Journal of Crystallization Process and Technology, 2012, 2, 16-20
http://dx.doi.org/10.4236/jcpt.2012.21003 Published Online January 2012 (http://www.SciRP.org/journal/jcpt)
Synthesis and Characterization of CaPd3O4 Crystals
Hiroaki Samata1*, Satoshi Tanaka1, Soichiro Mizusaki2, Yujiro Nagata2, Tadashi C. Ozawa3,
Akira Sato4, Kosuke Kosuda4
1Graduate School of Maritime Sciences, Kobe University, Kobe, Japan; 2College of Science and Engineering, Aoyama Gakuin Uni-
versity, Sagamihara, Japan; 3International Center for Materials Nanoarchitectonics, National Institute for Materials Science, Tsukuba,
Japan; 4Materials Analysis Station, National Institute for Materials Science, Tsukuba, Japan.
Received November 15th, 2011; revised December 21st, 2011; accepted December 29th, 2011
ABSTRACT
A new method for the crystal growth of alkaline-earth palladate CaPd3O4 was developed. The crystals were synthesized
on a voltage-applied electrode in a molten chloride solvent. The maximum length of the crystal was about 1.5 mm. The
X-ray diffraction data were refined well by assuming a cubic structure of the space group Pm 3n, and the lattice con-
stant a was 5.7471 (10) Å. The temperature dependence of the resistivity showed semiconductor-like characteristics
with a very small activation energy Ea of 0.45 meV at low temperatures, and the resistivity at 300 K was 0.1 ·cm. The
temperature dependence of the molar magnetic susceptibility showed the Curie-Weiss paramagnetic behavior with a
small molar Curie constant Cmol of 5.0(1) × 10–3 emu K/(mol·Oe), indicating the existence of localized spin moments.
Keywords: Single Crystal; Electrochemical Technique; CaPd3O4
1. Introduction
Alkaline-earth palladate CaPd3O4 has a cubic NaPt3O4-type
structure and exhibits a semiconductor-like temperature-
dependent resistivity [1,2]. Hase and Nishihara calculated
the band structure of CaPd3O4 by the FLAPW method
within LDA and suggested that CaPd3O4 is a potential
excitonic insulator in which electrons and holes bind as
excitons [3]. Ichikawa and Terasaki reported that CaPd3O4
was a degenerate semiconductor with a low carrier con-
centration [4]. Chemical substitution on insulating com-
pounds occasionally produces interesting physical prop-
erties, and insulator-metal transitions have been observed
in the Ca1–xNaxPd3O4 and Ca1–xLixPd3O4 systems [2,4-6].
The aliovalent ion-substituted CaPd3O4 is considered to
be a good representative of a thermoelectric material [4].
Thermoelectric energy conversion in solids has great po-
tential as an energy-saving technology, but this type of
conversion is not widely used due to the poor perform-
ance of the materials.
We have previously investigated the effects of aliova-
lent ion substitution on the properties of SrPd3O4, which
is isostructural with CaPd3O4, and we observed the exis-
tence of insulator-metal transitions and bipolar conduc-
tivity [7,8]. In this system, the substitutions of Na+ and
Bi3+ for Sr2+ introduced hole and electron carriers, re-
spectively. The power factor of Sr1–xNaxPd3O4 (x ≤ 0.2)
was not sufficient for practical use as a thermoelectric ma-
terial, but increasing the amount of substitution and de-
vising an appropriate synthesis method may enhance its
thermoelectric performance [8].
A strategy for searching for novel thermoelectric ma-
terials was suggested by the analysis of single-crystal data
for Ce-based compounds [9]. Moreover, the use of a high-
quality single crystal was effective for practical applica-
tions of thermoelectric materials [10]; therefore, the de-
velopment of new crystal growth techniques is important
for practical use as well as for the theoretical analysis of
thermoelectric materials.
The growth of CaPd3O4 crystals was previously attempted
using the conventional flux method; potassium hydroxide
was used as the flux, and the crystallographic properties
were characterized [11]. However, no other study has in-
vestigated the crystal growth, and the intrinsic properties
of this material have not been elucidated in detail. Elec-
trosynthesis in a molten flux effectively prepares single
crystals of transition metal oxides [12], and we recently re-
ported the crystal growth of certain oxides on electrodes
in molten chloride solvents [13]. In the current study, we
describe the results of the synthesis and the characteriza-
tion of CaPd3O4 crystals.
2. Experimental
Single crystals of CaPd3O4 were synthesized on the sur-
face of voltage-applied platinum rods submerged in mol-
*Corresponding author.

Synthesis and Characterization of CaPd3O4 Crystals 17
ten chloride solvent. A schematic of the apparatus used
for the crystal growth is shown in Figure 1 of [13]. An
alumina crucible with a capacity of 20 or 30 cc was used
to hold the solvent, which was a mixture of CaCl2 and
NaCl. An appropriate amount of PdO powder (99.9%) was
placed in the crucible, and the chloride mixture was lay-
ered over the PdO powder. The total amount of chlorides
was 20 or 30 g, and the molar ratio of NaCl/CaCl2 varied
in the range of 0 to 5.3; a PdO to CaCl2 molar ratio of
0.04 to 0.12 was used. The crucible was placed in an alu-
mina container, and two parallel platinum rods (0.5 mm
in diameter) were inserted into the solvent at a distance
of 7 mm from each other. The container was covered by
an alumina plate and placed in an electric furnace with an
ambient atmosphere. The depositions were performed at a
constant temperature in the range of 780˚C to 1050˚C,
which accounted for the melting points of CaCl2 (772˚C)
and NaCl (801˚C). After the solvent was melted at each
synthesis temperature, an electrical voltage in the range of
0.2 to 0.4 V was applied to the electrodes using a DC power
supply. The synthesis was carried out over 48 to 96 hours;
the applied voltage was then turned off, and the electrodes
were quickly extracted from the molten solvent. The crys-
tals were washed thoroughly in distilled water to remove
the solvent from the surface of the crystals, which were
then dried on a hot plate. Selected crystals were annealed
at 500˚C for 7 days under flowing oxygen gas at 1 atm.
The chemical composition of the crystals was charac-
terized by electron-probe microanalysis (EPMA; JEOL,
JXA-8500F). The crystal structure was identified by sin-
gle-crystal X-ray diffraction and powder X-ray diffract-
tion. Single-crystal X-ray diffraction data were acquired
on a Bruker SMART APEX S diffractometer equipped
with a pyrolytic graphite incident monochromator and a
CCD camera at room temperature. The crystal was mo-
unted on a glass fiber, and X-rays were generated using a
(a) (b)
Figure 1. Photographs of as-grown CaPd3O4 crystals syn-
thesized by applying a voltage of 0.2 V at 950˚C for 48 h: (a)
cry stals on the platin um anode rod and ( b) a crystal remove d
from the electrode.
Mo target at 50 kV and 30 mA. A total of 3597 reflections
were collected for the full sphere using a 0.3˚ ω-scan with
a 40 s exposure. The crystal structure was solved by di-
rect methods using SHELXS-97, and the structure was
refined utilizing the SHELXL-97 software package, which
used 130 unique reflections and 7 parameters [14]. Pow-
der X-ray diffraction data were acquired using Cu K

generated at 40 kV and 20 mA in a 2
 range of 20˚ - 80˚
at room temperature (Rigaku, RINT2000). The data were
refined by the Rietveld method using the values obtained
from the refinement of single-crystal X-ray diffraction
data [15].
The electrical resistivity of the crystals was measured
using a conventional DC two-probe method in the tem-
perature range of 10 to 300 K; a refrigerator was used to
maintain temperatures lower than room temperature. The
electrical contacts were established by attaching gold leads
onto the surface of the crystal using silver paste to form
an ohmic contact. Magnetic measurements were carried
out using a superconducting quantum interference device
magnetometer (Quantum Design, MPMS2) in the tempe-
rature range of 5 to 300 K in an applied magnetic field of
10 kOe.
3. Results and Discussion
Figure 1(a) shows a photograph of the CaPd3O4 crystals
grown by applying a voltage of 0.2 V at 950˚C for 48
hours; a molar ratio of PdO:CaCl2:NaCl of 0.06:1:2 was
used. The electrical current was approximately 1 mA at the
end of crystal growth. These conditions were the most sui-
table for the growth of large crystals within the range of
the present experimental conditions. Green-tinted black
crystals grew on the surface of the anode. Because the crys-
tals were grown in a molten flux, the natural surface of
the crystal was observed, as shown in the photograph. Fi-
gure 1(b) shows a photograph of a crystal removed from
the electrode. The maximum length of the crystal was about
1.5 mm. The EPMA measurements indicated that the Ca/Pd
molar ratio of the as-grown crystal was 0.35. This value
agrees with the calculated value of 0.33 that was assu-
med based on the chemical formula of CaPd3O4. The mea-
sured value of 0.35 did not vary in the direction of crystal
growth. Because an alumina crucible, platinum electrodes,
and chloride solvents were used in the synthesis, Al, Pt,
Na, and Cl could have potentially been incorporated into
the crystal; however, the EPMA measurements showed
no inclusion of these elements. Tiny CaPd3O4 crystals also
grew on the bottom of the crucible, but the size of each
crystal was 0.01 mm or less, which is significantly smaller
than the crystals grown on the anode. Based on this dif-
ference in size, we conclude that the application of the
electrical voltage to the electrodes had a significant effect
on the growth of large crystals. Because positive Pd ions

Synthesis and Characterization of CaPd3O4 Crystals
18
in the solvent were attracted by the cathode during the
synthesis, the cathode was covered with metallic palladium.
An increase in the amount of NaCl used in the synthesis
resulted in a decrease in the amount of metallic palladium
on the cathode; thus, NaCl seemed to inhibit the reduct-
ion of PdO.
The crystal structure of the CaPd3O4 crystal was char-
acterized by the refinement of single-crystal XRD data,
and the results are summarized in Tables 1 and 2. The
data were refined well by assuming a cubic structure of
the space group Pm 3n. The refined lattice constant a was
5.7471(10) Å, and this value agrees well with those re-
ported in previous studies [1-2,4-6]. The crystal structure
was also characterized by powder XRD. Figure 2 con-
tains the powder XRD profiles and the results of the re-
finement of the data. The data were refined well with re-
liability factors Rwp = 8.10, Re = 8.07, and S = 1.004, and
all peaks were indexed as a NaPt3O4-type structure. The
refined a of 5.7471 Å is in fair agreement with that ob-
tained by the refinement of the single-crystal XRD data.
This result demonstrates that all crystals have the same
structure. Figure 3 shows the crystal structure of CaPd3O4.
In this figure, palladium atoms are bonded with neighbor-
ing oxygen atoms; the distance between palladium and
oxygen atoms is 2.0319(2) Å, and each palladium atom is
surrounded by four coplanar oxygen atoms that form a
PdO4 square planar unit. These units are perpendicularly
connected to each other via shared corner oxygen and form
a three-dimensional framework. The large natural surface
shown in Figure 1(b) was confirmed as the (100) plane of
the cubic crystal.
Figure 4 shows the temperature dependence of electri-
cal resistivity, which was measured for an as-grown crys-
Table 1. Crystallographic data and structure refinement data
for CaP d3O4.
Formula weight 423.28 g/mol
Space group Pm3n (No. 223)
a 5.7471(10) Å
V 189.822(6) Å3
Z 2
Abs. coeff. 15.277 mm–1
F(000) 380
Crystal size 80 m  80 m  40 m
Final R indices [I > 2σ(I)] a R1 = 0.0199, wR2 = 0.0439
R indices (all data) R1 = 0.0203, wR2 = 0.0445
a
0c 0
1RFF=-
åå
F () () ; wR2 =
{}
12
22
22 2
00c
wF FwF
é
ùé ù
-
ê
úê ú
ë
ûë û
åå
,
(
)
(
)
2
122
0 0.00771.34wF P

-éù
=+ +
êú
ëû
Table 2. Atomic coordinates and isotropic displacement pa-
rameters (Å2) for CaPd3O4.
x y z Uiso
Ca 0 0 0 0.0078(3)
Pd 1/4 0 1/2 0.0053(2)
O 1/4 1/4 1/4 0.0075(6)
Figure 2. Powder X-ray diffraction profile of CaPd3O4 crys-
tals and the results of refinement of the diffraction data by
the Rietveld method.
Figure 3. Crystal structure of CaPd3O4. Palladium atoms
are bonded with four neighboring oxygen atoms, which are
represented by gray bars. The solid line shows the unit cell.
P, where ()
22
0c
max, 0 23.PF F
é
=+
ê
ëù
ú
û
tal with a size of 0.103 × 0.130 × 0.647 mm3. The current
was flowed parallel to the [100] direction of a rectangular
parallelepiped crystal. The resistivity showed semicon-
ductor-like characteristics; the resistivity decreased as the
temperature increased. The resistivity at 300 K was 0.10
·cm. Because this measurement was performed with a
single crystal, the result was free from the effects of the

Synthesis and Characterization of CaPd3O4 Crystals 19
grain boundary scattering. The inset of Figure 4 shows an
Arrhenius plot. The data at lower temperatures can be fit-
ted by the thermal-activation process expressed by
a
0
B
exp ,
E
kT




(1)
where 0

, Ea, and kB are the pre-exponential factor, the
activation energy, and the Boltzmann constant, respecti-
vely. An Ea of 0.45 meV was obtained; this value is
lower than that obtained for sintered material (1.1 meV)
[2] and is inconsistent with the result of the band calcula-
tion, which predicted a semimetallic band structure with
a zero band gap [3].
Figure 5 shows the temperature dependence of the mo-
lar magnetic susceptibility
 for as-grown crystals and the
crystals annealed at 500˚C for 7 days under an oxygen
atmosphere of 1 atm. The measurements were conducted
in an applied magnetic field of 10 kOe during the heating
process after zero-field-cooling (ZFC) and field-cooling
(FC) process. The FC data were nearly identical to the ZFC
data and are not shown. Both plots illustrate paramagnetic
behavior with an increase of the susceptibility at low tem-
peratures; χ can be fitted by the modified Curie-Weiss
law expressed by
mol
0,
C
TΘ
 
 (2)
where 0

, Cmol, and  are the temperature-independent
magnetic susceptibility, the molar Curie constant, and the
asymptotic Curie temperature, respectively. The solid line
shows the results which were obtained by fitting the mo-
dified Curie-Weiss formula to the data, using the Mar-
quardt-Levenberg algorithm [16]. The values 04.1



emu/(mol·Oe), Cmol = 5.0(1) × 03 emu K/
(mol·Oe), and
 = –1.4(2) K were obtained for the as-
grown crystals. The effective number of the Bohr mag-
neton per chemical formula unit p was calculated using
the formula

215 0

Figure 4. Temperature dependence of the electrical resistiv-
ity of an as-grown crystal. The inset of the figure shows the
Arrhenius plot.
Figure 5. Temperature dependence of the molar magnetic
susceptibility
  measured for as-grow n crystals and oxygen-
annealed crystals (10 kOe, ZFC).
22
B
mol
B
,
3
Np
Ck

 (3)
where N, μB, and kB are the Avogadro constant, the Bohr
magneton, and the Boltzmann constant, respectively. The
value of p was determined to be 0.20/f.u. In CaPd3O4, Pd
is the only magnetic ion. Based on the stoichiometry of the
crystal, the valence of the Pd ion is 2+, and the electron
configuration is . In this structure with PdO4 tetragons,
the
8
4d
22
x
y
d-energy level is the highest, and the 2
z
d, dxy, dyz,
and dzx levels are filled by electrons in a low-spin con-
figuration [8]. The band calculation also suggests that the
empty 22
x
y- forms the highest band [3]. Based on this
configuration, diamagnetism was expected for CaPd3O4,
and negative magnetic susceptibility was actually observed
at a high temperature range; however, the Curie-Weiss pa-
ramagnetic contribution was observed, suggesting the
existence of localized spin moments. Because the meas-
urements were performed for single crystals, the Curie-
Weiss paramagnetic behavior was not the effect of impu-
rity magnetic phases. Because the paramagnetic contribu-
tion didn’t decrease after the oxygen annealing of the
crystals, the Curie-Weiss paramagnetic behavior was not
the effect of the oxygen vacancies. The Curie-Weiss pa-
ramagnetic behavior is attributable to the existence of Pd3+,
which was induced by the existence of cation vacancies.
As previously reported, the molar ratio of Ca/Pd was de-
termined to be 0.35 by EPMA; this value is slightly lar-
ger than the value of 0.33 that is based on the stoichiome-
tric composition. This disparity suggests that Pd vacan-
cies are present, which would introduce Pd3+ in order to
maintain the charge neutrality of the system [4,6,17]. In
this case, the Curie-Weiss paramagnetic behavior would
be induced by the unpaired spin at the dxy level of Pd3+;
however, the exact cause of the Curie-Weiss paramagnetic
behavior remains unclear at present.
d

Synthesis and Characterization of CaPd3O4 Crystals
20
4. Conclusion [7] T. Taniguchi, Y. Nagata, T. C. Ozawa, M. Sato, Y. Noro,
T. Uchida and H. Samata, “Insulator-Metal Transition In-
duced in Sr1–xNaxPd3O4 for Small Na-Substitutions,”
Jour- nal of Alloys and Compounds, Vol. 373, No. 1-2,
2004, pp. 67-72. doi:10.1016/j.jallcom.2003.11.004
Single crystals of alkaline-earth palladate CaPd3O4 were
grown by an electrochemical technique that used voltage-
applied platinum electrodes inserted into a molten chlo-
ride solvent. Crystals with natural surfaces were obtained
on the anode, reflecting that the growth proceeded in a
liquid phase. The maximum length of the crystal was about
1.5 mm. The crystal structure parameters were refined well
by assuming a cubic structure of the space group Pm 3n,
and the lattice constant a was 5.7471(10) Å. The resistiv-
ity showed semiconductor-like characteristics with a very
small activation energy Ea of 0.45 meV at lower tempe-
ratures. The temperature dependence of the molar mag-
netic susceptibility showed the Curie-Weiss paramagne-
tic behavior, indicating the existence of localized spin mo-
ments.
[8] T. C. Ozawa, A. Matsushita, Y. Hidaka, T. Taniguchi, S.
Mizusaki, Y. Nagata, Y. Noro and H. Samata, “Synthesis
and Characterization of Electron and Hole Doped Ternary
Palladium Oxide: Sr1-xAxPd3O4 (A = Na, Bi),” Journal of
Alloys and Compounds, Vol. 448, No. 1-2, 2008, pp. 77-
83. doi:10.1016/j.jallcom.2007.03.137
[9] J. Kitagawa, T. Sasakawa, T. Suemitsu, Y. Echizen and T.
Takabatake, “Effects of Valence Fluctuation and Pseu-
dogap Formation on Phonon Thermal Conductivity of
Ce-Based Compounds with ε-TiNiSi-Type Structure,” Phy-
sical Review B, Vol. 66, No. 22, 2002, p. 224304.
doi:10.1103/PhysRevB.66.224304
[10] A. Saramat, G. Svensson, A. E. C. Palmqvist, C. Stiewe,
E. Mueller, D. Platzek, S. G. K. Williams and D. M.
Rowe, “Large Thermoelectric Figure of Merit at High Tem-
perature in Czochralski-Grown Clathrate Ba8Ga16Ge30,”
Journal of Applied Physics, Vol. 99, No. 2, 2006, p.
023708. doi:10.1063/1.2163979
5. Acknowledgements
The work performed at Kobe University and Aoyama Ga-
kuin University was supported by Grants-in-Aid for Sci-
entific Research from the Ministry of Education, Culture,
Sports, Science and Technology of Japan.
[11] P. L. Smallwood, M. D. Smith and H.-C. zur Loye, “Flux
Synthesis of Alkaline Earth Palladates,” Journal of Crystal
Growth, Vol. 216, No. 1-4, 2000, pp. 299-303.
doi:10.1016/S0022-0248(00)00432-2
REFERENCES [12] T. N. Nguyen and Hans-Conrad zur Loye, “Electrosyn-
thesis in Hydroxide Melts,” Journal of Crystal Growth,
Vo. 172, No. 1-2, 1997, pp. 183-189.
doi:10.1016/S0022-0248(96)00726-9
[1] R. C. Wnuk, T. R. Touw and B. Post, “The Crystal Struc-
ture of CaPd3O4,” IBM Journal of Research and Devel-
opment, Vol. 8, No. 2, 1964, pp. 185-186.
doi:10.1147/rd.82.0185 [13] H. Samata, Y. Saeki, S. Mizusaki, Y. Nagata, T. C. Oza-
wa and A. Sato, “Electrochemical Crystal Growth of Pe-
rovskite Ruthenates,” Journal of Crystal Growth, Vol.
311, No. 3, 2009, pp. 623-626.
doi:10.1016/j.jcrysgro.2008.09.042
[2] K. Itoh and N. Tsuda, “Metal to Semiconductor Like
Transition for Sintered Ca1–xNaxPd3O4,” Solid State
Com- munications, Vol. 109, No. 11, 1999, pp. 715-719.
doi:10.1016/S0038-1098(98)00549-3
[14] G. M. Sheldrick, SHELXTL, Version 6.10, Bruker AXS
Inc., Madison, 1997.
[3] I. Hase and Y. Nishihara, “CaPd3O4, as an Excitonic In-
sulator,” Physical Review B, Vol. 62, No. 20, 2000, pp.
13426-13429. doi:10.1103/PhysRevB.62.13426 [15] F. Izumi and T. Ikeda, “A Rietveld-Analysis Programm
RIETAN-98 and Its Applications to Zeolites,” Materials
Science Forum, Vol. 321-324, 2000, pp. 198-205.
doi:10.4028/www.scientific.net/MSF.321-324.198
[4] S. Ichikawa and I. Terasaki, “Metal-Insulator Transition
in Ca1–xLixPd3O4,” Physical Review B, Vol. 68, 2003, p.
233101. doi:10.1103/PhysRevB.68.233101
[16] W. H. Press, B. P. Flannery, S. A. Teukolsky and W. T. Vet-
terling, “Numerical Recipes,” Cambridge University Press,
Cambridge, 1986.
[5] K. Itoh, Y. Yano and N. Tsuda, “Metal to Insulator Tran-
sition for Ca1–xNaxPd3O4,” Journal of the Physical Soci-
ety of Japan, Vol. 68, No. 9, 1999, pp. 3022-3026.
doi:10.1143/JPSJ.68.3022 [17] S. J. Kim, S. Lemaux, G. Demazeau, J. Y. Kim and J. H.
Choy, “LaPdO3: The First PdIII Oxide with the Perovskite
Structure,” Journal of the American Chemical Society,
Vol. 123, No. 42, 2001, pp. 10413-10414.
doi:10.1021/ja016522b
[6] Y. Yano, M. Kanazawa, T. Fujii and N. Tsuda, “Magnetic
Susceptibility of Ca1-xNaxPd3O4,” Journal of the Physi-
cal Society of Japan, Vol. 70, No. 6, 2001, pp. 1772-1176.
doi:10.1143/JPSJ.70.1772