Electrochemical and Chemical Behavior of Extra Molybdenum Atoms into the Chevrel Phase Host Network

Article abstract of DOI:10.1006/jssc.1999.8213

In this paper, we present a comparative electrochemical study of copper insertion into series of Mo₆Se_8-xS_x phases which are obtained through different synthesis processes. We show that α-, β-, and HT-forms can reversibly accommodate copper at room temperature. The variation of the integral copper insertion rate is a function of the sulfur content and the synthesis type of pseudobinaries. While the mechanism of copper intercalation is quite similar for the different Mo₆Se_8-xS_x for a given x, there is a drastic difference for α- and β-Mo₆S₈ compound. This behavior is explained on the basis of structural considerations. If the extra Mo atoms, present in the lattice channels of the β form and HT-Mo₆Se_8-xS_x, are not displaced by this copper intercalation, they are shown to be mobile through cationic exchange reactions and chemical deinsertion at higher temperature.

Full text of DOI:10.1006/jssc.1999.8213

Journal of Solid State Chemistry 147, 199}210 (1999)
Article ID jssc.1999.8213, available online at http://www.idealibrary.com on
Electrochemical and Chemical Behavior of Extra Molybdenum
Atoms into the Chevrel Phase Host Network
Z. Kaidi,* S. Belin,- C. Boulanger,* J.-M. Lecuire,* M. Sergent,- and R. Chevrel-ꢀꢁ
*Laboratoire d+Electrochimie des Mate&riaux, UMR CNRS 7555, Universite& de Metz, Ile du Saulcy, 57045 Metz Cedex 01, France; and
-Laboratoire de Chimie du Solide et Inorganique Mole&culaire, UMR CNRS 6511, Campus de Beaulieu, Universite& de Rennes 1, 35042 Rennes Cedex, France
E-mail: chevrel@univ-rennes1.fr
Received November 25, 1998; accepted February 5, 1999
electrochemical topotactic redox reactions via electron/ion
transfer process (3}8):
In this paper, we present a comparative electrochemical study
of copper insertion into series of Mo6Se8؊xSx phases which are
obtained through di4erent synthesis processes. We show that a-,
b-, and HT-forms can reversibly accommodate copper at room
temperature. The variation of the integral copper insertion rate is
a function of the sulfur content and the synthesis type of
pseudobinaries. While the mechanism of copper intercalation is
\
L>
M Mo X !xne 0 Mo X #xM
.
V
ꢂ ꢃ
ꢂ ꢃ
This type of solid state reaction provides a way to obtain
the a-Mo Se
S
mixed binaries series for instance by
ꢂ
ꢃ\V V
quite similar for the di4erent Mo6Se8؊xSx for a given x, there is oxidation of ternary phase in an aqueous hydrochloric acid
a drastic di4erence for a- and b-Mo6S8 compound. This behavior
is explained on the basis of structural considerations. If the extra
Mo atoms, present in the lattice channels of the b form and HT-
Mo6Se8؊xSx, are not displaced by this copper intercalation, they
are shown to be mobile through cationic exchange reactions and
chemical deinsertion at higher temperature. ( 1999 Academic Press
Key Words: Chevrel phases; copper; electrochemistry; inter-
calation; cationic exchange.
solution (3). These binaries are then transformed into a new
series of phases by annealing, which are called b. It can be
noted that the usual high temperature ceramic synthesis
only gives Mo Se
phases. Recently, it has been shown that these b-
S
binaries for x45 so-called HT
ꢂ
ꢃ\V V
Mo Se S and HT-Mo Se S binaries correspond to
ꢂ
ꢃ\V V
ꢂ
ꢃ\V V
pseudo-ternaries with Mo guest atoms located in the chan-
nels of the host network (9}11).
Two main ideas emerged to study the mobility of these
molybdenum atoms (Mo(2)) in the host. First, we searched
some process to remove these extra Mo(2) atoms from the
channels. Second, we studied the accommodation capacity
of the new host network with extra Mo(2) atoms through
the overinsertion of cations. The Chevrel phases can be
characterized by the valency electron concentration (VEC),
i.e., the number of valence electrons on the metal}metal
I. INTRODUCTION
The Chevrel phases M Mo X (M"metal ion; X"S,
V
ꢂ ꢃ
Se, Te) present interesting physical properties (superconduc-
tivity and high critical magnetic "elds, coexistence of super-
conductivity and magnetism) (1, 2). These properties have
been found to strongly depend upon the nature of chalcogen
X, cation M, and its content x. These compounds are
characterized by a framework matrix built-up from quasi-
bondings of the Mo cluster, which takes into account the
ꢂ
number of valence electron that a metal ion can transfer
to the Mo cluster. The host network Mo S has a
rigid Mo X molecular units. The three-dimensional pack-
ꢂ ꢃ
ꢂ ꢃ
VEC"20e/Mo , and Mo Se has a VEC"21e/Mo (12).
ing of these Mo X units leads to a system of channels
ꢂ
ꢂ ꢃ
ꢂ
ꢃ
ꢂ
ꢂ ꢃ
The host Mo X network can theoretically accommodate
along the rhombohedral axes which can accommodate the
metal ions until a VEC value as high as 24e/Mo . The
ternary metal atoms (M).
In addition to these physical properties, these phases
ꢂ
recently reported b- and HT-Mo Se S series (9}11) were
found to have a VEC inferior or equal to 22e/Mo , so we
can expect to overinsert some cations in such compounds as
ꢂ
ꢃ\V V
exhibit
a remarkable chemical reactivity. The high
ꢂ
mobility of the guest atom in the open framework type
structure allows to the synthesis of a series of new meta-
stable compounds by means of low-temperature chemical or
it was already shown in other ternary Chevrel phases (13).
Two &&soft'' routes are here investigated. An important
electrochemical study is carried on these materials since
electrochemistry is most appropriate to investigate the ac-
commodation capacity of a host network. Soft chemistry
ꢁTo whom correspondence should be addressed.
199
0022-4596/99 $30.00
Copyright ( 1999 by Academic Press
All rights of reproduction in any form reserved.

200
KAIDI ET AL.
studies (cationic exchange solid state reaction and chemical pounds, that is, a ceramic route from the starting elements
deintercalation process) are also used to characterize the or binaries (11).
mobility of the Mo(2) guest atoms.
We report in this context the results of our investigations.
II.3. Controls
It was of interest, therefore, to perform an electrochemical
study on topotactic properties of these di!erent binary and
pseudo-ternary systems in order to compare their reactivity.
Since the copper systems were the subject of a large amount
of theoretical and experimental works because the possible
use of Mo X and Cu Mo X as cathode material in room
X-ray di!raction data are obtained on an Inel C.P. 120
di!ractometer with a curved localization detector.
The stoichiometry was determined with an electron
microprobe microanalyzer (Cameca SX 50) after embedding
and polishing the microcrystals. The elements Cu, Mo, S,
and Se were analyzed in twelve di!erent microcrystals of all
the samples. The stoichiometry corresponds to the average
of these 12 values and is calculated with a number of
chalcogen (S#Se) per compound assigned and equal to 8
and with a con"dence rate to 95%.
ꢂ ꢃ
ꢄ
ꢂ ꢃ
temperature secondary lithium batteries (14}18), we have
chosen to study more speci"cally this Cu cation intercala-
tion. Finally, a new chemical Mo deintercalation process is
reported.
II. EXPERIMENTAL
III. ELECTROCHEMICAL DEINSERTION
The topotactic reaction is experimentally followed using
the curve i"f (E) with a classical three-electrode device in
a thermostated cell (603C) and under an inert atmosphere
(argon HP).
BEHAVIOR OF b-Mo6S8
Taking into account the action of anhydrous HCl gas (9),
the b-Mo S samples were electrochemically oxidized in
ꢂ ꢃ
1 M aqueous Na SO at ambient temperature. Figure 1
ꢄ
ꢅ
II.1. Electrodes and Electrolytes
shows the anodic scans. Attempts to remove the Mo(2)
located in the cavities (1) of the channels by an oxidation at
#700 mV have been unsuccessful. This anodic treatment
does not induce a change in the Mo composition of the
solid. The comparison between the voltamperograms ob-
tained with this pseudo-ternary and those corresponding to
the a-binary proves a similarity which can be attributed to
the destruction of the host lattice. Similar behavior in aque-
ous medium and at room temperature were observed in the
Crystallite bed electrodes are prepared according to the
already described procedure (7). The powdery samples are
applied to the surface of a glassy carbon rod (Societe Le
Carbone Lorraine) by means of a colloidal graphite slurry
(Societe Siceron KF). The electrode so realized was kept
immobile in the regularly stirred solution. All potentials are
controlled and expressed by reference to the aqueous KCl
saturated calomel electrode (SCE). The counter electrode
was a platinum wire.
The electrolytes were prepared from copper sulfate (ana-
lytical grade Prolabo).
A computer-driven Radiometer IMT1 electrochemical
interface and a PJT 24-1 potentiostat/galvanostat were used
for cyclic voltametric analysis.
ꢄ>
>
case of phases with large cations such as Pb , Ag located
in the same cavity (1). The di$culty to separate the charac-
teristic signal of electrochemical deinsertion and that of the
host lattice oxidation is in a relation with channel size.
Moreover, the structure of b-Mo S reveals that pairs of
ꢂ ꢃ
Mo atoms bonded together are present in the cavity (1).
ꢅ>
These (Mo)
framework at room temperature.
species may be not very mobile inside the
ꢄ
II.2. Samples
Di!erent process are used to synthesize a-Mo X , b-
ꢂ ꢃ
IV. ELECTROCHEMICAL INSERTION
Mo X , and HT-Mo X compounds. The former ones are
OF CADMIUM INTO b-Mo6S8
ꢂ ꢃ
ꢂ ꢃ
obtained by the reaction of stoichiometric mixtures of el-
emental Ni and Mo, and binary MoSe and/or MoS pow-
We tried to obtain the a-Mo S metastable binary from
ꢄ
ꢄ
ꢂ ꢃ
ders into sealed evacuated silica tubes brought to 12003C the b-Mo S pseudo-ternary by a Mo electrochemical dis-
ꢂ ꢃ
over 1 week. This leads to the formation of Ni Mo X
placement. We had shown, in early work, that the
W
ꢂ ꢃ
ternaries, which after hydrochloric acid leaching, yield the In Mo Se phase can be converted to Mo Se by an
ꢄ
ꢁꢆ ꢁꢇ
ꢁꢆ ꢁꢇ
ꢁꢆ ꢁꢇ
a-Mo X phases.
overintercalation of cadmium. The In Mo Se phase is
ꢂ ꢃ
ꢄ
b-Mo X compounds result from annealing under vac- not the only one for which the ternary element (M) can be
uum of the a-series at appropriated temperatures during electrochemically displaced. Tarascon et al. have studied the
ꢂ ꢃ
1 or 2 days (9, 10).
displacement of M"Ag (19), In, or Tl (5) by lithium in
The process used to obtain the HT-Mo X compounds is MMo S . A similar type of reaction occurs with small
the same as that described for the Ni Mo X ternary com- cations (17, 20).
ꢂ ꢃ ꢂ ꢃ
ꢂ ꢃ
W

BEHAVIOR OF EXTRA Mo ATOMS IN THE CHEVREL PHASE
201
FIG. 1. Voltamperogram for b-Mo S compounds in 1 M Na SO medium, supporting disk area, s"6.2 mmꢄ, temperature ¹"603C, scan rate
ꢂ ꢃ
ꢄ
ꢅ
v"10 mV min\ꢁ.
ꢄ>
The curves of the electrochemical insertion of cadmium sponding Cd intercalation peak in a-Mo S . Additional
are shown in Fig. 2. Several features can be observed. The voltametric scans reveal a transformation of the voltam-
ꢂ ꢃ
"rst potentiodynamic cycle exhibits only one signal perograms. An extra peak located at !450 mV shows up.
(E3"!550 mV) which is small as compared to the corre- This peak slightly grows on subsequent cycles while the "rst
FIG. 2. Voltamperogram for b-Mo S in 1 M CdSO medium, supporting disk area, s"6.2 mmꢄ, temperature ¹"603C, scan rate v"10 mV min\ꢁ.
ꢂ ꢃ
ꢅ

202
KAIDI ET AL.
one decreases. The voltamperogram features remain con- (350(E(250 mV) gives rise to higher stability phases
stant after 10 cycles. However, one should mention that the than those obtained in the range (250(E(0 mV). This
electrochemical signals have very low intensities. This ob- result may be explained in terms of site energy which deter-
servation suggests that only the surface sites are concerned mine the occupancy of these sites by copper. Dudley et al.
by the intercalation of cadmium and that the cation does have shown that the large cavities (site 1) are the lowest
ꢄ>
not progress through the channels. The migration of Cd
energy sites which begin to be populated.
ꢅ>
into the sites would be blocked by (Mo}Mo) species. The
The experimental curve suggests the successive occu-
ꢄ
available sites in MoMo S are certainly not large enough pancy of two types of available sites. The "rst signals be-
ꢂ ꢃ
ꢄ>
to accommodate large ions such as Cd . X-ray di!raction tween 350 and 250 mV correspond to the occupancy of the
measurements evidenced the absence of structural changes sites 1 while the second potential range indicates lower free
within the b-Mo S .
energy reactions thus the "lling of site 2. To substantiate
ꢂ ꢃ
The present results clearly indicate that no extra these suggestions, X-ray di!raction measurements were per-
molybdenum (2) atoms can be removed from the host formed. We have found (Table 1) that the lattice constants of
lattice by an electrochemical treatment (oxidation or dis- the ternary compounds obtained in the "rst potential range
placement by cadmium). The low mobility of Mo atoms into agree with phases having a copper content inferior to 1.5
the latter phase is uncompatible with a work at ambient while the second one leads to the formation of richer copper
temperature.
compounds. Moreover, it can be noted that there is a broad
We then extended the study to the copper intercalation correspondence between the energy level model proposed
for phases of a- and b-Mo Se S systems. The family by Dudley and the crystallographic data given by Yvon (22).
ꢂ
ꢃ\V V
of these di!erent compounds could constitute an ideal Our results are in good agreement with the results of struc-
example to illustrate the relationship between structure and tural investigations con"rming the above suggestions.
electrochemistry.
We now present the electrochemical and structural results
for the a-Mo Se S (x"7, 6, 4, 2, 1, 0). The electrochemi-
ꢂ
ꢃ\V V
V. ELECTROCHEMICAL INSERTION OF
cal behavior is reported in Figs. 3b}g, the X-ray results are
shown on Table 1. The given stoichiometry (Table 1) for
each binary is obtained for the "nal step, just before the
formation of metallic copper by imposing a constant poten-
COPPER INTO Mo6Se8؊xSx
V.1. a-Mo Se
S
ꢂ
ꢃ\V V
Mo S , x"8. The electrochemical copper insertion tial to the cathodes during 2 h.
ꢂ ꢃ
curve (Fig 3a) shows several signals. The voltamperogram
Mo Se S , x"7. The general shape of the i"f (E)
ꢁ ꢉ
curve (Fig. 3b) is similar to that exhibited by a a-Mo S
feature depends on the potential sweep rate. A slow sweep
ꢂ
\ꢁ
rate (10 mV min ) allows successive steps of the insertion
ꢂ ꢃ
electrode although the signals are not so well de"ned. Four
intercalation signals can be observed. The cathodic peak
potentials are slightly shifted in the potential scale
to be precisely distinguished. The curve exhibits one "rst
signal which corresponds to a superposition of two signals
ꢀ
(an important one E "330 mV and
a
small one
#
(E "300, 225, 160, and 100 mV), but only two global oxi-
ꢁ
E "300 mV) followed by two other perfectly de"ned
#
#
dation peaks are distinguished at anodic sweep even if we
ꢀꢀꢀ
ꢀꢀꢀꢀ
#
(E "190 mV, E "140 mV).
#
can suppose the existence of supplementary steps which
could certainly be seen by using a slower scan in potential.
A copper content of 2.65 for the limit phase is determined
leading logically to an increase of cell volume.
Our study has shown that the stoichiometries of the
successive phases are Cu Mo S , Cu Mo S ,
ꢁꢈꢄꢉ ꢂ ꢃ ꢁꢈꢅꢆ ꢂ ꢃ
Mo S . We can bring these inser-
Cu
Mo S , and Cu
ꢄꢈꢂꢊ
ꢂ ꢃ ꢋꢈꢆꢊ
ꢂ ꢃ
tion rates closer to those obtained by Schollhorn in an
intensiostatic process (14). The phase limits given by this
author are Cu Mo S , Cu Mo S , and Cu Mo S . Sev-
Mo Se S , x"6. The trend initiated, in the previous
ꢂ
ꢄ ꢂ
case, by the substitution with selenium in the Mo S unit is
ꢂ ꢃ
ꢁꢈꢄ
ꢂ ꢃ
ꢄ
ꢂ ꢃ
ꢋꢈꢄ
ꢂ ꢃ
eral de"nite changes were also reported for 0.7(Cu(1.2.
The number of steps is homogeneous in the two electro-
chemical process if the small and poorly di!erentiated signal
is attributed to the mentioned changes. As is expected, the
cation rates in the intensiostatic insertion are slightly in-
ferior to those obtained in potentiostatic process.
Another feature arises from the examination of the signals
position. In a previous study, we reported that the exploita-
tion of voltamperograms allows the calculation of *G free
energy of formation of the ternary phases according to
con"rmed for this binary, namely poorly di!erentiated in-
sertion signals (Fig. 3c) and two main oxidation areas (the
"rst one centered at 250 mV and the second one at
#350 mV). The shift to a more cathodic intercalation po-
tential becomes more pronounced. The stoichiometry of
limit phase is reaching a "nal value of 2.52 Cu by Mo Se S
ꢂ
ꢄ ꢂ
unit. The cell parameters (Table 1) reveal the presence of
cations in channels.
Mo Se S , x"4. From this selenium content, the vol-
ꢂ
ꢅ ꢅ
yCu3#Mo S PCu Mo S (21). On the basis of equation, tamperograms are less complex (Fig. 3d). The insertion
ꢂ ꢃ
W
ꢂ ꢃ
the electrochemical insertion in the "rst potential range arises in two successive steps. The "rst one takes place from

BEHAVIOR OF EXTRA Mo ATOMS IN THE CHEVREL PHASE
203
FIG. 3. Voltamperograms for a-Mo Se S compounds in 1 M CuSO medium, supporting disk area, s"6.2 mmꢄ, temperature ¹"603C, scan
ꢂ
ꢃ\V V
ꢅ
rate v"10 mV min\ꢁ; (a) x"8, (b) x"7, (c) x"6, (d) x"4, (e) x"2, (f) x"1, (g) x"0.

204
KAIDI ET AL.
TABLE 1
Hexagonal Lattice Parameters, Unit Cell Volumes, Analytical Data of a-Mo6Se8؊xSx Binaries and a-CuyMo6Se8؊xSx Ternaries
Prepared by Cathodic Reduction in Copper Aqueous Electrolyte
Parameters of initial binary
(A) (A) (Aꢋ)
Parameters of intercalated compund
(A) (A) (Aꢋ)
Observations and
Cu conent estimation
from literature data
Cu content determined
by microprobe analysis
Phase
a
c
<
a
c
<
)
)
)
)
)
)
250 mV
11.32
100 mV
10.22
10.35
10.47
10.60
10.63
10.75
a-Mo S
9.20
10.88
797
9.40
809
1.13
1 (1)
ꢂ ꢃ
9.69
9.79
9.72
9.94
9.96
9.97
831
860
855
907
914
926
3.72
2.52
1.67
2.26
2.12
2.14
2.7 (1)
Intercalation
(2 (24)
2.6 (23)
Intercalation
2 (12)
a-Mo Se S
ꢄ ꢂ
a-Mo Se S
ꢅ ꢅ
a-Mo Se S
ꢂ ꢄ
9.33
9.42
9.48
9.53
9.54
10.82
10.90
11.03
11.10
11.21
816
837
859
873
884
ꢂ
ꢂ
ꢂ
a-Mo Se S
ꢂ
a-Mo Se
ꢂ
ꢉ ꢁ
ꢃ
280 to 200 mV while the second step is centered at 125 mV. copper atoms to enter the structure) is strongly correlated to
The anodic scan gives evidence for only one wide signal with the sulfur content (x) of the cathode material. For instance,
a little shoulder. A kinetic phenomenon is probably respon- the maximum of copper content decreases from 4 to 2.14 as
sible for this electrochemical behavior. The insertion rate x decreases from 8 to 0, except for x"4 where an insertion
(Table 1) stabilizes at a value of 1.67.
rate of 1.67 is observed. This latter result is di$cult to
explain rationally. Nevertheless, the shape of the i"f (E)
curves depictes this trend, namely, the increase of insertion
rate follows the increase of insertion signal number.
Mo Se S , x"2. The potentiodynamic curve (Fig. 3e)
ꢂ
ꢂ ꢄ
presents an obvious similarity with that corresponding to
the precedent binary, both in shape and in number of
signals, though it does not agree in potential scale. The
intercalation systems (either in reduction or in oxidation)
occur at rather lower values of potential and arises the
The lower potentials for the "rst intercalation systems
observed in richer selenium compounds might basically be
due either to a higher kinetic hindrance of the copper ions
transport in the solid phase or to a lower thermodynamic
stability of selenide phases compared to the sul"de com-
pounds. Since the peak gaps for a step (de"ned as the
di!erence between anodic and cathodic peak potentials)
remain approximatively constant whatever the selenium
content, the intercalation phenomenon does not seem to be
directly related to the cation di!usion in the solid. On this
basis, one would retain the second possibility. Among the
chalcogenides compounds, a higher free energy of reaction
has always been observed for the sul"des than for the
selenides. Our results con"rm these observations.
formation of Cu
gives clear evidence of a structural modi"cation (Table 1).
Mo Se S . The crystallographic study
ꢄꢈꢄꢂ
ꢂ ꢂ ꢄ
The hexagonal lattice constants observed for this limit
phase agree with literature data (Cu Mo Se
scribed by Mc Kinnon (23)).
S
de-
ꢄꢈꢂ ꢂ ꢂꢈꢆ ꢁꢈꢆ
Mo Se S , x"1. The same general observations (Fig.
ꢂ
ꢉ ꢁ
3f) can be noted for this value of x. The copper content for
the Mo Se S is equal to 2.12 for an intercalated phase
ꢂ
ꢉ ꢁ
obtained at 100 mV.
Mo Se , x"0. The voltamperogram (Fig 3g) exhibits
The results of the structural investigations for the di!er-
a signi"cant signal (E"150 mV). Additional potential ent intercalated phases reveal changes in lattice parameters
ꢂ
ꢃ
sweeps reveal a supplementary but very small step at (an increase of a concomitant with a decrease of c ). The
250 mV. Nevertheless, the existence of two steps is in this intercalation logically induces an increase of cell volume.
case, homogeneous with that has been observed for the Another point to note is the good correlation between the
&
&
lower values of selenium.
copper content determined by X-ray and microanalysis
The stoichiometry determined by the microanalysis data.
shows a homogeneous intercalation with a content equal to
2.14 in agreement with the X-ray di!raction pattern.
V.2. b-Mo Se
S
ꢂ
ꢃ\V V
A partial substitution with selenium in the Mo S unit
ꢂ ꢃ
logically modi"es the voltamperograms. They give evidence
The b-Mo Se
S
pseudo-ternaries are obtained by
ꢂ
ꢃ\V V
for a progressive change from four steps observed for heating the a-equivalent ones. Indeed, from 4703C, a partial
Mo S to the rather unique one in the case of the Mo S
binary. The maximum uptake of guest ions (number of denum, simultaneously followed by a Mo self-intercalation
decomposition of a-compounds in Mo Se S and molyb-
ꢂ ꢃ ꢂ ꢃ
ꢂ
ꢄ\X X

BEHAVIOR OF EXTRA Mo ATOMS IN THE CHEVREL PHASE
205
leads to the new sulfoselenide series. Note that MoSe
phases in the "nal products are always, of course, observed one obtained with a-Mo S . The intercalation systems
(1, 9, 10). Before undertaking a comparative study on the are displaced to lower potentials and indicate lower free
S
is the fact that the curve does not surimpose on the "rst
ꢄ\X X
ꢂ ꢃ
copper intercalation, we had to ensure that the electro- energy reactions. These observations led us to conclude the
chemical response was not disturbed by the presence of di$culty even the inability of this b-Mo S material
ꢂ ꢃ
MoSe S . A simple way to test this eventuality is to cycle to intercalate copper not only on surface sites but also
ꢄ\X X
a MoSe S cathode in a copper solution. We observed the inside the crystallites. This is clearly demonstrated by
ꢄ\X X
absence of detectable signal whatever the tested compounds the X-ray di!raction. There are very slight changes in
(MoS , MoSe S , MoSe ). In our conditions, the electro- lattice parameters (Table 2). The result of the microanalysis
ꢄ
ꢁ ꢁ
ꢄ
chemical insertion of copper into MoSe
S
can be indicates a low content of 0.28 copper by Mo S unit
ꢄ\X X
ꢂ ꢃ
completely ruled out. X-ray patterns substantiate this while the content of extra Mo does not vary. This behavior
suggestion.
may simply result from structural factors as we will discuss
Electrochemical data from electrodes based on b- below.
Mo Se S are plotted in Figs. 4a}e.
ꢂ
ꢃ\V V
ꢂ ꢃ
b-Mo Se S , x"6. In contrast with the b-Mo S bi-
ꢂ
ꢄ ꢂ
ꢂ ꢃ
b-Mo S , x"8. A striking feature is the di!erent be- nary, the cyclic voltammetry investigations (Fig. 4b) on the
havior of this material with respect copper intercalation. Cu/b-Mo Se S system demonstrate a higher ionic mobility
First, we had to decrease the potential sweep rate in order to of Cu ions in the Mo Se S framework. The curve exhibits
bring evidence for a possible insertion. The electrode re- several signals, though they are poorly di!erentiated.
sponse with a potential sweep rate similar to the ones The appearance of successive steps is probably due to the
previously used is not very representative of a phenomenon existence of di!erent types available lattice sites. Moreover,
because of less well-de"ned peaks and of the signal inten- the reduction process begins at 300 mV. The electrochemi-
sities that are very small. However, the signals show a cal formation of b-Cu Mo Se S is easier than that of
ꢂ
ꢄ ꢂ
ꢂ
ꢄ ꢂ
W
ꢂ ꢄ ꢂ
reversible intercalation process. Another interesting point b-Cu Mo S . The value of copper content (1.68) and
W
ꢂ ꢃ
FIG. 4. Voltamperograms for b-Mo Se S compounds in 1 M CuSO medium, supporting disk area, s"6.2 mmꢄ, temperature ¹"603C, scan
ꢂ
ꢃ\V V
ꢅ
rate v"5 mV min\ꢁ (for a) and v"10 mV min\ꢁ (for b}e): (a) x"8, (b) x"6, (c) x"4, (d) x"2, (e) x"1.

206
KAIDI ET AL.
TABLE 2
Hexagonal Lattice Parameters, Unit Cell Volumes, Analytical Data of b-Mo6Se8؊xSx Pseudo-Ternaries and
b-CuyMo6Se8؊xSx Compounds Prepared by Cathodic Reduction in Copper Aqueous Electrolyte
Parameters of initial binary
(A) (A) (Aꢋ)
Parameters of intercalated compound
(A) (A) (Aꢋ)
Observations and
Cu conent estimation
from literature data
Cu content determined
by microprobe analysis
Phase
a
c
<
a
c
<
)
)
)
)
)
)
b-Mo S
9.09
9.39
11.27
10.68
804
816
9.08
9.56
11.29
10.39
807
821
0.28
1.68
+0
Interaction
ꢂ ꢃ
b-Mo Se S
(Cu content in
a-equivalent component
(2 (24)
ꢂ
ꢄ ꢂ
b-Mo Se S
9.47
9.56
10.77
10.88
836
861
9.77
9.84
10.48
10.62
866
889
1.67
1.52
ꢂ
ꢅ ꢅ
ꢂ ꢄ
b-Mo Se S
2.6 (12)
ꢂ
structural data (Table 2) con"rm the above trend and agree phases present another di!erence with the previous binary,
quite well with the electrochemical data.
the guest molybdenum content is much lower (Mo "0.41
ꢀ!ꢁ
for x"6).
b-Mo Se S , x"4. Figure 4c shows the voltampero-
ꢂ
ꢅ ꢅ
As copper intercalation into b-Mo S proceeds, the cop-
ꢂ ꢃ
grams of b-Mo Se S . The curve is less complex in compari-
ꢂ
ꢅ ꢅ
per ions could "rst occupy the small tetrahedral sites in the
son with the precedent one. The insertion arises in two
successive steps. The second signal (150 mV) is the most
important. In the course of the oxidation process, one global
oxidation peak with a shoulder is distinguished. The posi-
tion and the importance of signals are homogeneous with
those observed for the a-equivalent binary consistent with
a similar behavior for both phases. This is con"rmed by
either the crystallographic (Table 2) or analytical data
(equivalent cationic content).
cavities (2), but their migration into the channels will be
blocked by Mo atoms (amount larger than one) sitting on
the cavity (1). Furthermore, the existence of a large energy
barrier between outer sites of neighboring clusters accord-
ing to the model proposed by Dudley enhances this impossi-
bility of migration. Besides, the size of copper ion is certainly
not large enough to liberate the inner sites occupied by Mo.
The intercalation of copper atoms into b-Mo S induces no
ꢂ ꢃ
changes in the unit cell volume. This implies that the
b-Mo Se S , x"2. The shape of the i"f (E) curve Mo Mo S matrix does not expand and the displacement
ꢂ
ꢂ ꢄ
ꢁꢈꢃ
ꢂ ꢃ
ꢄ>
(Fig. 4d) is closely similar to that for Cu/a-Mo Se S system of (Mo) does not take place by an overintercalation of
ꢂ
ꢂ ꢄ
ꢄ
(two insertion steps and one wide anodic signal). Micro- copper contrary to the case with lithium alluded to above.
analysis con"rms the formation of an intercalated com- The electrochemical data (absence of quantitative signals,
pound but with a copper content slightly inferior to that of lower potentials) are consistent with these observations.
a-Mo Se S . A smaller variation of the lattice parameters Our analytical data which show that the copper content in
(Table 2) is observed from a "9.84 A and c "10.62 A b-Mo S is close to zero (0.28) and the molybdenum one is
ꢂ ꢃ
(< "889 A ) to a "9.94 A and c "10.60 A (< " constant strongly support the above assumptions.
907 A ) for the b- and a-Cu Mo Se S compounds, respec-
ꢂ
ꢂ ꢄ
&
&
ꢋ
&
&
&
&
ꢋ
For the b-Mo Se S (xO8), the situation is totally
di!erent. Crystallographic studies on Cu Mo Se
W
ꢂ
ꢂ ꢄ
ꢂ
ꢃ\V V
tively.
S
W
ꢂ
ꢃ\V V
The b-sulfoselenide compounds show intercalation pro- (y"1.8) compounds obtained by ceramic way have shown
cess resulting in compounds of formula b-Cu Mo Se
(or better (Cu Mo )Mo Se S ). A simple displacement copper is electrochemically introduced into the a-host lat-
between copper and molybdenum (2) is to be ruled out tice, it can sit on the available cavity (cavity (1)). In con-
S
that copper atoms "rst occupy the inner site (cavity (1)). As
W
ꢂ
ꢃ\V V
W
X
ꢂ
ꢃ\V V
because the initial molybdenum (2) content remains con- tinuation of the reduction process, it appears that the guest
stant after copper intercalation.
cation would have to occupy tetrahedral sites (cavity (2))
The Mo X phases present two types of sites available for which are partially "lled with Mo atoms. It is conceivable
ꢂ ꢃ
intercalation, a large pseudo-cubic cavity (site 1) com- that the presence of atom in this second cavity slows down
pressed along the 3-fold axis and six other small and very but does not prevent the copper transport. On this basis,
distorted cavities (site 2) forming a kind of octahedron one would expect an electrochemical intercalation into b-
around the 3-fold axis. If both sites are empty in a-binaries, Mo Se S to be less easy than into a-phases leading to
ꢂ
ꢃ\V V
they are partially occupied in b-Mo Se S series. The a shift to the lower free energy potentials. This conclusion is
ꢂ
ꢃ\V V
cavity 1 (inner site) is "lled by Mo atom in b-Mo S leading veri"ed through voltamperograms. Another remarkable
Mo S . In contrast, the ternary molybdenum "nding arises from the copper content. Since sites 2 were
atom sits in the outer site 2 for b-Mo Se S (xO8). These equally occupied in pseudobinaries, one would expect the
ꢂ ꢃ
to Mo
ꢁꢈꢁꢃ
ꢂ ꢃ
ꢂ
ꢃ\V V

BEHAVIOR OF EXTRA Mo ATOMS IN THE CHEVREL PHASE
207
ion content intercalated into these phases, on the one
The electrochemical curves are shown all together in
hand, to be independent of selenium substitution and, on Figs. 5a, 5b, and 5c for Cu Mo Se S where x"0, 2,
W
ꢂ
ꢃ\V V
the other hand, to be smaller than that observed in a- and 4.
homologue phases. This is clearly demonstrated by the
A common feature is the shape of i"f (E) which is quite
analytical data, within their accuracy. The maximum up- similar to that exhibited by b-Mo Se S electrodes (two
take of guest ions in these materials does not vary with the less well-de"ned intercalation steps and one wide oxidation
selenium content (1.52 to 1.68, inferior, in all cases, to the signal). Their position in potential scale shifts again but
ꢂ
ꢃ\V V
a-insertion rate).
smoothly. The observed intensities would represent a com-
plete transformation of initial binary. The X-ray patterns
con"rm the existence of intercalated compounds (Table 3).
The lattice parameters, particularly the unit cell volumes,
V.3. HT-Mo Se
S from Direct Synthesis
ꢂ
ꢃ\V V
We now brie#y present the results of similar studies on led us to conclude that the uptake of copper atoms into the
Cu/high temperature HT-Mo Se S systems. Only richer ceramic phases is slightly lower. The microprobe analysis is
ꢂ
ꢃ\V V
selenium phases (x45) can be synthesized directly. The in good agreement with this prediction.
high temperature phases present structural characteristics
Another point arises from the tiny di!erence in signals
similar to those of b-homologue Mo Se S . The cell para- de"nition for a- and HT-Mo X curves. This result may be
ꢂ
ꢃ\V V
ꢂ ꢃ
meters are very close together. On this basis, one might explained in terms of granulometry. We have shown that the
expect molybdenum atoms to be incorporated into the grain size parameter of samples can in#uence the shape of
lattice of high temperature binaries. In a recent study, S. voltamperograms. The a-samples result from a chemical
Belin et al. have reported the structure of these latter phases oxidation (by dilute hydrochloric acid) of ternary phases.
(11). They have shown that Mo atoms are present in the This deintercalation causes some damage to the crystallites.
same cavity in an equivalent proportion. Both "ndings Their size is decreasing. This fact is favorable to an enhance-
suggest that the high temperature host materials behave as ment of the di!usional phenomenon taking place inside the
the b-phases with respect to copper intercalation.
crystal.
FIG. 5. Voltamperograms for high-temperature HT-Mo Se
ature ¹"603C, scan rate v"10 mV min\ꢁ: (a) x"0, (b) x"1, (c) x"2, (d) x"4.
S
compounds in I M CuSO medium, supporting disk area, s"6.2 mmꢄ, temper-
ꢂ
ꢃ\V V
ꢅ

208
KAIDI ET AL.
TABLE 3
Hexagonal Lattice Parameters, Unit Cell Volumes, Analytical Data of High Temperture HT-Mo6Se8؊xSx Pseudo-Ternaries and
HT-CuyMo6Se8؊xSx Compounds Prepared by Cathodic Reduction in Copper Aqueous Electrolyte
Parameters of initial binary
(A) (A) (Aꢋ)
Parameters of intercalated compound
(A) (A) (Aꢋ)
Observations and
Cu content determined Cu conent estimation
Phase
a
c
<
a
c
<
)
by microprobe analysis
from literature data
)
)
)
)
)
HT-Mo Se S
9.53
9.57
10.67
10.87
838
862
9.75
9.83
10.45
10.61
860
888
1.47
1.58
(2 (24)
(2.6 (23)
Intercalation
ꢂ
ꢅ ꢅ
ꢂ ꢄ
HT-Mo Se S
ꢂ
HT-Mo Se S
9.58
10.99
874
9.88
10.70
905
1.42
( Cu content in
a-equivalent compound
ꢂ
ꢉ ꢁ
The chemistry of the intercalation into the Chevrel phase molybdenum atoms). The present work has revealed that
Mo X , for a "xed cation, depends upon the nature, the the voltamperograms and their amplitude are also a!ected
ꢂ ꢃ
size, and the occupancy of the cavities as well as the elec- by these extra atoms in the channels and can be used to
tronic structural factors. The following paragraph will deduce structural informations. For instance, since the a-
adress the outlines of the electrochemical behavior of the and HT-Mo Se compounds have the same structure, one
ꢂ
ꢃ
di!erent binaries series with respect to copper insertion. would expect the same electrochemical characteristics. Our
Figure 6 shows the variation of the integral copper insertion study has revealed di!erences between these two binaries in
rate as function of the sulfur content and the synthesis type their general shape of i"f (E) curves and their intercalated
of pseudobinaries. This "gure provides evidence that the cation rate. This anomaly can be ascribed either to an e!ect
more a-rich sulfur phases have better intercalation capabili- of granulometry like alluded to above or to the existence of
ties than the richer selenium binaries. Such results may very low extra molybdenum content into channels for the
indicate the predominance of electronic factors since both thermal phase not easily detectable.
available sites are empty. This point is clearly con"rmed by
This work has shown that the study of electrochemical
the electrochemical curves. The number of intercalation reactivity of host lattices, through the examination of the
steps increases and the position of signals shifts smoothly voltamperograms, accounts for the structural states of dif-
from lower to higher energy reaction with increasing sulfur ferent binaries forms and provides supplementary informa-
content in the unit Mo Se S . The evolution of electro- tions about the solid material which support even con"rm
ꢂ
ꢃ\V V
chemical reactivity of a-pseudobinaries is homogeneous the analytical and crystallographic data.
with the variation of S/Se ratio.
We have also found that molybdenum can never be
Another striking fact is that the maximum ternary ele- removed from the b- and thermal binaries either by an
ment content in materials not only correlates strongly to electrochemical oxidation or by an electrochemical overin-
the presence of extra molybdenum atoms but also to the tercalation displacement at room temperature. Since such
nature of occupied cavities. The b-Mo S is a perfect temperatures are obviously uncompatible with a relatively
ꢂ ꢃ
illustration. Moreover, the insertion rate of b- and direct high mobility of molydenum, the use of nonaqueous electro-
synthesis binaries are likely to be similar in agreement with lytes that allow higher temperatures should improve the
the structural considerations (occupancy of cavity (2) by di!usional phenomenon.
VI. CHEMICAL BEHAVIOR OF b-Mo6S8
VI.1. Chemical Deinsertion
The most used process to deintercalate cations from ter-
nary Chevrel phases is the hydrochloric acid leaching. Of
course, we "rst attempt this process on b-Mo X and HT-
ꢂ ꢃ
Mo X . The X-ray di!raction patterns recorded on the
ꢂ ꢃ
products treated with this method show no change, and the
&&washing solutions'' analyzed by absorption spectroscopy
do not contain any molybdenum element.
So, we were interested in the method used by Potel et al.
to deinsert indium from InMo S (3). The method is based
on the use of HCl gas #ow passing on powdered compounds
FIG. 6. Variation of copper content in a-, b-, and HT-Mo Se
binary and pseudo-ternary compounds.
S
ꢂ ꢃ
ꢂ
ꢃ\V V

BEHAVIOR OF EXTRA Mo ATOMS IN THE CHEVREL PHASE
209
heated at 3003C. At this temperature, InCl is very volatile, added and closely mixed with b-Mo S . The mixture was
ꢋ
ꢂ ꢃ
so this reaction is very interesting to produce pure a-Mo S . slowly raised to 4003C for several days with intermediate
ꢂ ꢃ
In the same way we studied b-Mo S but at about 6003C rehomogenization. Most of time to complete the reaction
during 6 h. Quickly a release of molybdenum chloride is we need to increase the temperature until 6003C. When the
ꢂ ꢃ
shown. The compound is then cooled under HCl gas to reaction is "nished, the X-ray di!raction pattern always
prevent the b-Mo S from re-forming. A loss of weight with exhibits a mixture of Cu Mo S compounds with Mo ele-
ꢂ ꢃ
W
ꢂ ꢃ
a Mo/Mo S ratio of &1.08 can be detected on the com- ment and MoS binary (the latter comes from the initial
ꢂ ꢃ
ꢄ
pound after such a treatment. The X-ray di!raction pattern b-Mo S formation).
shows the obtained compound to be a-Mo S .
Elsewhere we have reported that this method was also b-Mo S because the other b- or HT-Mo Se S com-
e$cient on the whole b- and HT-series. So, to remove Mo(2) pounds do not contain enough Mo(2) atoms to detect Mo
guest atoms from the host we just need to increase the element by X-ray di!raction pattern if it is removed. b-
ꢂ ꢃ
The cationic exchange reaction was only undertaken on
ꢂ ꢃ
ꢂ ꢃ
ꢂ
ꢃ\V V
temperature of the method because molybdenum chloride is Mo S was the best example to simply carry out this study.
ꢂ ꢃ
less volatile than InCl3, and moreover these Mo(2) atoms
We also tried nickel cations. The result is still the same,
are bonded with either both to form quadruple bond (it is Cu or Ni substitute for Mo(2) atoms in the cavity (1) of the
the b-Mo S particular case) or with Mo(1) atoms to form channels.
ꢂ ꢃ
kind of Mo clusters (case of b-Mo Se S (xO8) and
Incomplete reactions at lower temperatures show that
small amounts of Cu (or Ni) can be overinserted in b-
ꢉ
ꢂ
ꢃ\V V
HT-Mo Se S series) (10,11).
ꢂ
ꢃ\V V
In ¹RMo X (¹R"rare earth element), PbMo X , and Mo S . E!ectively, we observe at the beginning of the
ꢂ ꢃ
ꢂ ꢃ
ꢂ ꢃ
SnMo X (X"Se or S), we recently showed that the cation reaction a mixture of two Chevrel phases: for instance, the
ꢂ ꢃ
can be removed with this last method at the same temper- reaction Cu#b-Mo S at 4003C for 2 days gives one Chev-
ꢂ ꢃ
ature. In general, the deintercalation activation energy de- rel phase with unit cell parameters very close to ones of b-
pends on the inserted metal ion (unfavorable steric e!ects) Mo S ((Cu& Mo)Mo S : a "9.10 A, c "11.34}11.35 A,
ꢂ ꢃ
ꢊꢈꢄ
ꢂ ꢃ
&
&
ꢋ
and also on the bonding strengths acting on the metal ion. < "813}815 A ) plus a second Cu Mo S phase (y&
&
W
ꢂ ꢃ
ꢋ
In the case of Mo(2) guest atoms, the latter predominates, 2.6}2.7: a "9.66 A, c "10.213 A, < "825 A ) with
&
&
&
while for Pb, Sn, or ¹R elements it is the size of cation which equivalent amounts. The longer the reaction and/or the
matters.
higher the temperature, the more Cu Mo S amount in-
W
ꢂ ꢃ
creases with y decreasing to the detriment of the &b-Mo S
and Mo impurities appear. At 6003C, 48 h, the close b-phase
disappears completely and we then observe only one
ꢂ ꢃ
VI.2. Cationic Exchange
We have studied the overinsertion of small cations in Cu Mo S phase (y&2: a "9.59 A, c "10.235 A, < "
W
ꢂ ꢃ
&
&
&
ꢋ
b-Mo S via ceramic routes at low enough temperature. 815 A ) with much more Mo and, of course, MoS .
ꢂ ꢃ
The electrochemical study has shown that it is very di$cult
ꢄ
We observe exactly the same phenomena with the 2 Cu,
to insert copper cation in this compound. It seems that and more with the same close b-phase and Cu Mo S
W
ꢂ ꢃ
the highest copper amount should be 0.28 Cu/Mo S
(y&2.8}2.0). From these observations, we can conclude
ꢂ ꢃ
unit. We have already reported that in b-Mo S , VEC that a very few copper atoms can be overintercalated in the
ꢂ ꢃ
is close to 22e/Mo exactly as for PbMo S . It was network but the overinsertion of copper atoms involves the
established that compounds with such a VEC should be removal of molybdenum atoms. These results con"rm the
ꢂ
ꢂ ꢃ
very stable. Previous studies (13) have shown that overinter- electrochemical investigations previously obtained on this
calation in PbMo S was very low; so, for b-Mo S
a
compound.
ꢂ ꢃ
ꢂ ꢃ
similar behavior could be expected. Then we attempt to
exchange Mo(2) atoms for small cations as Cu or Ni,
because it was reported in ternary Chevrel phases that small
cations can expel bigger ones (for instance copper can be
VII. CONCLUSION
We reported the electrochemical insertion of copper
exchanged for lead). In the "rst part of this paper, we have atoms into three series of Mo Se S phases. We showed
ꢂ
ꢃ\V V
seen that no Mo(2) atoms could be remove electrochemi- that a-, b-, and HT-forms can accommodate copper. The
cally from b-Mo X and HT-Mo X series because copper content decreases from a- binaries to b- and HT-
ꢂ ꢃ
ꢂ ꢃ
the method uses an aqueous electrolyte and therefore the ones. A striking di!erence has been observed in the behavior
reaction temperature is limited at ambient conditions. How- of b-Mo S . This phase practically does not intercalate with
ꢂ ꢃ
ever, the electrochemical route is very e$cient for overinser- copper. We showed that such a behavior might be explained
tion in these materials.
in terms of available vacant sites. In general, structural
So we have carried out some experiments with copper changes and analytical data correlate perfectly with the
element on b-Mo6S8 samples. Metallic copper powders in electrochemical curves. Finally, the voltamperograms can
ratios of 0.5, 1, 2, or 3 copper atoms per Mo S unit were be used as &&chimie douce'' potential photographs for any
ꢂ ꢃ

210
KAIDI ET AL.
material. The electrochemical curves can evidence the reac-
tivity and re#ect the structural state of host lattices.
9. S. Belin, R. Chevrel, and M. Sergent, Mater. Res. Bull. 33(1), 43 (1998).
10. S. Belin, L. Burel, M. Sergent, and R. Chevrel, Mater. Res. Bull., to be
published.
11. S. Belin, R. Chevrel, and M. Sergent, J. Solid State Chem. accepted for
publication.
12. K. Yvon, in &&Current Topics in Materials Science'' (E. Kaldis, Ed.),
Vol. 3, p. 53. North Holland, Amsterdam, 1979.
13. M. Sergent, R. Chevrel, C. Rossel, and ". Fischer, J. ¸ess-Common
Met. 58, 179 (1978). R. Chevrel, C. Rossel, and M. Sergent, J. ¸ess-
Common Met. 72, 31 (1980). F. Y. Fradin, and J. W. Downey, Mater.
Res. Bull. 14, 1525 (1979).
Chemical copper insertions in b-Mo S show cationic
ꢂ ꢃ
exchange reactions if the reaction is complete otherwise
a slight overinsertion of copper cations in the host. Chem-
ical methods, such as HCl gas #ow, indicate that Mo(2)
guest atoms can be deintercalated from the b-Mo X and
ꢂ ꢃ
HT-Mo X series. All these studies exhibit the high enough
ꢂ ꢃ
mobility of the Mo(2) guest atoms in the Chevrel phases.
14. R. Schollhorn, M. Kumpers, A. Lerf, E. Umlauf, and W. Schmidt,
Mater. Res. Bull. 14, 1039 (1979).
REFERENCES
15. M. Tovar, L. Delong, D. G. Johnson, and M. B. Maple, Solid State
Comm. 30, 551 (1979).
16. G. J. Dudley, K. Y. Cheung, and B. C. H. Steele, J. Solid State Chem. 32,
259 (1980).
1. R. Chevrel and M. Sergent, in &&Topics in Current Physics'' (". Fischer
and M. B. Maple, Eds.), Vol. 32, p. 25. Springer Verlag, Berlin, 1982.
2. ". Fischer, Appl. Phys. 16, 1 (1978).
3. R. Chevrel, M. Sergent, and J. Prigent, Mater. Res. Bull. 9, 1487 (1974). 17. P. J. Mulhern and R. R. Haering, Can. J. Phys. 62, 527 (1984).
M. Potel, P. Gougeon, R. Chevrel, and M. Sergent, Rev. Chim. Mineral. 18. T. Uchida, K. Watanabe, M. Wakihara, and M. Taniguchi, Chem. ¸ett.
21, 509 (1984).
1095 (1985). M. Wakihara, T. Uchida, T. Morishita, H. Wakamatsu,
and M. Taniguchi, J. Power Sources 20, 199 (1987). M. Wakihara, T.
Uchida, K. Suzuki, and M. Taniguchi, Electrochim. Acta 34, 867 (1989).
T. Uchida, Y. Tanjo, M. Wakihara, and M. Taniguchi, J. Electrochem.
Soc. 137, 7 (1990). S. Yamaguchi, T. Uchida, and M. Wakihara,
J. Electrochem. Soc. 138, 687 (1991).
4. R. Schollhorn, M. Kumpers, and J. O. Besenhard, Mater Res. Bull. 12,
781 (1977). R. Schollhorn, Angew. Chem. 92, 1015 (1980).
5. J. M. Tarascon, J. V. Waszczak, G. W. Hull, F. J. Di Salvo, and L. D.
Blitzer, Solid State Commun. 47, 973 (1983). J. M. Tarascon, F. J. Di
Salvo, D. W. Murphy, G. W. Hull, E. A. Rietman, and J. V. Waczczak,
J. Solid State Chem. 54, 204 (1984). J. M. Tarascon, F. J. Di Salvo, 19. J. M. Tarascon, T. P. Orlando, and M. J. Neal, J. Electrochem. Soc. 135,
D. W. Murphy, G. W. Hull, and J. V. Waszczak, Phys. Rev. B 29, 172, 804 (1988).
(1984). J. M. Tarascon, F. J. Di Salvo, J. V. Waszczak, and G. W. Hull, 20. W. R. Mc Kinnon and J. R. Dahn, Solid State Commun. 52, 245 (1984).
Phys. Rev. B 31, 1012 (1985). J. M. Tarascon, J. V. Waszczak, F. J. Di 21. C. Boulanger and J. M. Lecuire, Electrochim. Acta 33(11), 1577 (1988).
Salvo, and D. C. Johnson, Solid State Commun. 53, 849, (1985).
22. K. Yvon, A. Paoli, R. Flukiger, and R. Chevrel, Acta Crystallogr. B 33,
6. G. J Dudley and B. C. H. Steele, J. Solid State Chem. 32, 233 (1980).
3066 (1977).
7. C. Boulanger and J. M. Lecuire, Electrochim. Acta 32(2), 345 (1987). 23. L. S. Selwyn, W. R. Mc Kinnon, J. R. Dahn, and Y. Le Page, Phys. Rev.
C. Boulanger, and J. M. Lecuire, Solid State Ionics, 25, 45 (1987). B 33, 6405 (1986).
8. Umarji, G. V. Subba Rao, M. P. Janawadkar, and T. S. Radhakrish- 24. L. S. Selwyn, W.R. Mc Kinnon, and Y. Le Page, Phys. Rev. B 42, 10427
nan, J. Phys. Chem. 41, 421 (1980). (1990).