing x,
there is no obvious variation in the grain morphology. Also,
there is no obvious evidence of the presence of any secondary
phase for the compositions of xo0.20, while the EDS results
indicate some composition inhomogeneity (obvious Ca/Ti-rich
regions) for x 5 0.20, which might be related to the CaTiO3 sec-
ondary phase.
Fig. 2. X-ray diffraction patterns for Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics
sintered at 14001C in air for 3 h.
value increases rapidly with x until x 5 0.025, followed by a
slight decrease at x 5 0.05, and then a gradual increase until x
reaches the maximum (93 400 GHz) at x 5 0.15, which is about
1.4 times of that for x 5 0 (68 600 GHz). For x 5 0.2, the Q ꢁ f0
value begins to decrease again. The variations of the dielectric
constant er and the temperature coefficient of the resonant fre-
quency tf of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics sintered at
14001C for 3 h as a function of x are shown in Fig. 6. Both
the dielectric constant er and the temperature coefficient of the
resonant frequency tf ascend linearly with increasing x. The er
value increases from 18.0 to 20.2, and tf is adjusted from ꢀ13.8
to 1 ppm/1C with increasing x. At the composition of x 5 0.15,
the optimum combination of microwave dielectric characteris-
tics is achieved (er 5 19.5, Q ꢁ f0 5 93 400 GHz and tf 5 ꢀ2
ppm/1C).
The effects of Ca/Ti co-substitution upon the Q ꢁ f value
of CaNdAlO4 ceramics should be discussed from both intrinsic
and extrinsic viewpoints. Considering the structure of Ca11x
Nd1ꢀxAl1ꢀxTixO4 ceramics, tetragonal ABCO4 crystal with a
K2NiF4 structure is composed by one perovskite layer alternat-
ing with one rock salt layers. Owing to its layer structure, it will
be treated as a juxtaposition, perpendicular to z, of successive
layers: –AlO2–(Ca, Nd)O–(Ca, Nd)O–AlO2–(Ca, Nd)O–(Ca,
Nd)O–AlO2–(Ca, Nd)O, and the AlO2 plane has one negative
Figure 5 shows the Q ꢁ f0 values of Ca11xNd1ꢀxAl1ꢀxTixO4
ceramics sintered at 14001C for 3 h as a function of x. The Q ꢁ f0
Fig. 1. Relative density of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics as a func-
tion of sintering temperature.
Fig. 3. Cell parameters of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics as a func-
tion of x.

2288
Journal of the American Ceramic Society—Yuan et al.
Vol. 92, No. 10
Fig. 4. Backscattered electron images of thermally etched surfaces of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics sintered at 14001C for 3 h: (a) x 5 0,
(b) x 5 0.025, (c) x 50.05, (d) x 5 0.10, (e) x 5 0.15, and (f) x 5 0.20.
charge while the neighboring plane (Ca, Nd)O has a 0.5 positive
charge. As a consequence of these polarizations, by the neigh-
boring layers, the constituent ions move along z in the direction
imposed by the resulting electric field until they reach their equi-
librium positions.12 By Ca/Ti co-substitution, interlayer polar-
ization decreases, so the Q ꢁ f0 value ascends with increasing x
because of the decreased interlayer stress due to interlayer po-
larization. In addition, the measure of the stability of the
K2NiF4-type structure is the tolerance factor t, and that some-
what reflects the size adaptability between the perovskite and
rocksalt layers. The tolerance factor t of Ca11xNd1ꢀxAl1ꢀx
TixO4 ceramics can be defined as
ꢀ
ꢁ
ð1 þ xÞRCa2þ þ ð1 ꢀ xÞRNd3þ =2 þ RO2ꢀ
pffiffi
t ¼
ꢀ
ꢁ
(1)
2 ð1 ꢀ xÞRAl3þ þ xRTi4þ þ RO2ꢀ
Fig. 5. Q ꢁ f0 value of Ca11xNd1ꢀxAl1ꢀxTixO4 dense ceramics as a
function of x.
Fig. 6. er and tf of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics as a function of x,
measured at around 9 GHz.

October 2009
Structure and Microwave Dielectric Characteristics
2289
where RCa2þ , RNd3þ , RAl3þ , and RTi4þ , RO2ꢀ are the effective radii
of Ca, Nd, Al, Ti, and O ions, respectively. According to Eq. (1),
the tolerance factor t of Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics de-
creases linearly as x increases The decreased tolerance factors t
would lead to the increased interlayer mismatch, and subse-
quently the increased interlayer stress and the decrease in Q ꢁ f0
value. Therefore, from the viewpoint of intrinsic factors, the
Q ꢁ f0 value should increase at first and then decrease with in-
creasing x in Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics, as the competi-
tion balance result. On the other hand, however, the extrinsic
factors also affect the Q ꢁ f0 value significantly, and they may
even play the dominating role in some cases. The primary ex-
trinsic factors affecting the Q ꢁ f0 value include the secondary
phase, point defects, impurities, inhomogeneities, and grain
boundaries, etc. At first, the appearance of high loss CaTiO3
secondary phase should be the dominating factor for the de-
creased Q ꢁ f0 value. The effects of grain boundary can be con-
sidered as the effects of grain boundary defects and
inhomogeneity, and they can be discussed through the varia-
tion of grain morphology and grain size. If the oxygen vacancy
is absent or its effects are ignorable, the large grain size with a
uniform distribution would contribute to the improvement of
the Q ꢁ f0 value because there are less grain boundaries and
subsequently less grain boundary defects and inhomogeneities.
Actually, MRAlO4 ceramics with a tetragonal K2NiF4 structure
tend not to form oxygen vacancy, and therefore the effects of
oxygen are absent, and the effects of grain morphology and
grain size can simply reflect the effects of grain boundary. The
average grain size has been determined on the polished and
thermally etched surfaces using the linear intercept tech-
nique.16,17 As shown in Fig. 7, the average grain size of
Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics generally increases with in-
creasing x and indicates a maximum at x 5 0.025, which sug-
gests the extrinsic reason for the rapid increase at x 5 0.025. The
inhomogeneity is generally harmful to the Q ꢁ f0 value, and it is
an important issue to improve the homogeneity to optimize the
Q ꢁ f0 value in the present ceramics. Moreover, the Zr-impurity
induced during ball milling should also affect the Q ꢁ f0 value.
The structure dependence of the dielectric constant and tem-
perature coefficient of resonant frequency is discussed as fol-
lows. The dielectric constant er can be evaluated based on the
microscopic Clausius-Mosotti relationship,
dielectric constant er depends on the composition through ion
polarizability and molar volume;18,19 so increasing the dielectric
constant in a ceramic can be achieved by the substitution of a
cation with a greater ionic polarizability. Although, by Ca/Ti co-
substitution, the unit cell volume Vcell slightly increases with in-
3
˚
creasing titanium content, which is from 164.88 A at x 5 0 to
3
166.79 A at x 5 0.20 (as shown in Fig. 3), the ion polarizability
˚
of Ti41 (2.94 A ) is much larger than that of Al (0.78 A ), and
3
31
3
˚
˚
Ca21 (16.80 A ) is the same as Nd (16.80 A ). So the di-
electric constant er increases with increasing x.
The temperature coefficient of resonant frequency tf is given
by the following formula:
3
31
3 17
˚
˚
ꢂ
ꢃ
te
2
tf ¼ ꢀ
þ a
(3)
where te is the temperature coefficient of dielectric constant, and
a is the linear thermal expansion coefficient. At first, the toler-
ance factor decreases with increasing x in Ca11xNd1ꢀxAl1ꢀx
TixO4 ceramics, and the decreased tolerance factor will lead to
the increased magnitude of te,20 and the details need further in-
vestigating. Considering that the te for a paraelectric phase is
generally negative, the increased magnitude of te in the present
ceramics means that the increase of tf is towards the positive
side. On the other hand, the linear thermal expansion coefficient
a will decrease with forming a solid solution, and this should
also result in the variation of tf towards the positive side. Be-
cause the contributions from both te and a lead to the increase
of tf towards the positive side, tf increases with increasing x in
Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics.
Compared with the previous work on Ca11xSm1ꢀxAl1ꢀx
TixO4,11 the intrinsic effects of Ca/Ti co-substitution upon the
microwave dielectric characteristics of CaNdAlO4 ceramics
are more obvious because there is no secondary phase in the
unsubstituted CaNdAlO4 ceramics and the K2NiF4 structure in
Ca11xNd1ꢀxAl1ꢀxTixO4 is more stable than that in Ca11xSm1ꢀx
Al1ꢀxTixO4.
IV. Conclusions
Dense Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics with the K2NiF4-type
solid solution single phase were obtained for xo0.20 by sinte-
ring, while a small amount of CaTiO3 secondary phase was de-
tected for x 5 0.20. With Ca/Ti co-substitution in CaNdAlO4
ceramics, the dielectric constant (er) increased with increasing x,
and the temperature coefficient of resonant frequency (tf) was
adjusted from negative to positive, while the Q ꢁ f0 value in-
creased significantly at first and reached an extreme value at
x 5 0.025 and the maximum at x 5 0.15. The best combination
of microwave dielectric characteristics was achieved at x 5 0.15
(er 5 19.5, Q ꢁ f0 5 93400 GHz, te 5 ꢀ2 ppm/1C). The improve-
ment of the Q ꢁ f0 value primarily originates from the reduced
interlayer polarization with Ca/Ti co-substitution, while the de-
creased tolerance factor and the subsequent increased interlayer
stress bring negative effects upon the Q ꢁ f0 value. Moreover,
the appearance of a secondary phase also indicates negative ef-
fects. The final effects of Ca/Ti co-substitution upon the Q ꢁ f0
value of CaNdAlO4 ceramics are the competing results of these
processes.
1 þ 2baD=Vm
er ¼
(2)
1 ꢀ baD=Vm
3
˚
where Vm is the molar volume A (one-half of the unit cell
volume) and aD is the ion polarizability. The variation of the
References
1X. M. Chen, Y. Xiao, X. Q. Liu, and X. Hu, ‘‘SrLnAlO4 (Ln 5 Nd and Sm)
Microwave Dielectric Ceramics,’’ J. Electroceram., 10 [2] 111–5 (2003).
2X. Q. Liu, X. M. Chen, and Y. Xiao, ‘‘Preparation and Characterization of
SrLaAlO4 Microwave Dielectric Ceramics,’’ Mater. Sci. Eng. B, 103 [3] 276–80
(2003).
3Y. Xiao, X. M. Chen, and X. Q. Liu, ‘‘Microstructures and Microwave Di-
electric Characteristics of CaReAlO4 (Re5 Nd, Sm and Y) Ceramics with Te-
tragonal K2NiF4 Structure,’’ J. Am. Ceram. Soc., 87 [11] 2143–6 (2004).
4R. Sobolewski, P. Gierlowski, W. Kula, S. Zarembihki, S. J. Lewandowski, M.
Berkowski, A. Pajqczkowska, B. P. Gorshunov, D. B. Lyudmirsk, and O. I. Siro-
tinski, ‘‘High-Tc Thin Films on Low Microwave Loss Alkaline-Rare-Earth-Al-
uminate Crystals,’’ IEEE Trans. Magnet., 27 [2] 876–9 (1991).
Fig. 7. Variation
Ca11xNd1ꢀxAl1ꢀxTixO4 ceramics sintered at 14001C in air for 3 h.
of
average
grain
size
with
x
in

2290
Journal of the American Ceramic Society—Yuan et al.
Vol. 92, No. 10
5R. Brown, V. Pendrick, D. Kalokitis, and B. H. T. Chai, ‘‘Low-Loss Substrate
13A. Magrez, M. Cochet, O. Joubert, G. Louarn, M. Ganne, and O. Chauvet,
‘‘High Internal Stresses in Sr1ꢀxLa11xAl1ꢀxMgxO4 Solid Solution (0rxr0.7)
Characterized by Infrared and Raman Spectroscopies Coupled with Crystal Struc-
ture Refinement,’’ Chem. Mater., 13 [11] 3893–8 (2001).
for Microwave Application of High-Temperature Superconductor Films,’’ Appl.
Phys. Lett., 57 [13] 1351–3 (1990).
6I. M. Reaney and D. Iddles, ‘‘Microwave Dielectric Ceramics for Resonators
and Filters in Mobile Phone Networks,’’ J. Am. Ceram. Soc., 89 [7] 2063–72 (2006).
7D. Cruickshank, ‘‘1–2 GHz Dielectrics and Ferrites: Overview and Perspec-
tives,’’ J. Eur. Ceram. Soc., 23 [14] 2721–6 (2003).
14B. W. Hakki and P. D. Coleman, ‘‘A Dielectric Resonant Method of Mea-
suring Inductive Capacitance in the Millimeter Range,’’ IRE Trans. Microwave
Theory Tech., 8, 402–10 (1960).
8N. Ichinose and T. Shimada, ‘‘Effect of Grain Size and Secondary Phase on
Microwave Dielectric Properties of Ba(Mg1/3Ta2/3)O3 and Ba([Mg,Zn]1/3Ta2/3)O3
Systems,’’ J. Eur. Ceram. Soc., 26, 1755–9 (2006).
15X. C. Fan, X. M. Chen, and X. Q. Liu, ‘‘Complex Permittivity Measurement
on High Q Materials via Combined Numerical Approaches,’’ IEEE Trans. Mi-
crowave Theory Tech., 53 [10] 3130–4 (2005).
9P. K. Davies, J. Z. Tong, and T. Negas, ‘‘Effects of Ordering-Induced Domain
Boundaries on Low-Loss Ba(Zn1/3Ta2/3)O3-BaZrO3 Perovskite Microwave Di-
electrics,’’ J. Am. Ceram. Soc., 80, 1727–40 (1997).
16E. E. Underwood, Quantitative Stereology. Addison-Wesley, Reading, MA,
1970.
17L. C. Srearns and M. P. Harmer, ‘‘Particle-inhibited Grain Growth in Al2O3–
SiC: I, Experimental Results,’’ J. Am. Ceram. Soc., 79 [12] 3013–9 (1996).
18E. R. Kipkoech, F. Azough, and R. Freer, ‘‘Microstructural Control of Mi-
crowave Dielectric Properties in CaTiO3-La(Mg1/2Ti1/2)O3 Ceramics,’’ J. Appl.
phys., 97, 064103 (2005).
10M. R. Varma and M. T. Sebastian, ‘‘Effect of Dopants on Microwave Di-
electric Properties of Ba(Zn1/3Nb2/3)O3 Ceramics,’’ J. Eur. Ceram. Soc., 27, 2827–
33 (2007).
11X. C. Fan and X. M. Chen, ‘‘Effects of Ca/Ti Co-substitution upon Micro-
wave Dielectric Characteristics of CaSmAlO4 Ceramics,’’ J. Am. Ceram. Soc., 92
[2] 433–8 (2009).
19D. Shannon, ‘‘Dielectric Polarizabilities of Ions in Oxides and Fluorides,’’ J.
Appl. phys., 73, 348–66 (1993).
12A. Benabbas, ‘‘Review of Polyhedral Distortions as a Multi-Scale Minimiza-
tion of the Electric Polarization and their Correlation with Physical Properties,’’
Acta Cryst., B62, 9–15 (2006).
20I. M. Reaney, P. Wise, R. Ubic, J. Breeze, N. McN. Alford, D. Iddles, D.
Cannell, and T. Price, ‘‘On the Temperature Coefficient of Resonant Frequency in
Microwave Dielectrics,’’ Philos. Mag. A, 81 [2] 501–10 (2001).
&