ARTICLE IN PRESS
M. Takeda et al. / Journal of Solid State Chemistry 179 (2006) 2823–2826
2824
2
. Experimental
We prepared powders of CaB and SrB via borothermal
reduction of metal oxides, i.e. CaO and SrO, expressed as
6
6
MO þ 7B ! MB6 þ BO " . (2)
As for preparing BaB6 powder, we synthesized the
powder according to the following reaction using BaCO3
instead of MO:
BaCO3 þ 7B ! BaB6 þ BO " þCO2 " .
(3)
Mixtures of the oxide or carbonate and boron powders
were pressed into pellets, and heated at 1743 K in a
vacuum. The resultant pellets were ground into powder,
and subsequently the powder was sintered at 2073–2273 K
and 50 MPa for 20 min using the ‘‘Pulse Electric Current
Sintering’’ system. We also synthesized Ca1 Sr B alloys
ꢀx
x 6
from the mixture of CaO, SrO and boron powders by the
same procedure as described above except for the sintering
condition. The sintering temperature was 2073 K, and
holding time was prolonged by 50 min to ensure the
uniformity of the specimen. CaB and SrB , namely x ¼ 0
6
6
Fig. 1. Temperature dependence of (a) the Seebeck coefficient and
b) electrical conductivity for alkaline-earth hexaborides.
and 1, respectively, were prepared by the same condition as
that of alloys for comparison. We refer to the two
specimens hereafter as CaB * and SrB * to distinguish
(
6
6
Hall coefficient, R , on the basis of the free electron
H
them from the CaB6 and SrB6 prepared by shorter
sintering. X-ray diffraction (XRD), scanning electron
microscopy (SEM), and energy dispersive X-ray spectro-
scopy (EDX) were used to characterize the specimens.
Electrical conductivity (s) and Seebeck coefficient (a)
were measured by a standard four-probe method and a
steady-state temperature gradient method, respectively.
The thermal conductivity (k) was calculated from the
thermal diffusion coefficient and heat capacity measured
by the laser flash method. Those measurements were
carried out from room temperature to 1073 K. Hall
measurements were made using the Van der Pauw method
at room temperature by applying a magnetic field up
to 1 T.
approximation, R ¼ 1/en. Fig. 2 shows the a, s, and
H
2
power factor given by a s as a function of n. These values
were obtained at room temperature. The a decreases
linearly with the increasing logarithm of n. The s lies in a
straight line with a slope of 1, indicating almost the same
Hall mobility for the three specimens. These results suggest
that TE properties of the alkaline-earth hexaboride depend
largely on carrier concentration, namely, on the quality of
the specimen. Consequently, the power factor of the
hexaboride appears to have a maximum value at a carrier
26 ꢀ3
concentration of around 2 ꢁ 10 m
.
Thermal conductivity (k) of the alkaline-earth hexabor-
ide was higher than 10 W/m K even at 1073 K, which is
relatively high compared with other TE materials, and
lattice contribution to the k (klatt) was dominant [12]. The
k, particularly klatt, should be reduced to improve TE
performance of the hexaboride. Alloying is an effective way
to reduce klatt, and has been employed to improve the TE
performance of materials such as Si–Ge system [17]. As
discussed above, electrical properties of the alkaline-earth
hexaboride appeared to be largely unaffected by constitu-
ent metal atoms, and hence, alloying of hexaboride using
two or more alkaline-earth metals is expected to reduce k
without seriously lowering the electrical performance.
3
. Results and discussion
Fig. 1 shows the temperature dependence of a and s for
CaB , SrB , and BaB . The negative a indicates that the
6
6
6
hexaborides prepared in this study are n-type materials.
The s decreases with increasing temperature for all
specimens, indicating metallic behavior. Although theore-
tical study [13] and angle-resolved photoemission measure-
ment [14] suggested that the CaB6 is essentially a
semiconductor, such metallic behavior of s has been
reported [14–16], and the values of s are scattered among
the reports. The origin of the n-type carrier and differences
in s appear to be attributable to impurity and defects in the
hexaborides. As a result, the differences in a and s for the
alkaline-earth hexaborides synthesized in the present study
can be attributed to the difference in carrier concentration,
n. We calculated n of the hexaborides from the measured
We synthesized (Ca,Sr)B alloys and confirmed that the
6
alloys had hexaboride structures according to XRD
spectra, and no phase separation was observed by SEM
and EDX. Fig. 3 shows the temperature dependence of a
and s for the (Ca,Sr)B alloys together with those for
6
CaB * and SrB * synthesized by the same condition as the
alloys. The a and s of the alloys are intermediate between
6
6
those of the CaB * and the SrB *, and the values for all
6
6
Takeda, Masatoshi
