Full text of DOI:10.1557/JMR.2002.0161

High-temperature thermoelectric properties of the
Ca1−xBixMnO3 system
Gaojie Xu, Ryoji Funahashi,a) Ichiro Matsubara, Masahiro Shikano, and Yuqin Zhou
National Institute of Advanced Industrial Science and Technology, 1-8-31 Midorigaoka, Ikeda,
Osaka 563-8577, Japan
(Received 6 September 2001; accepted 14 February 2002)
Polycrystalline samples of Ca1−xBixMnO3 (0.02 ഛ x ഛ 0.20) were studied by means of
x-ray diffraction, electrical resistivity (␳), thermoelectric power (S), and thermal
conductivity (␬) at high temperature. Bi doping leads to the lattice parameters a, b, and
c increasing. And the ␳ and the absolute value of S decrease rapidly with Bi doping.
The largest power factor, S2/␳, is obtained in the x ס 0.04 sample, which is
3.6 × 10−4 W m−1 K−2 at 400 K. The figures of merit (Z ס S2/␳␬) for this sample and
1.0 × 10−4 and 0.86 × 10−4 K−1 at 600 and 1000 K, respectively.
I. INTRODUCTION
on each other, optimization of the TE performance
always requires a compromise between them. The adjust-
ment of carrier concentration by doping is a feasible
route to optimizing the TE performance. Our experimen-
tal results indicate that the highest power factor (PF),
S2/␳, and the figure of merit, Z, were reached at an op-
timal carrier density near x ס 0.04 for Ca1−xMxMnO3
(M: Bi3+, La3+, etc.).22 And also the Bi-doped CaMnO3
system was found to show the best TE performance
among these systems. In this paper, a series of samples of
Ca1−xBixMnO3 were synthesized and the composition-
dependent thermoelectric properties at high temperature
were investigated.
Thermoelectrics, namely energy conversion between
heat and electricity via thermoelectric phenomena in sol-
ids, has long been a fundamental issue in condensed mat-
ter physics and is currently attracting a renewed interest
as a promising energy-conversion technology that is not
harmful to the environment. Performance of thermoelec-
tric (TE) materials can be evaluated by a figure of merit
defined as Z ס S2/␳␬ (or ZT ס S2T/␳␬; T, temperature),
where S, ␳, and ␬ are the Seebeck coefficient, electrical
resistivity, and thermal conductivity, respectively. In
comparison with conventional TE materials such as
metal chalcogenides,1,2 transition metal disilicides,3–5
and Si–Ge alloys,6–8 metal oxides, due to their high ther-
mal and chemical stability, have attracted more and more
attentions. Among the oxide TE materials, cobalt oxides
with layered structure [such as NaCo2O4 (Co-124),
Ca3Co4O9 (Co-349), and Bi2Sr2Co2Oy (BC-222)] have
been found to be strong p-type candidates.9–13 For n-type
TE materials, some candidates, including (ZnO)5In2O3,
In2O3 и MO (M ס Cr, Mn, Ni, Zn, Y, Nb, Sn), (Zn,
Al)O, and (Ba, Sr)PbO3, have been systematically stud-
ied.14–17 As a member of the large family of perovskite-
type oxides, the electron-doped CaMnO3 system has
also been suggested to be potential material for
high-temperature thermoelectrics.18,19 Furthermore, the
one oxide TE module had been fabricated using
Ca0.92La0.08MnO3 n-type legs and Ca2.75Gd0.25Co4O9 p-
type legs by our group20 and has been proved to be
operable for a few weeks in air showing high durability.
However, for the electron-doped CaMnO3 system, only a
few systematical studies about carrier concentration
dependence of high-temperature TE properties were re-
ported.21 Since the parameters S, ␳, and ␬ are dependent
II. EXPERIMENTAL
Polycrystalline samples of Ca1−xBixMnO3
(0.02 ഛ x ഛ 0.2) were synthesized by the conventional
solid-state reaction using powders CaCO3 (99.99%),
Bi2O3 (99.9%), and Mn2O3 (99.9%) in stoichiometric
ratio. Bi2O3 was added in 10% excess to compensate for
the loss of bismuth oxide due to sublimation. Initially the
appropriate mixture of these powders was mixed well
and calcined at 1173 K for 12 h and 1373 K for 12 h in
air with an intermediate grinding. Then the calcined
samples were reground and pressed into dish-shaped pel-
lets and sintered in air at 1473–1523 K for 15 h. Finally
the pellets were cooled to room temperature in the
furnace.
X-ray powder diffraction (XRD) analysis was carried
out with Rigaku diffractometer using Cu K␣ radiation.
Cationic composition analysis was made with a Horiba
EMAX-5770 energy-dispersive x-ray analysis (EDX)
system. The lattice parameters were determined from the
d values of XRD peaks using a standard least-squares
refinement method. ␳ was measured using a dc stan-
dard four-probe method within temperature range of
a)Address all correspondence to this author.
e-mail: funahashi-r@aist.go.jp
1092
J. Mater. Res., Vol. 17, No. 5, May 2002
© 2002 Materials Research Society
http://journals.cambridge.org
Downloaded: 18 Mar 2015
IP address: 129.93.24.73

G. Xu et al.: High-temperature thermoelectric properties of the Ca1−x Bix MnO3 system
300–1100 K in air. Pt paste was used for the terminal
connections. S was obtained from the slope of the rela-
tion between thermoelectromotive forces and tempera-
ture differences with an error less than 3%. ␬ was
determined from the thermal diffusivity and the specific
heat capacity measured by a laser flash technique using
Rigaku FA-8510B thermal constant measurement system
with a total error less than 5%.
III. RESULTS AND DISCUSSION
XRD analysis indicates that all the samples of
Ca1−xBixMnO3 (0.02 ഛ x ഛ 0.2) are single phase with
orthorhombic symmetry. EDX result shows that the cat-
ion stoichiometries are very close to the target values.
Figure 1 shows the lattice parameters a, b, and c and the
cell volume V as a function of x for Ca1−xBixMnO3. With
Bi doping, the parameters a, b, and c all increase mo-
notonously, and the cell volume increases from 208 to
214 Å3 as x changes from 0.02 to 0.2. This result can be
well understood on the basis of the fact that the ionic
radius of Bi3+ is larger than that of Ca2+.
Figure 2 presents the temperature dependence of ␳
within 300–1100 K. It is known that the undoped sample
CaMnO3 is an n-type semiconductor with rather high ␳
(approximately 10 ⍀ cm) at room temperature. However,
Bi doping, even as x ס 0.02, can cause a marked
FIG. 2. Electrical resistivity (␳) as a function of temperature of
Ca1−x Bix MnO3 (0.02 ഛ x ഛ 0.20).
decrease in ␳ in 2–3 orders of magnitude at room tem-
perature. With further increasing Bi content, the ␳ de-
creases gradually. Obviously, substituting Bi for Ca
introduces carrier into the ␴* conduction band and leads
to a corresponding decrease in ␳ as the carrier concen-
tration increases.23 On the other hand, the increase of
density with Bi doping (3.73 and 4.52 g cm−3 for
x ס 0.02 and 0.20, respectively) is another reason for the
decrease in ␳.
Figure 3 shows the temperature dependence of S for
the samples of Ca1−xBixMnO3. The S values are all nega-
tive, indicating n-type conduction. For all the samples,
FIG. 1. Lattice parameters a, b, and c and the cell volume V of
Ca1−x Bix MnO3 (0.02 ഛ x ഛ 0.20).
FIG. 3. Temperature dependence of thermoelectric power (S) of
Ca1−x Bix MnO3 (0.02 ഛ x ഛ 0.20).
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
1093
http://journals.cambridge.org
IP address: 129.93.24.73

G. Xu et al.: High-temperature thermoelectric properties of the Ca1−x Bix MnO3 system
the absolute values of S (|S|) increase with increasing
temperature over the whole measured temperature range.
Bi doping results in a considerable decrease in the |S|,
which is in qualitative agreement with the decrease in ␳.
Even so, the |S| is still moderate as 100–200 ␮V K−1 at
1000 K for the x ഛ 0.1 samples. For x ס 0.02 and 0.04
samples, the |S| values are 220 and 176 ␮V K−1, respec-
tively at 1000 K.
The samples with 0.02 ഛ x ഛ 0.1 all have rather low ␳
and large |S| as oxides, which is required for thermoelec-
tric applications. Due to the ␳ and |S| varying in a
consistent way, optimizing the carrier concentration by
modifying the compositions is needed to get the best TE
performance. The calculation of PF can help us to evalu-
ate the electrical components of the TE performance.
Figure 4 shows the PF versus temperature for
Ca1−xBixMnO3. Clearly, the x ס 0.04 sample possesses
the highest PF, i.e., 3.6 × 10−4 and 2.7 × 10−4 W m−1 K−1
at 400 and 1000 K, respectively. An increase or a
decrease in x results in a marked decrease in the PF,
which is similar to the results observed in Ca1−xLaxMnO3
system.22
Figure 5 shows the temperature dependence of ␬. For
comparison, the data for the undoped CaMnO3 from the
work of Ohtaki et al.18 were plotted in this figure. The ␬
value of the x ס 0.04 sample at 300 K is 3.5 W m−1 K−1,
slightly less than that of the undoped sample. At the same
temperature the ␬ of the x ס 0.15 sample is only
3.0 W m−1 K−1, markedly less than that of the x ס 0 and
0.04 samples. The ␬ of the x ס 0.15 sample increases
gradually from 3.0 to 3.6 W m−1 K−1 as the temperature
increases from 300 to 1000 K. While for the other two
samples, the ␬ decreases with increasing temperature
FIG. 5. Temperature dependence of thermal conductivity (␬) of
Ca1−x Bix MnO3.
within the whole measured temperature range. With Bi
doping the slope, d␬/dT, changes from negative to posi-
tive. It is known that ␬ can be expressed by the sun of a
lattice component (␬1) and an electronic component (␬e)
as ␬ ס ␬ + ␬e.24,25 ␬ is proportional to T−1 above
1
1
Debye temperature and ␬e is proportional to ␦T (␦: elec-
trical conductivity). For the Ca1−xBixMnO3 system, on
one hand, the doping of heavy element Bi can effectively
decrease ␬1; on the other hand, the increase of carrier
concentration induced by Bi doping leads to the increase
in ␬e. At room temperature the ␬ is small and the ␬
e
l
should dominate the ␬. So the markedly decrease of
␬ leads to the ␬ decreases with increasing x. With in-
l
creasing temperature the ␬ decreases and ␬ increases
l
e
(the ␴T increases with increasing temperature for all the
Ca1−xBixMnO3 samples). In consequence, the d␬/dT
changes to the positive value. In fact, the change of car-
rier concentration is one factor to cause the d␬/dT
changing from negative to positive. The true situation,
however, may be more complicated. Because we have
estimated the value of ␬ according to Wiedermann–
e
Franz’s law using Lorenz number for free electrons
(2.45 × 10−8 W ⍀ K−2) and found the ␬ is too small to
e
change the ␬ so much at high temperature. Some other
factors, such as density, may influence ␬. This study
indicates that the quite large difference in density be-
tween the x ס 0.04 (3.78 g/cm3) and x ס 0.15 (4.41 g/
cm3) samples mainly originates from different closed
pores. More closed pores could influence ␬ very much
and lead to a different temperature dependence of ␬.26
Table I summarizes the results of electrical and ther-
mal transport measurements at 300 and 1000 K. It clearly
shows that the parameters S, ␳, and ␬ are dependent on
FIG. 4. Power factor as a function of temperature of Ca1−x Bix MnO3.
1094
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
http://journals.cambridge.org
IP address: 129.93.24.73

G. Xu et al.: High-temperature thermoelectric properties of the Ca1−x Bix MnO3 system
TABLE I. Resistivity (␳ and ␳1000), thermoelectric power (S300 and S1000), thermal conductivity (␬ and ␬1000) at 300 and 1000 K, power
300
300
factor (PF), and dimensionless figure of merit (ZT ) at 1000 K of the Ca1−xBixMnO3 (0.02 ഛ x ഛ 0.20) system.
␳
␳
S300
S1000
␬
␬
1000
(W m−1 K−1
PF1000
ZT
(1000 K)
300
1000
300
x
(m⍀ cm)
(m⍀ cm)
(␮V K−1
)
(␮V K−1
)
(W m−1 K−1
)
)
(10−4 W m−1 K−2
)
0.02
0.04
0.06
0.08
0.10
0.15
0.20
20.6
5.9
4.8
4.6
4.2
3.1
3.3
36.5
11.5
10.1
8.9
8.0
5.8
−174
−142
−115
−84
−77
−37
−220
−177
−140
−112
−99
и и и
3.6
и и и
и и и
и и и
3.0
и и и
3.1
и и и
и и и
и и и
3.6
1.35
2.64
1.97
1.41
1.24
0.90
0.95
и и и
0.086
и и и
и и и
и и и
−70
−66
0.025
и и и
4.6
−25
и и и
и и и
ACKNOWLEDGMENT
This work was supported in part by the Industrial
Technology Research Grant Program in 2001 from the
New Energy and Industrial Technology Development
Organization (NEDO) of Japan.
REFERENCES
1. J.F. Nakahara, T. Takeshita, M.J. Tschetter, B.J. Beaudry, and
K.A. Gshneidner Jr., J. Appl. Phys. 63, 2331 (1988).
2. A. Boyer and E. Cisse, Mater. Sci. Eng. B 113, 103 (1992).
3. R.M. Ware and D.J. McNeill, Proc. IEEE 111, 178 (1964).
4. T. Kojima, Phys. Status Solidi A 111, 233 (1989).
5. I. Nishida, Phys. Rev. B 7, 2710 (1970).
6. G.A. Slack and M.A. Hussain, J. Appl. Phys. 70, 2694 (1991).
7. C.B. Vining, J. Appl. Phys. 69, 331 (1991).
8. F.D. Rosi, Solid State Electron. 11, 833 (1968).
9. I. Terasaki, Y. Sasago, and K. Uchinokura, Phys. Rev. B 56,
R12685 (1997).
10. R. Funahashi, I. Matsubara, H. Ikuta, T. Takeuchi, U. Mizutani,
and S. Sodeoka, Jpn. J. Appl. Phys. 39, L1127 (2000).
11. S. Li, R. Funahashi, I. Matsubara, K. Ueno, S. Sodeoka, and
H. Yamada, Chem. Mater. 12, 2424 (2000).
12. R. Funahashi, I. Matsubara, and S. Sodeoka, Appl. Phys. Lett. 76,
2385 (2000).
13. R. Funahashi and I. Matsubara, Appl. Phys. Lett. 79, 362 (2001).
14. M. Kazeoka, H. Hiramatsu, W.S. Seo, and K. Koumoto, J. Mater.
Res. 13, 523 (1998).
15. M. Ohtaki, D. Ogura, K. Eguchi, and H. Arai, J. Mater. Chem. 4,
653 (1994).
16. M. Ohtaki, T. Tsubota, K. Eguchi, and H. Arai, J. Appl. Phys. 79,
1816 (1996).
17. M. Yasukawa and N. Murayama, J. Mater. Sci. Lett. 16, 1731
(1997).
18. M. Ohtaki, H. Koga, T. Tokunaga, K. Eguchi, and H. Arai,
J. Solid State Chem. 120, 105 (1995).
FIG. 6. Temperature dependence of figure of merit Z (᭿) and ZT (᭡)
of Ca0.96Bi0.04MnO3.
each other and a compromise between them is reached at
an optimal carrier concentration. The best thermoelectric
performance was obtained in the x ס 0.04 sample.
The relationship between figure of merit Z and tem-
perature of Ca0.96Bi0.04MnO3 is presented in Fig. 6. The
maximum Z values obtained here are 1.0 × 10−4 and
0.86 × 10−4 K−1 at 600 and 1000 K, respectively. Ac-
cordng to the trends shown in Fig. 6, the value of ZT can
reach 0.1 at 1173 K. The good TE performance and the
excellent high-temperature durability in air of
Ca1−xBixMnO3 suggest the oxides are of interest as po-
tential high-temperature TE materials.
19. T. Kobayashi, H. Takizawa, T. Endo, T. Sato, H. Taguchi, and
M. Nagao, J. Solid State Chem. 92, 116 (1991).
20. I. Matsubara, R. Funahashi, T. Takeuchi, and S. Sodeoka, Appl.
Phys. Lett. 78, 3627 (2001).
21. S. Hashimoto and H. Iwahara, Mater. Res. Bull. 35, 2253 (2000).
22. Gaojie Xu (unpublished).
23. P.N. Santhosh, J. Goldberger, P.M. Woodward, T. Vogt,
W.P. Lee, and A.J. Epstein, Phys. Rev. B 62, 14928 (2000).
24. T. Tsubota, M. Ohtaki, K. Eguchi, and H. Arai, J. Mater. Chem. 7,
85 (1997).
25. M. Yasukawa and N. Murayama, J. Mater. Sci. 32, 6489 (1997).
26. R. Funahashi (unpublished).
IV. CONCLUSION
The high-temperature thermoelectric properties of
Ca1−xBixMnO3 (0.02 ഛ x ഛ 0.20) polycrystalline
samples were investigated carefully. Bi doping causes a
rapid decrease in ␳. The samples with 0.22 < x ഛ 0.10
show rather low ␳ and fairly large |S| (100–200 ␮V K−1),
suggesting promising TE materials. The highest PF was
reached in the x ס 0.04 sample, and Z for this sample is
1.0 × 10−4 K−1 at 600 K and 0.86 × 10−4 K−1 at 1000 K.
J. Mater. Res., Vol. 17, No. 5, May 2002
Downloaded: 18 Mar 2015
1095
http://journals.cambridge.org
IP address: 129.93.24.73