Synthesis and single-crystal growth of Ca2-xSrxRuO4

Article abstract of DOI:10.1006/jssc.2000.8953

For the study of the quasi-two-dimensional Mott transition system Ca_2-xSr_xRuO₄, we have succeeded in synthesizing polycrystalline samples and also growing single crystals by a floatingzone method. Details of the preparations for the entire solution range are described. The structural, transport, and magnetic properties of both polycrystalline and single-crystal samples are fully in agreement.

Full text of DOI:10.1006/jssc.2000.8953

Journal of Solid State Chemistry 156, 26}31 (2001)
doi:10.1006/jssc.2000.8953, available online at http://www.idealibrary.com on
Synthesis and Single-Crystal Growth of Ca Sr RuO
2
؊x
x
4
Satoru Nakatsuji* and Yoshiteru Maeno*ꢀ-
*
Department of Physics, Kyoto University, Kyoto 606-8502, Japan; and -CREST, Japan Science and Technology Corporation, Japan
Received May 23, 2000; in revised form August 8, 2000; accepted September 5, 2000; published online December 21, 2000
In this paper, we will present details of the synthesis of
polycrystalline samples and also the single-crystal growth of
For the study of the quasi-two-dimensional Mott transition
system Ca2؊xSr RuO , we have succeeded in synthesizing poly-
crystalline samples and also growing single crystals by a 6oating-
Ca
Sr RuO in the entire range of x. Especially for the
x
4
ꢁ
\V
V
ꢂ
preparation of the polycrystalline samples, we discuss the
systematic change of reaction conditions, which require
zone method. Details of the preparations for the entire solution
range are described. The structural, transport, and magnetic careful attention to tune the heating temperature and atmo-
properties of both polycrystalline and single-crystal samples are sphere for each x. The results of structural analysis by X-ray
fully in agreement. ( 2001 Academic Press
di!raction for both polycrystalline and single-crystalline
samples agree well with each other. In addition, the consist-
ency between polycrystalline and single-crystalline samples
in transport and magnetic properties demonstrates that the
Key Words: Ca2؊xSr
x
RuO
4
;
Mott transition; Sr
2
4
RuO ;
Ca RuO ; spin-triplet superconductor; Mott insulator; p-wave
2
4
superconductivity; ruthenate; 6oating zone method; quasi-two-
dimensional system.
intrinsic properties of Ca
Sr RuO can be well studied
ꢁ
\V
V
ꢂ
using polycrystalline samples.
INTRODUCTION
PREPARATION OF POLYCRYSTALS
AND SINGLE CRYSTALS
First synthesized in 1959 (1), Sr RuO was known
ꢁ
ꢂ
simply as a paramagnetic conductor, until its superconduc-
The conditions for the synthesis of polycrystalline
Sr RuO allow only a narrow window for the appro-
tivity was revealed in 1994 (2). Quite interestingly, it Ca
shares the same K NiF -type structure with one of the priate heating temperature and atmosphere that is heavily
ꢁ
\V
V
ꢂ
ꢁ
ꢂ
best-studied high transition temperature (¹ ) superconduc- dependent on x. We synthesized polycrystalline Ca
#
ꢁ\V
tors La
Sr CuO . In contrast with the spin-singlet Sr RuO from CaCO (purity 4N), SrCO (4N), and
ꢁ\V
V
ꢂ
V
ꢂ
ꢃ
ꢃ
d-wave paring in high-¹ cuprates, it has recently been RuO (3N) mixed in the stoichiometric proportion. The mix-
demonstrated that Sr RuO is a spin-triplet super- tures were pressed into 10-mm-diameter pellets under about
#
ꢁ
ꢁ
ꢂ
conductor (3).
2000 kgf/cmꢁ. A tubular furnace, which allows controlled
The discovery of the superconductivity in Sr RuO has gas #ow, was used for sintering. We explored heating tem-
ꢁ
ꢂ
stimulated searches for superconductivity in related new peratures between 1100 and 15703C, and O partial pres-
ꢁ
ruthenates (4, 5). The Ca analogue, Ca RuO , has recently sures between 0 and 0.01 bar in Ar#O atmosphere with
ꢁ
ꢂ
ꢁ
been synthesized (5, 6). In contrast with the superconductive the total pressure of 1 bar. In order to minimize the evapor-
Sr RuO , Ca RuO is a Mott insulator (5, 7, 8). Naively, it is ation of RuO at high temperatures, we inserted pellets
ꢁ
ꢂ
ꢁ
ꢂ
ꢁ
expected that 4d electrons in the highly extended orbitals directly into a hot furnace, and after heating, pulled them
should have only a weaker correlation in comparison with out of the furnace hot and quenched them in air. Heating
the 3d case. Yet, the existence of this Mott insulator con- times of 12 and 16 h were employed, but they yielded no
"
rms that strong correlation plays a signi"cant role in this signi"cant di!erence.
4
d-electron system, and probably in the mechanism of the Figures 1a and 1b show the dependence of the major
superconductivity in Sr RuO . To explore the correlation phase in a pellet on the reaction temperature after an initial
ꢁ
ꢂ
\V
e!ects in the ruthenates, Ca
Sr RuO is an ideal system heating in a 99% Ar#1% O atmosphere and a 99.9%
ꢁ
V
ꢂ
ꢁ
that should bridge a new type of the Mott-transition route Ar#0.1% O atmosphere, respectively. The stability range
between the Mott insulator Ca RuO and the spin-triplet of Ca
superconductor Sr RuO (9). We have succeeded in syn- oxygen partial pressure. Furthermore, even in the same
thesizing polycrystalline samples and in growing single crys- atmosphere, we see the signi"cant and systematic variations
ꢁ
Sr RuO (214 phase) changes remarkably with
ꢁ
ꢂ
ꢁ\V
V
ꢂ
ꢁ
ꢂ
tals of this solid solution system.
of the stable temperature range with x for the 214 phase.
2
6
0
022-4596/01 $35.00
Copyright ( 2001 by Academic Press
All rights of reproduction in any form reserved.

SYNTHESIS AND SINGLE-CRYSTAL GROWTH OF Ca
Sr RuO
27
ꢁ\V
V
ꢂ
perature and too strongly a reducing atmosphere prevents
the formation of the 214 phase, but encourages the reduc-
tion of RuO into Ru metal. Once Ru metal appears in a
ꢁ
pellet, we have never succeeded in obtaining the 214 phase
by repeating the same heating process. This condition deter-
mines the upper limit of the annealing temperature. Second,
although the stoichiometric Sr RuO and Ca RuO can be
ꢁ
ꢂ
ꢁ
ꢂ
obtained by annealing at relatively low temperature near
3003C, the reaction temperature for the solution system
1
Ca
Sr RuO increases as the Sr content x approaches 1.
ꢁ
\V
V
ꢂ
Instead the Ca
Sr Ru O (327) phase becomes stable up
ꢃ\V
V
ꢁ ꢄ
to a higher temperature, which may be correlated with the
additional di!usion process to mix Ca and Sr in the solution
system, in contrast with the endmembers (x"0 and 2.0).
On the basis of a series of phase diagrams including
Figs. 1 and 2, we have established the x dependence of the
optimized reaction temperature and atmosphere, as sum-
marized in Table 1. Grinding and sintering were repeated
several times to obtain more homogeneous samples with
better crystallinity. As mentioned above, we introduced
pellets into a hot furnace at an appropriate temperature,
and after heating, quenched them to room temperature.
This procedure also helps avoid the inclusion of
Ca
Sr RuO (113) phase and/or 327 phase. However, to
ꢅ\V
V
ꢃ
obtain well-sintered samples, we noticed that it is better to
FIG. 1. Major phases in polycrystalline pellets as a function of the
heating temperature and the Sr concentration x under the atmosphere of
"rst synthesize a mixed phase of 214 with a little amount of
(
a) 99% Ar#1% O and (b) 99.9% Ar#0.1% O . The circle, triangle, and 327 as a precursor, by reacting the raw material at a temper-
ꢁ ꢁ
cross indicate that the fraction of the 214 phase is approximately more than
0%, around 50%, and less than 30%, respectively. 327, 113, and Ru
ature slightly lower (by about 503C) than the stable one
for 214. This preceding treatment yielded better sintered
samples.
Contamination of samples was minimized. In particular,
Ca in a sample readily reacts with an Al O crucible and
8
indicate that the dominant phase is Ca
Ru metal, respectively. The solid and broken lines are guides to the eye.
Sr Ru O , Ca
Sr RuO , and
ꢃ\V
V
ꢁ
ꢄ
ꢅ\V
V
ꢃ
ꢁ
ꢃ
Generally, the atmosphere becomes more reducing at higher
temperatures as well as with decreasing oxygen partial pres-
sure. Taking into account the physical phase diagram of
magnetic and electrical properties described below, Figs. 1a
and 1b suggest that the metallic 214 phase is more stable
under relatively reducing conditions, re#ecting the expected
stronger covalency of the metallic Ru}O bond.
Although single-phase samples were obtained using both
atmospheres, we sometimes had di$culty in obtaining
well-sintered samples for the region 0.24x41.0. This is
most likely due to the heavy evaporation of RuO under
ꢁ
rather high heating temperatures as in Figs. 1a and 1b.
In order to lower the temperature for the synthesis, we
carefully investigated the atmosphere and temperature
dependence of the major phase in pellets with x"0.2, 0.4,
0
.5, 0.8, and 1.0. Figure 2 presents the result for the case with
FIG. 2. Major phases in polycrystalline pellets as a function of the
x"1.0.
reaction temperature and the oxygen partial pressure P(O ) in the mixture
The systematic change of the reaction temperature in
ꢁ
of Ar and O with the total pressure of 1bar. The nominal composition is
Figs. 1a, 1b, and 2 indicates the following two competing
aspects. First, while the metallic 214 phase may favor
ꢁ
CaSrRuO (x"1.0). The circle, triangle, and cross indicate that the frac-
ꢂ
tion of the 214 phase is approximately more than 80%, around 50%, and
a rather reducing condition as noted above, too high a tem- less than 30%, respectively. The solid lines are guides to the eye.

2
8
NAKATSUJI AND MAENO
TABLE 1
STRUCTURAL ANALYSIS
Optimized Reaction Temperature and Atmosphere for
Polycrystalline Ca2؊xSr RuO , with Various Values of x
x
4
The crystal structures at room temperature of both poly-
crystalline and single-crystalline samples were studied by
powder X-ray (CuKa) di!raction measurements. In Fig. 3,
x
Temperature(3C)
Atmosphere
we summarize the results of polycrystalline Ca
synthesized under the condition listed in Table 1. All of the
X-ray spectra are basically consistent with K NiF -type
structure and well indexed with tetragonal symmetry in the
region 0.24x42, and with orthorhombic symmetry in the
region 04x(0.2. Each sample is revealed to be a single
phase, except for a small inclusion of CaO probably due to
Sr RuO
0
0
0
0
0
0
0
0
0
1
1
1
1
1
2
.00ꢆ
.05
.10
.15
.20
.25
.40
.50
.80
.00
.30
.50
.80
.90
.00
1360}1380 (1360)
1320}1340 (1340)
1360}1380 (1380)
1360 (1360)
1380}1395 (1380)
1400}1410 (1410)
1520 (1520)
1520 (1520)
1500}1505 (1505)
1503 (1503)
1380}1400 (1400)
1300}1400 (1400)
1300}1400 (1400)
1300}1400 (1400)
1100}1400 (1400)
Ar#Oꢁ 1.0%
ꢁ
\V
V
ꢂ
Ar#O 1.0%
ꢁ
Ar#O 1.0%
ꢁ
ꢁ
ꢂ
Ar#O 0.3%
ꢁ
Ar#O 0.3%
ꢁ
Ar#O 0.3%
ꢁ
Ar#O 1.0%
ꢁ
Ar#O 1.0%
ꢁ
Ar#O 0.3%
the evaporation of RuO . The results for the single-crystal-
ꢁ
ꢁ
Ar#O 0.3%
line samples show spectra essentially identical to those for
ꢁ
Ar#O 0.1%
ꢁ
the polycrystalline samples. Moreover, for the case of the
single crystals, no impurity phase was detected for the entire
range of Sr contents.
Ar#O 0.1%
ꢁ
Ar#O 0.1%
ꢁ
Ar#O 0.1%
ꢁ
Ar#Oꢁ 0.1%
In order to clarify the systematic variation of the cell
parameters, we considered here a unit cell with I4/mmm like
that of Sr RuO for the tetragonal lattice, while in the
Note. For the "nal sintering, the samples were heated at the temper-
atures indicated in parentheses.
ꢁ
ꢂ
region 04x(0.2, we show a"a /(2 and b"b /(2,
ꢀ
ꢀ
where a and b are the in-plane cell parameters for the
ꢀ
ꢀ
orthorhombic unit cell as in Ref. (11). The cell parameters
forms an eutectic Ca}Al}O compound with a melting point based on this assignment are plotted in Fig. 4. The solid and
lower than 13003C. In order to avoid this reaction as well as open squares correspond to the single-crystal and polycrys-
the contamination by di!usion of impurities, we placed two talline samples, respectively. Parameters of the single crys-
or three pellets of the same composition on top of a pellet of tals were displayed using the Sr content of the starting
Sr RuO in an Al O crucible, and used only the top of the materials. They agree well with those of the polycrystalline
ꢁ
ꢂ
ꢁ ꢃ
two or the middle of the three for measurements.
samples, which con"rms the homogeneous mixing of Ca
We have recently succeeded in growing single crystals of and Sr. The volume increases almost linearly with x re#ect-
Ca
Sr RuO in the whole region of x by a #oating zone ing the larger size of Srꢁ> than of Caꢁ>. In the region
ꢁ
\V
V
ꢂ
method with Ru self-#ux. As in the polycrystalline sample 04x40.2, however, the cell parameters a and b split and
preparation, we "rst made a pellet from a mixture of decrease with x, while c increases rapidly. These changes
CaCO , SrCO , and RuO . Here, the volatility of RuO is
also a serious problem. To compensate for this loss, we
prepared a Ru-rich starting material with a composition of
ꢃ
ꢃ
ꢁ
ꢁ
Ca
Sr Ru
O . After presynthesizing the pellet, we
ꢁ
\V
V
ꢅꢇꢅꢆ ꢂ
ground it again and pressed it under a pressure of
1
500 kgf/cmꢁ into a rod of 6}8 cm in length and about 5 mm
in diameter. The rod was sintered at 10003C for 12 h in air.
Contamination was avoided, especially during the heating.
We took the same precautions as for the polycrystalline
samples described above.
In the #oating zone method, the focused infrared radi-
ation melts the rod without any physical contact with a cru-
cible, so that no contamination is expected during the
crystal growth. We used a commercial infrared image fur-
nace (NEC Machinery, Model SC-K15HD) and performed
the crystal growth typically under a 90% Ar#10% O gas
ꢁ
mixture with a total pressure of 10 bar and a growth speed
of 20 mm/h. The average size of single crystals obtained so
far is about 5;3;3 mmꢃ, still substantially smaller than
those of Sr RuO . Details of the single-crystal growth of the
FIG. 3. X-ray di!raction spectra at room temperature of polycrystal-
line Ca
Sr RuO for various Sr concentrations x. The asterisk indicates
ꢁ
ꢂ
ꢁ\V
V
ꢂ
superconductor Sr RuO is presented elsewhere (10).
a small inclusion of CaO.
ꢁ
ꢂ

SYNTHESIS AND SINGLE-CRYSTAL GROWTH OF Ca
Sr RuO
29
ꢁ\V
V
ꢂ
neutron di!raction measurements at room temperature
have indeed con"rmed that these simple expectations are
correct in this system (14).
PHYSICAL PHASE DIAGRAM
Here, we brie#y review the phase diagram of the physical
properties of Ca
lowing three regions:
Sr RuO (9), which consists of the fol-
ꢁ
\V
V
ꢂ
I. 04x(0.2. Insulating phase with canted antifer-
romagnetic ordering appears at low temperatures, in con-
trast with paramagnetic metal phase at high temperatures.
Therefore, a metal/nonmetal transition occurs by varying
temperature. Ca RuO (x"0) is the Mott insulator up to
ꢁ
ꢂ
3
50 K (8, 14).
II. 0.24x40.5. Antiferromagnetically correlated metal-
lic region (we call this &&magnetic metallic region'') appears at
low temperatures.
III. 0.5(x42.0. Paramagnetic metallic phase spreads in
this wide region. Itinerant magnetism changes continuously
with x from Curie}Weiss like paramagnetism to Pauli para-
magnetism at x"2.0. Superconductivity occurs at x"2.0
below ¹ "1.5 K.
#
ELECTRICAL PROPERTIES OF POLYCRYSTALLINE
SAMPLES
We measured the resistivity of polycrystalline samples by
a standard four-probe method. The resistivity curves o(¹)
for the region I measured on cooling are shown in Fig. 5.
For x"0, Ca RuO shows insulating behavior as a Mott
ꢁ
ꢂ
insulator. However, a small Sr substitution for Ca dramati-
cally changes the temperature dependence of the resistivity.
FIG. 4. Cell parameters a, b, c and unit cell volume < of Caꢁ Sr
\V
V
RuO against the Sr concentration x of their starting materials. The open
ꢂ
and solid squares respectively represent those of polycrystalline and single- For x"0.05, 0.10 and 0.15, o(¹) is almost constant at high
crystalline samples.
temperatures like the metallic Sr RuO . Below some char-
ꢁ
ꢂ
acteristic temperature ¹+ , however, the temperature de-
ꢉ,+
pendence suddenly changes into an insulating one. This
indicate that a structural transition between an orthorhom- change indicates that a metal/nonmetal (M/NM) transition
bic phase and a tetragonal phase occurs at around x"0.2 occurs on cooling at this point. The inset of Fig. 5 illu-
and room temperature.
strates thermal hysteresis observed in o(¹) for x "0.15,
The structures of the endmembers have previously been which con"rms that the transition is of the "rst order. The
determined by neutron di!raction measurements (11, 12). transition temperature was determined from
Ca RuO has a highly distorted K NiF structure with the intersection of the extrapolations of low- and
¹
+ꢉ,+
ꢁ
ꢂ
ꢁ
ꢂ
both rotation and tilt of the RuO octahedra (11). On the high-temperature logo}¹ curves measured on cooling.
ꢈ
other hand, Sr RuO exhibits neither rotational nor tilt
¹
+ꢉ,+
decreases rapidly with x, and "nally becomes zero at
ꢁ
ꢂ
distortion (12), although phonon dispersion measurements around x"0.2. All of the observations here are consistent
have indicated that it is close to a rotational instability (13). with those for the single-crystalline samples (9).
Given these results, it is naturally expected that Ca substitu-
Figure 6 shows the resistivity data for polycrystalline
tion in Sr RuO "rst stabilizes the rotation of RuO oc- samples in regions II and III. For comparison, we also show
ꢁ
ꢂ
ꢈ
tahedra in the region 2.0'x50.2. In fact, the occurrence the data for x"0.15. In contrast with the insulating
of the maximum in the c-axis parameter at x"1.5 is sugges- Ca RuO , the o(¹) curves in regions II and III have small
ꢁ
ꢂ
tive of the onset of the static rotational distortion. The large absolute values and little temperature dependence. Even for
reduction in the c parameter across the structural transition a good metal such as Sr RuO (the in-plane resistivity
ꢁ
ꢂ
near x"0.2 marks the onset of additional tilt. Recent o +120 k)cm and the out-of-plane resistivity o +
?@
A

3
0
NAKATSUJI AND MAENO
FIG. 6. Temperature dependence of the resistivity o(¹) on cooling for
0.154x42.0 of polycrystalline Caꢁ\V
Sr RuO .
V
ꢂ
around 100 K for x"2.00 in Fig. 6 is most probably re#ect-
ing the maximum in o (2). Given these e!ects, the observed
A
temperature independence or weekly semiconducting be-
havior in polycrystalline samples is considered consistent
with the transport properties observed in the single crystals.
Finally, we note the fact that the magnetic properties of
both the polycrystalline and single-crystalline samples also
agree qualitatively and even quantitatively well with each
other. The absolute values of the susceptibility and its tem-
perature dependence show identical behavior to each other in
all the regions, as shown in Fig. 4 of Ref. (9) and Fig. 7 of
Ref. (15). This consistency in the measured physical properties
between polycrystalline and single-crystalline samples dem-
FIG. 5. Temperature dependence of the resistivity o(¹) on cooling for
the region I (04x40.2) of polycrystalline Caꢁ Sr RuO . ¹
is the
\V
V
ꢂ
+ꢉ,+
metal/nonmetal transition temperature. The inset shows the hysteresis in
o(¹) for x"0.15.
3
0 m)cm at 300 K), the polycrystalline sample shows al- onstrates that we are able to deduce the intrinsic properties of
most temperature independence and at high temperatures this system by investigating the polycrystalline samples.
weakly semiconducting resistivity. This di!erence between
the data for single-crystal and polycrystalline samples is
attributable to poor conduction of the grain boundaries (2).
Taking this into account, the intrinsic behavior of the region
CONCLUSION
We describe the syntheses of polycrystalline samples and
0
.24x42.0 may well be metallic at all the temperatures. also of single crystals of the quasi-two-dimensional Mott
To con"rm this, we measured o?
of single-crystal samples transition system Ca
Sr RuO for the entire region of x.
@
ꢁ\V
V
ꢂ
with x"0.2, 0.5, 0.9, 1.5, and 2.0. Indeed, all of them show We report the details of their preparations. Especially for
metallic temperature dependence with positive do/d¹ at all the polycrystalline samples, we discuss the systematic
the temperatures (15). On the other hand, o in this system is change of the reaction conditions, which requires careful
A
almost temperature independent and sometimes shows non- tuning of the heating temperature and atmosphere for each
metallic behavior, re#ecting the quasi-two-dimensional concentration, x. The results of a structural analysis on both
structure (15). Therefore, in addition to the resistance of its polycrystalline and single-crystalline samples show excellent
grain boundary material such as CaO or SrO, as discussed agreement. Moreover, measurements of transport and mag-
in Ref. (16), this highly anisotropic transport property netic properties of polycrystals and single crystals demon-
should also be responsible for the weakly semiconductive strate that polycrystalline samples are suitable for
behavior of polycrystalline samples. In fact, a broad peak at investigations of the intrinsic properties of this system.

SYNTHESIS AND SINGLE-CRYSTAL GROWTH OF Ca
Sr RuO
31
ꢁ\V
V
ꢂ
ACKNOWLEDGMENTS
6. G. Cao, S. McCall, M. Shepard, J. E. Crow, and R. P. Guertin, Phys.
Rev. B 56, R2916 (1997).
7
8
. H. Fukazawa, S. Nakatsuji, and Y. Maeno, Physica B 281/282, 613 (2000).
. C. S. Alexander, G. Cao, V. Dobrosavljevic, S. McCall, J. E. Crow,
E. Lochner, and R. P. Guertin, Phys. Rev. B 60, R8422 (1999).
. S. Nakatsuji and Y. Maeno, Phys. Rev. ¸ett. 84, 2666 (2000).
The authors acknowledge T. Ishiguro for his support. They thank T.
Ando, H. Fukazawa, and M. Minakata for their technical support and
valuable discussions. This work has been supported in part by a Grant-in-
Aid for Scienti"c Research on Priority Areas from the Ministry of Educa-
tion, Science, Sports and Culture of Japan. One of the authors (S.N.) has
been supported by JSPS Research Fellowships for Young Scientists.
9
1
1
1
0. Z. Q. Mao, Y. Maeno, and H. Fukazawa, Matt. Res. Bull. 35 (2000), in
press.
1. M. Braden, G. Andre
H , S. Nakatsuji, and Y. Maeno, Phys. Rev. B 58, 847
(1998).
REFERENCES
2. M. Braden, A. H. Moudden, S. Nishizaki, Y. Maeno, and T. Fujita,
Physica C 273, 248 (1997).
1
2
. J. J. Randall and R. Ward, J. Am. Ceram. Soc. 81, 2629 (1959).
. Y. Maeno, H. Hashimoto, K. Yoshida, S. Nishizaki, T. Fujita, J. G.
Bednorz, and F. Lichtenberg, Nature 372, 532 (1994).
. K. Ishida, H. Mukuda, Y. Kitaoka, K. Asayama, Z. Q. Mao, Y. Mori,
and Y. Maeno, Nature 396, 658 (1998).
13. M. Braden, W. Reichardt, S. Nishizaki, Y. Mori, and Y. Maeno, Phys.
Rev. B 57, 1236 (1998).
14. O. Friedt, M. Braden, S. Nakatsuji, and Y. Maeno, submitted for
publication.
15. S. Nakatsuji and Y. Maeno, Phys. Rev. B 62, 6458 (2000).
16. R. J. Cava, H. W. Zandbergen, J. J. Krajewski, W. F. Peck, Jr., B.
Batlogg, S. Carter, R. M. Fleming, O. Zhou, and L. W. Rupp, Jr.,
J. Solid State Chem. 116, 141 (1995).
3
4
5
. S. Ikeda, Y. Maeno, and T. Fujita, Phys. Rev. B 57, 978 (1998).
. S. Nakatsuji, S. Ikeda, and Y. Maeno, J. Phys. Soc. Jpn. 66, 1868
(1997).