Materials Transactions, Vol. 47, No. 6 (2006) pp. 1458 to 1463
Special Issue on Thermoelectric Conversion Materials II
#2006 The Thermoelectrics Society of Japan
Thermoelectric Properties of Cesium–Graphite Intercalation Compounds
Rika Matsumoto1 , Noboru Akuzawa2 and Yoichi Takahashi3
1
Faculty of Engineering, Tokyo Polytechnic University, Atsugi 243-0297, Japan
Department of Chemical Science and Engineering, Tokyo National College of Technology, Tokyo 193-0997, Japan
3
Department of Applied Chemistry, Chuo University, Tokyo 112-8551, Japan
2
The thermoelectric properties of cesium–graphite intercalation compounds (Cs-GICs) and some other GICs were studied. The electrical
conductivities of the Cs-GICs prepared from Grafoil were around 106 1 m1 , about 10 times higher than that of the host graphite, and the
thermal conductivities were 50–100 Wm1 K1 , about 1/4–1/2 of that of the host graphite. While the Seebeck coefficient of the host graphite
was nearly zero, those of the Cs-GICs were found to be around 30 mV K1 . These properties were determined also for the Cs-GICs prepared
from well-oriented PGS graphite sheets. Consequently, the figures of merit of the Cs-GICs were of the order of 105 K1 , approximately 105
times higher than that of the host graphite. The power factor of the Cs-GICs reached the order of 103 Wm1 K2 , with the best datum at
6:5 103 Wm1 K2 . The carrier density of the GICs increased as the concentration of intercalated species increased, whereas their lattice
thermal conductivities decreased. It is suggested that highly concentrated GICs might be promising candidate thermoelectric materials.
[doi:10.2320/matertrans.47.1458]
(Received November 20, 2005; Accepted April 23, 2006; Published June 15, 2006)
Keywords: graphite intercalation compounds, thermoelectric, electrical conductivity, thermal conductivity, Seebeck coefficient, power factor,
figure of merit, cesium
1.
Introduction
Intercalation
Recently, interest in thermoelectric materials has been
growing and there have been extensive efforts to discover
excellent materials. Many candidate thermoelectric materials, such as oxides, silicides, heuslers, chalcogenides,
skutterudites, and clathrates etc., have been investigated. In
this paper, a novel candidate thermoelectric material, a
graphite intercalation compound, is introduced.
Graphite intercalation compounds (GICs) are graphite
derivatives that contain other chemical species (intercalates)
in their interlayer spaces, as shown in Fig. 1. More than 100
intercalate species have been found so far,1) such as alkali
metals, halogens, metal chlorides, etc. As charge transfer
occurs between the graphite planes and the intercalates, GICs
usually have high electrical conductivities. GICs have
attracted much attention for many years as lightweight
conductive materials. In addition to the high electrical
conductivity, GICs have relatively high Seebeck coefficients2–5) and low thermal conductivities2,6–8) compared with
their host graphite materials. It is quite likely that GICs can
be used as thermoelectric materials. GICs have characteristics derived from their host graphite materials; thus, GICs
are lightweight and anisotropic, and come in arbitrary shapes,
such as sheets, blocks and powders. Also, both n-type and
p-type GICs are available depending on the types of
intercalates. We feel that these characteristics make GICs
particularly attractive as thermoelectric materials.
However, since GICs have rarely been investigated as
thermoelectric materials, there are only a few reports of
thermoelectric data on GICs. Generally, it is very difficult to
obtain thermoelectric data on GICs due to their instability in
air. In this study, relatively air-stable GICs were chosen and
their thermoelectric properties were measured; then, their
potential as thermoelectric materials was discussed.
Usually, the thermoelectric performance of a material is
evaluated using the figure of merit, Z ( 2 =), the
Graphite
Intercalate
Fig. 1
GIC
Formation of GIC.
dimensionless figure of merit, ZT ( Z T), or the power
factor, P ( 2 ), where is the electrical conductivity, is the Seebeck coefficient, and is the thermal conductivity.
Since the working temperature is important for the practical
use of thermoelectric materials, the dimensionless figure of
merit, ZT, is often used. The power factor, P, expresses only
the electronic contribution among the thermoelectric properties, so it is effective when the power conversion efficiency
need not be considered. In this study, we evaluated the ability
of Cs-GICs using the figure of merit and power factor,
because their thermoelectric properties were measured only
at room temperature.
In this paper, we report the thermoelectric properties of
some GICs, mainly Cs-GICs, and their potential as thermoelectric materials. First, we discuss the thermoelectric
properties of Cs-GICs and the dependence of these thermoelectric properties on the intercalates and host graphite
materials; then we show their dependence on the intercalate
concentration. Finally, we propose a method of producing
GICs as promising thermoelectric materials.
2.
Experimental
2.1 Preparation of GICs
Exfoliated graphite sheets prepared from natural graphite,
Grafoil (UCAR Co.), were mainly used as the host graphite
material. Grafoil sheets, 0.30 mm thick, were cut into
3:0 30-mm2 rectangles, and were heat-treated under vacu-
Thermoelectric Properties of Cesium–Graphite Intercalation Compounds
: Graphite plane
: Intercalate plane
1459
Heat shield (Ag)
Heater
DC Power source
Insulated Cu wire
Voltmeter
Sample
Thermocouple Voltmeter
Voltmeter
1st. Stage
CsC8
2nd. Stage 3rd. Stage
CsC24
CsC36
4th Stage
CsC48
Heat sink (Cu)
Vacuum
Fig. 2 Stage structure of Cs-GIC.
Fig. 4 Experimental set-up for measuring thermal conductivity.
Absorption
Binary GIC
Molecule
273 K
Ternary GIC
Fig. 3 Absorption of molecules into GIC.
um at 900 C before use. Well-oriented PGS graphite shees
(Matsushita Electric Industrial Co.), 0.10 mm thick and cut
into 2:0 30-mm2 rectangles, was also used.
GICs were prepared by allowing the graphite to react with
the vapor of the intercalates at specific temperatures under
vacuum (two-bulb method).1,9) In the case of Cs-GICs, their
compositions were CsC8 CsC48 and regulated by the stage
structure,1) as shown in Fig. 2. Non-stoichiometric constitutions such as CsC12 , CsC27 , CsC30 and CsC39 have a mixed
stage structure.
Ethylene (C2 H4 ), benzene (C6 H6 ) or acrylonitrile
(C2 H3 CN) was then absorbed into the vacant space in the
interlayer of the Cs-GICs to improve their air stability, as
shown in Fig. 3, and ternary GICs9–11) such as Cs-ethyleneGIC, Cs-benzene-GIC and Cs-acrylonitrile-GIC were
formed. Their compositions were CsCx (C2 H4 )y , where x ¼
12{48, y ¼ 0:32{1:2, CsC24 (C6 H6 )2:0 and CsC27 (C2 H3 CN)1:2 , respectively. The remarkable air stability of
CsC24 (C2 H4 )x>1:0 was recognized.9–12) It has been found that
the electrical conductivities of these ternary GICs are
somewhat lower than those of their original binary GICs
because of the partial electron back-donation from the
graphite planes to the Cs atoms.13–15)
For the purpose of comparison, K-ethylene-GIC,16)
K-benzene-GIC17,18) and FeCl3 -GIC19,20) were prepared and
studied in a similar manner.
2.2 Measurement of thermoelectric properties
The in-plane electrical conductivities, Seebeck coefficients
and thermal conductivities of the above GICs and their host
graphites were measured at room temperature. Samples cut
into 25-mm-long strips were used for these measurements.
The electrical conductivities were measured by the fourterminal method in air. The thermal conductivities and
Seebeck coefficients were measured under vacuum using
almost the same system as that shown in Fig. 4. The thermal
conductivities were measured by the four-terminal method,
in a way similar to the electrical conductivity measurement.
That is, the electrical current was replaced by a heat flow and
the voltage was replaced by a temperature difference. The
Seebeck coefficients were measured by mounting voltage
terminals on the samples after measuring the thermal
conductivity. The reliability of this apparatus was confirmed
using an isotropic graphite block as the standard, comparing
the values determined by this method and applying the laserflash method.
3.
Results and Discussion
3.1 Thermoelectric properties of Cs-GICs
3.1.1 Cs-GICs prepared from Grafoil
The typical data obtained for the thermoelectric properties
of the Cs-GICs prepared from Grafoil are listed in Table 1,
together with the evaluated values of the power factors and
figures of merit.
(1) Electrical conductivity For the Cs-GICs, the highest electrical conductivity
observed was 1:5 106 1 m1 , over 10 times higher than
that of Grafoil. Table 2 shows our previous results11,21) for
the carrier densities and mobilities estimated from the values
of the Hall coefficients, magnetoresistances, and electrical
conductivities of Grafoil and Cs-ethylene-GIC prepared from
Grafoil. Since the magnetoresistance of Cs-ethylene-GIC
was too small to be detected, it was concluded that the
carriers of electrical conduction on Cs-ethylene-GIC were
almost exclusively electrons. The electron density of Csethylene-GIC was 5:7 1020 cm1 , which is approximately
700 times that of Grafoil, and its electron mobility was
approximately 1/60. This remarkable enhancement of the
electron density is due to the electron transfer from the Cs
atoms to the graphite planes. It has also been observed that
the electrical resistivity of Cs-ethylene-GIC has a metallic
temperature dependence, as shown in Fig. 5—this is a typical
feature of GICs22–24)—while Grafoil has a semiconductive
dependence.14,25,26)
The relatively low electrical conductivities of Cs-benzeneGIC and Cs-acrylonitril-GIC shown in Table 1 were attributed to their lower air stability, because the electrical
conductivity of a sample is strongly affected by its decomposition in air.10,21)
1460
R. Matsumoto, N. Akuzawa and Y. Takahashi
Table 1 Electrical conductivity (), Seebeek coefficient (), thermal conductivity (), power factor (P), and figure of merit (Z) for CsGICs prepared from Grafoil.
/1 m1
Composition
5
(Grafoil)
1:2 10
CsC12 (C2 H4 )0:32
1:1 106
29
6
1:5 10
CsC30 (C2 H4 )0:95
1:0 10
CsC36 (C2 H4 )0:98
6:7 10
CsC39 (C2 H4 )0:99
9:2 10
3:5 1011
50
8:4 104
1:7 105
70
4
7:0 106
3
1:9 105
4
1:1 105
4
8:2 106
4
9:7 106
4
4:9 10
1:2 10
9:9 10
72
30
5
5:9 10
80
29
Z/K1
6:9 10
93
32
5
9
63
29
6
P/Wm1 K2
195
27
5:8 10
CsC24 (C2 H4 )x<1:0
/Wm1 K1
0:24
5
CsC24 (C2 H4 )1:1
/mV K1
7:8 10
CsC48 (C2 H4 )0:82
5
7:6 10
30
94
7:0 10
7:4 106
CsC24 (C6 H6 )2:0
1:7 105
33
71
1:9 104
2:6 106
75
4
4:7 106
5
CsC27 (C2 H3 CN)1:2
2:1 10
41
3:5 10
Table 2 Values of electrical conductivity (), Hall coefficients (RH ), magnetoresistance (=), electron and hole densities ðne ; nh Þ, and,
electron and hole mobilities ðe ; h Þ for Grafoil and Cs-ethylene-GIC prepared from Grafoil at 77 K.
/1 m1
RH /cm3 C1
(at 1T)
=
(at 1T)
ne /cm3
nh /cm3
e /cm2 V1 s1
h /cm2 V1 s1
Grafoil
7:5 104
0:12
0.084
8:3 1017
7:8 1017
2900
2900
CsC24 (C2 H4 )1:4
4:6 105
0:011
—
5:7 1020
—
ρ / µ Ω cm
1000
Grafoil
Cs-ethylene-GIC
100
0
100
200
300
T/K
Fig. 5 Temperature dependence of electrical resistivity for Grafoil and
CsC24 (C2 H4 )1:4 prepared from Grafoil.
(2) Seebeck coefficient The Seebeck coefficient of the host graphite is nearly
zero,27) since the number of electrons and number of holes
(conductive carriers) are almost equal. Since Cs-GICs are ntype materials, their Seebeck coefficients became negative,
approximately 30 mV K1 . The absolute values of the
Seebeck coefficients of GICs at around room temperature
are similar, 25–40 mV K1 ,2–5) although the reported values
for the Seebeck coefficients are considerably fewer than the
51
—
reported electrical conductivity values. These values are of
the order of 1/10 times smaller than those of the other
candidate thermoelectric materials.
(3) Thermal conductivity The thermal conductivities of the Cs-GICs were 50–
100 Wm1 K1 , about 1/4–1/2 of that of Grafoil. The
thermal conductivity of GICs has been investigated in
detail,2,6,7) as described below.
Although the thermal conduction in host graphites is
dominated by the lattice contribution, that in GICs is the
result of both the lattice and carrier contribution. While the
carrier contribution of GICs is higher than that of the host
graphites, the lattice contribution is smaller, because the
intercalation process causes an increase in the number of
lattice defects, as well as an increase in the carrier density. It
has also been found that the lattice contribution dominates in
GICs at around room temperature. As a result, the thermal
conductivities of GICs at around room temperature are low as
compared to those of the host graphites.
(4) Power factor P, and figure of merit Z
Although the Seebeck coefficients of the Cs-GICs were
appreciably smaller than those of the other candidate
thermoelectric materials, the electrical conductivities were
sufficiently higher; therefore, the power factors of the CsGICs were relatively high. The highest value for the Cs-GICs
prepared from Grafoil was 1:2 103 Wm1 K2 , which is
comparable to the values of promising candidate thermoelectric materials.
On the other hand, the figures of merit of the Cs-GICs were
one to two orders of magnitude lower than those of the other
candidates, because the thermal conductivities of the CsGICs were of the order of 10 times higher than those of the
other candidates. The figures of merit of the Cs-GICs were of
the order of 105 K1 , 105 times higher than that of Grafoil.
Thus, in order to improve the thermoelectric performance
Thermoelectric Properties of Cesium–Graphite Intercalation Compounds
-20
-1
α / µV K
-1
(a)
-30
-40
150
-1
κ / Wm K
-1
(b)
100
50
1.0
-3
-1
P / 10 Wm K
-2
(c)
0.5
(d)
1.5
-5
Z / 10 K
-1
2.0
1.0
0.5
0.0
0
5
10
5
-1
15
-1
σ / 10 Ω m
Fig. 6 Variation of (a) Seebeck coefficient , (b) thermal conductivity ,
(c) power factor P, and (d) figure of merit Z with electrical conductivity ,
for Cs-ethylene-GIC ( ) with Cs-benzene-GIC ( ), Cs-acrylonitril-GIC
( ), K-ethylene-GIC ( ), and K-benzene-GIC ( ) prepared from Grafoil.
1461
acrylonitril-GIC, K-ethylene-GIC, and K-benzene-GIC prepared from Grafoil.
As seen in Fig. 6(a), the Seebeck coefficients remain
almost constant with increasing electrical conductivity above
¼ 5:8 105 1 m1 , and the absolute values of the
Seebeck coefficients are somewhat larger below ¼ 5:1 105 1 m1 . As seen in Fig. 6(b), the thermal conductivities
appear to decrease slightly with increasing electrical conductivity.
As mentioned in section 3.1.1, the electrical conductivity
is strongly affected by the decomposition of the sample in air,
unlike the Seebeck coefficient and thermal conductivity.
Thus, the effect of the difference in the air stability of each
sample is more or less reflected in the results shown in
Figs. 6(a) and (b).
It should be noted that the power factor and figure of merit
generally increased lineally with increasing electrical conductivity, as shown in Figs. 6(c) and (d). This suggests that
the performance of GICs as thermoelectric materials is
governed mainly by their electrical conductivity.
3.1.3 Cs-GICs prepared from PGS graphite sheet
The thermoelectric properties of Cs-GICs prepared from
PGS graphite sheets, instead of Grafoil, are listed in Table 3.
PGS graphite sheets28) have higher electrical and thermal
conductivity than Grafoil.
For the Cs-GICs prepared from the PGS graphite sheet, the
Seebeck coefficients were about 2/3 and the thermal
conductivities were a few times higher than those of the
Cs-GICs prepared from Grafoil. The electrical conductivities
were, however, about 10 times higher than those of the CsGICs prepared from Grafoil, and thus the figures of merit
were much higher than those of the Cs-GICs prepared from
Grafoil.
It should be noted that the power factor of the Cs-GICs
prepared from the PGS graphite sheet reached 6:5 103 Wm1 K2 , which is comparable to those of promising
candidate thermoelectric materials.
It is considered that the most effective way of improving
the figure of merit and power factor of Cs-GICs, is to increase
the electrical conductivity of the sample, because it is
apparently difficult to control the Seebeck coefficient.
3.2
of Cs-GICs, it is necessary to increase the Seebeck
coefficients and decrease the thermal conductivities.
3.1.2 Other GICs prepared from Grafoil
In Fig. 6, the thermal conductivities, Seebeck coefficients,
and figures of merit are plotted against the electrical
conductivity, for Cs-ethylene-GIC, Cs-benzene-GIC, Cs-
Dependence of the thermoelectric properties of CsGICs on the intercalate concentration
Figure 7 shows the dependence of the electrical conductivity, Seebeck coefficient, and thermal conductivity on the
Cs concentration in Cs-ethylene-GICs prepared from Grafoil.
In this figure, an increase in x implies a decrease in the Cs
concentration.
As shown in Fig. 7(a), the electrical conductivity tends to
Table 3 Electrical conductivity (), Seebeek coefficient (), thermal conductivity (), power factor (P), and figure of merit (Z) for CsGICs prepared from PGS graphite sheet.
Composition
/1 m1
5
/mV K1
/Wm1 K1
P/Wm1 K2
Z/K1
500
6
9:1 10
1:8 108
300
6:2 103
2:1 105
3
(PGS graphite sheet)
4:3 10
CsC24
1:3 107
CsC24
6
6:8 10
22
270
3:2 10
1:2 105
CsC24 (C2 H4 )1:4
1:2 107
23
240
6:5 103
2:7 105
4:6
22
1462
R. Matsumoto, N. Akuzawa and Y. Takahashi
Table 4 Electrical conductivity (), Seebeek coefficient (), thermal
conductivity (), and figure of merit (Z) for FeCl3 -GICs prepared from
Grafoil.
(a)
5
/1 m1
/mV K1
/Wm1 K1
Z/K1
C173 FeCl3
1:7 105
23
200
4:4 107
C8:8 FeCl3
5
31
69
1:0 105
Composition
-1
σ / 10 Ω m
-1
15
10
7:3 10
5
Seebeck coefficients (Fig. 7(b)) were almost constant with
increasing Cs concentration, except in one of the samples,
CsC48 (C2 H4 )y .
On the other hand, the thermal conductivity clearly
correlated with the Cs concentration, as shown in Fig. 7(c).
The increase in the Cs concentration causes a decrease in the
thermal conductivity in a manner similar to that described in
section 3.1.1: The intercalation process causes an increase in
the number of lattice defects, as well as an increase in the
carrier density. Therefore, with the increase in the Cs
concentration, the increase in electrical conductivity and
the decrease in thermal conductivity occurred simultaneously
in the Cs-GICs. In this case, the Wiedemann-Franz rule,
which represents one of the difficulties in finding an excellent
thermoelectric material, apparently does not apply.
The figure of merit showed a similar dependence on the
electrical conductivity, as shown in Fig. 7(d). This indicates
that the performance of Cs-GICs as thermoelectric materials
is principally regulated by their electrical conductivity, as has
been pointed out.
The same trend as that in the Cs-GICs was also observed in
the FeCl3 -GICs prepared from Grafoil; the experimental
results are given in Table 4. Since FeCl3 -GICs are p-type
materials, the electrons of the graphite planes are transferred
to the FeCl3 intercalates. With the increase in the FeCl3
concentration, the electrical conductivity increased and the
thermal conductivity decreased, resulting the increase of the
figure of merit.
(b)
α / µVK
-1
-20
-30
-40
(c)
100
-1
κ / Wm K
-1
120
80
60
40
(d)
-5
Z / 10 K
-1
2.0
1.5
1.0
0.5
4.
10
20
x
30
40
50
in CsCx(C2H4)y
Fig. 7 Cs concentration dependence of (a) electrical conductivity ,
(b) Seebeck coefficient , (c) thermal conductivity , and (d) figure of
merit Z for Cs-GICs prepared from Grafoil.
increase slightly with increasing the Cs concentration. We
have reported previously that the electrical conductivities of
K-GICs prepared from Grafoil increase with increasing K
concentration.13)
Some reports2–4) have indicated that the Seebeck coefficients of GICs depend slightly on their stage structure, that
is, on the intercalate concentration. For example, in the case
of SbCl5 -GIC at room temperature,2) the Seebeck coefficients
of the stage-2 and stage-3 compounds are almost equal
( 25 mV K1 ), and that of the stage-10 compound is
approximately 40 mV K1 . The stage structure dependence
of the Seebeck coefficient appears weak compared with that
of the electrical and thermal conductivity. In this study, the
Conclusion
The thermoelectric properties of Cs-GICs and other GICs
were investigated and their potential as thermoelectric
materials was discussed. At present, although the electrical
conductivities of Cs-GICs are sufficiently high compared
with those of the other candidates, their Seebeck coefficients
are rather small and thermal conductivities are high. As a
result, the figures of merit of these GICs are not comparable
to those of the other candidate thermoelectric materials,
although the power factors are comparable. In order to
improve the thermoelectric performance of GICs, it is
necessary to increase their Seebeck coefficients and decrease
their thermal conductivities. However, it appears difficult to
control the Seebeck coefficients of GICs.
The simultaneous increase in the carrier density and
decrease in the lattice thermal conductivity were caused by
the increase in the intercalate concentration. We consider this
to be a great advantage for improving the thermoelectric
performance of GICs. In particular, it is important to devise a
method to decrease the thermal conductivity by increasing
the intercalate concentration—this is one of the most
important subjects for further study.
Thermoelectric Properties of Cesium–Graphite Intercalation Compounds
REFERENCES
1) M. S. Dresselhaus and G. Dresselhaus: Adv. Phys. 30 (1981) 139–326.
2) M. Elizinga, D. T. MoRelli and C. Uher: Phys. Rev. B 26 (1982) 3312–
3319.
3) K. Sugihara: Phys. Rev. B 28 (1983) 2157–2165.
4) K. Kobayashi, H. Oshima, K. Sugihara and T. Tsuzuku: J. Phys. Soc.
JPN. 61 (1992) 596–601.
5) J.-P. Issi, J. Boxus, B. Poulaert, H. Mazurek and M. S. Dresselhous:
J. Phys. C 14 (1981) 307–310.
6) J.-P. Issi, J. Heremans and M. S. Dresselhaus: Phys. Rev. B 27 (1983)
1333–1347.
7) L. Piraux, J.-P. Issi, J. F. Maêché and E. McRae: Synth. MET. 30
(1989) 245–249.
8) J. Boxus, B. Poulaert, J.-P. Issi, H. Mazurek and M. S. Dresselhaous:
Solid. State. Commun. 38 (1981) 1117–1118.
9) H. Pilliere, Y. Takahashi, T. Yoneoka, T. Otosaka and N. Akuzawa:
Synth. MET. 59 (1993) 191–199.
10) R. Matsumoto, Y. Takahashi and N. Akuzawa: Mol. Cryst. Liq. Cryst.
340 (2000) 43–48.
11) R. Matsumoto, Y. Takahashi, K. Watanabe and N. Akuzawa: Synth
MET. 125 (2002) 142–146.
12) R. Matsumoto, N. Akuzawa and Y. Takahashi: TANSO 216 (2005) 34–
40.
13) N. Akuzawa, Y. Amari, T. Nakajima and Y. Takahashi: J. Mater. Res. 5
(1990) 2849–2853.
14) N. Akuzawa, S. Shimura, K. Oi, T. Terai and Y. Takahashi: Mater. Sci.
1463
Forum 91–93 (1992) 459–464.
15) N. Akuzawa, S. Kondow, Y. Kaburagi, Y. Hishiyama and Y.
Takahashi: Carbon 31 (1993) 963–968.
16) R. Matsumoto, Y. Takahashi, T. Yoshioka, M. Osaki, I. Kumamoto and
N. Akuzawa: TANSO 184 (1998) 194–198.
17) G. Merle, I. Rashkov, C. Mai and et J. Gole: Mater. Sci. Eng. 31 (1977)
39–44.
18) H. Kurata, T. Kobayashi and N. Uyeda: Synth. MET. 20 (1987) 109–
118.
19) M. Ohira, T. Terai and Y. Takahashi: TANSO 125 (1986) 45–52.
20) R. C. Croft: Aust. J. Chem. 9 (1956) 184–193.
21) R. Matsumoto, Y. Takahashi and N. Akuzawa: TANSO 198 (2001)
129–133.
22) D. G. Onn, G. M. T. Foley and J. E. Fischer: Phys. Rev. B 19 (1979)
6474–6483.
23) E. Mcrae, D. Billaud, J. F. Maêché and A. Hérold: Phys. B 99 (1980)
489–493.
24) H. Suematsu, K. Higuchi and S. Tanuma: J. Phys. Soc. JPN. 48 (1980)
1541–1549.
25) C. Uher: Cryogenics 20 (1980) 445–447.
26) N. Akuzawa, M. Murakami, M. Nakano, Y. Soneda, R. Matsumoto and
Y. Takahashi: Mol. Cryst. Liq. Cryst. 340 (2000) 59–84.
27) C. Uher: Phys. Rev. B 25 (1982) 4167–4172.
28) M. Murakami, N. Nishiki, K. Nakamura, J. Ehara, H. Okada, T.
Kouzaki, K. Watanabe, T. Hoshi and S. Yoshimura: Carbon 30 (1992)
255–262.