Two cation disulfide layers in the WxMO(1-x)S2 lamellar solid solution

Article abstract of DOI:10.1006/jssc.2001.9208

W_xMo_(1-x)S₂ compounds with both molybdenum and tungsten present in every disulfide layer, i.e., as an intralayer solid solution, can be obtained. Several synthesis routes have been used but all of them do not allow this solid solution, which evidences the decisive role of the nature of the precursor, to be obtained. We never observed formation of a W_xMo_(1-x)S₂ solid solution resulting from stacking of homocationic [WS₂] and [MoS₂] layers (interlayer solid solution).

Full text of DOI:10.1006/jssc.2001.9208

Journal of Solid State Chemistry 160, 147}155 (2001)
doi:10.1006/jssc.2001.9208, available online at http://www.idealibrary.com on
Two Cation Disulfide Layers in the W Mo
S
x
(1؊x) 2
Lamellar Solid Solution
Ce
H
cile Thomazeau,* Christophe Geantet,* Michel Lacroix,* Virginie Harl eH ,- Simone Benazeth,?
Christiane Marhic,A and Michel DanotA
*
Institut de Recherches sur la Catalyse, UPR CNRS 5401, 2 Avenue Albert Einstein, 69626 Villeurbanne cedex, France; -Institut Franmais du P e& trole,
1
}4 avenue de Bois-Pr e& au, 92852 Rueil-Malmaison cedex, France; ?Laboratoire pour l+Utilisation du Rayonnement Electromagn e& tique, UMR 130,
CNRS}Universit e& Paris Sud (Paris XI), bat. 209 D, BP 34, 91898 Orsay cedex, France; and AInstitut des Mat e& riaux Jean Rouxel, UMR 6502,
CNRS}Universit e& de Nantes, 2 Rue de la Houssini e% re, BP 32229, 44322 Nantes cedex 3, France
E-mail: michel.danot@cnrs-imn.fr
Received December 16, 2000; in revised form March 29, 2001; accepted April 12, 2001; published online June 11, 2001
By contrast, only sparse literature data concerning the
W
x
Mo(1؊x)
S
2
compounds with both molybdenum and tungsten synthesis of solid solutions such as M?M@
S were re-
V
(ꢂ\V) ꢀ
present in every disul5de layer, i.e., as an intralayer solid solu- ported, even if the M?S and M@S sul"des both crystallize
ꢀ
ꢀ
tion, can be obtained. Several synthesis routes have been used but with the same site geometry (octahedron or trigonal prism).
all of them do not allow this solid solution, which evidences the
decisive role of the nature of the precursor, to be obtained. We
Such a solid solution was evidenced for the Nb}Mo disulf-
ide system (5, 6) and proposed for the Mo}W one (7, 8).
However, the atomic arrangement was determined only in
the "rst case (6). The present work was undertaken in order
to determine the structural characteristics of a possible
never observed formation of a W
ing from stacking of homocationic [WS
x
Mo(1؊x)
S
2
2
solid solution result-
] and [MoS ] layers
2
(
interlayer solid solution). ( 2001 Academic Press
Key Words: molybdenum disul5de; tungsten disul5de; solid
solution; EXAFS.
W Mo
S solid solution for reasons that will be
V
(ꢂ\V) ꢀ
indicated in the following section.
a
x
b
I. INTRODUCTION
II. WHICH STRUCTURE FOR AN M M
S2 SOLID
(1؊x)
SOLUTION?
In MS lamellar transition metal disul"des, the layer is
ꢀ
constituted by three hexagonal atomic planes, a cationic one
Two structural models may be envisioned for
&
&sandwiched'' by two anionic ones. Within the layer, every a M?M@
S layered solid solution: (i) every layer could
V
(ꢂ\V) ꢀ
M atom is surrounded by a sulfur octahedron (M"Ti, for contain only one type of cation, the phase resulting from the
instance) or a sulfur trigonal prism (M"Mo or W for stacking of homocationic [M?S ] and [M@S ] layers (inter-
ꢀ
ꢀ
instance) with a coordination number of 6 in both cases. The layer solid solution), or (ii) the two cations could be present
crystals result from the stacking of such layers along the in every layer, the phase resulting from the stacking of
c axis according to the scheme [S}M}S] . Di!erent stack- heterocationic [M?M@
S ] layers (intralayer solid solu-
ing sequences exist, leading to various structures that are tion). In both cases, the Vegard's law should be obeyed, at
ꢁ
V
(ꢂ\V) ꢀ
designated as 1T (one layer per unit cell, trigonal symmetry), least approximately. However, except for ordered arrange-
2
H (two layers per unit cell, hexagonal symmetry), 3R (three ments for which superstructure re#ections would provide
layers per unit cell, rhombohedral symmetry), etc. (1). The valuable information, the position of the di!raction lines
anisotropy of such structures results from the coexistence of cannot make it possible to distinguish between the two
strong iono-covalent metal}sulfur bonds within the S}M}S models due to the statistical character of the X-ray di!rac-
layer and of weak bonds between adjacent layers (these tion technique, and only local methods, such as EXAFS for
weak bonds correspond to the interlayer space known as the instance, may provide relevant information. In a lamellar di-
van der Waals gap). This particular arrangement explains sul"de, the Fourier transform of the EXAFS spectrum re-
the extensive work devoted to these systems with regards to corded at the cation edge exhibits two intense peaks, the
their synthesis, properties, and applications as catalysts (2), "rst one corresponding to the "rst sulfur neighbors of
lubricating agents (3), or host compounds for the intercala- the absorbing cation and the second to its "rst in-plane ca-
tion of various species (4).
tionic environment (Fig. 1) without any multiple-scattering
1
47
0
022-4596/01 $35.00
Copyright ( 2001 by Academic Press
All rights of reproduction in any form reserved.

1
48
THOMAZEAU ET AL.
FIG. 1. Fourier transform obtained at the Mo K-edge for MoS : the "rst peak corresponds to the "rst sulfur neighbors of molybdenum, the second
ꢀ
peak to its "rst in-plane cationic environment.
contribution (9). Consequently, for an interlayer solid solu- periodicity, another possibility being the formation of
tion, the M? cation would &&see'' only M? cationic neighbors an uncommensurate mis"t arrangement, with notice-
and the M@ cation only M@ cationic neighbors, just as in able strains too. In the case of the the Nb Mo
S
V
(ꢂ\V) ꢀ
their own M?S and M@S disul"des, while for an intralayer system, the observed intralayer model thus appears ener-
ꢀ
ꢀ
solid solution the M? cations and the M@ cations would getically more favorable since it allows &&dilution'' of the
&
&see'' both M? and M@ neighbors. Therefore EXAFS is strains.
a powerful technique for discriminating between the two
The Mo}W disul"de system also seems propitious for the
possibilities. This was fairly illustrated in the case of the obtention of a solid solution since MoS and WS (10}12)
ꢀ
ꢀ
Nb Mo
S compounds: cell-parameter evolution versus are isostructural, with the same sulfur coordination ge-
V
(ꢂ\V) ꢀ
composition (x) showed that this system is e!ectively a solid ometry of the cation (trigonal prismatic), the same space
solution (5), and EXAFS results evidenced the presence of group (P6 /mmc), and the same stacking sequence (2H).
both Nb and Mo cations in every [MS ] layer (6) according Moreover, in contrast to the Nb}Mo system, the a cell
ꢃ
ꢀ
to the intralayer solid solution model.
parameters of the WS and MoS partners are nearly identi-
ꢀ
ꢀ
The in-plane periodicity is signi"cantly di!erent for cal (a "3.161 A
ꢀ1
and son, packing of alternate homocationic [WS ] and [MoS ]
>
and a "3.154 A
>
) (13). For this rea-
+
ꢀ
51
ꢀ
[
a
NbS ] and [MoS ] layers (a "3.330 A
,"1
). For this reason, the stacking of alternate layers can be expected to be possible without noticeable dis-
>
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
"3.161 A
>
+
ꢀ1
ꢀ
homocationic [NbS ] and [MoS ] layers (interlayer solid tortions (and strains) being induced. From this point of
S system appears ideal for obtaining an
in order to adapt both types of layers to a common a interlayer solid solution, at least for particular compositions
ꢀ
ꢀ
solution) would induce important distortions (and strains) view, the W Mo
V
(ꢂ\V) ꢀ

MOLYBDENUM}TUNGSTEN DISULFIDE LAMELLAR SOLID SOLUTION
149
(
x"0.50 for instance), that could allow simple ordered
arrangements along the c axis. Our study of the
W Mo
S system aimed to determine whether this geo-
V
(ꢂ\V) ꢀ
metrical factor is determinant concerning the nature of
a lamellar disul"de solid solution. Lastly it is to be noted
that the Mo}W disul"de system is very favorable for
EXAFS identi"cation of a possible solid solution because
backscattering amplitudes and phases are very di!erent for
molybdenum and tungsten.
III. EXPERIMENTAL
III.1. Synthesis
W Mo
S compounds were prepared using the am-
FIG. 2. X-ray di!raction patterns of WS , MoS , and W
Mo
S
V
(ꢂ\V) ꢀ
ꢀ
ꢀ
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
prepared at 673 K by treatment, under H S #ow, of respectively
monium tetrathiomolybdate (NH ) MoS and ammonium
tetrathiotungstate (NH ) WS salt precursors. Ammonium
tetrathiomolybdate was prepared according to the proced-
ure previously described (14). This method consists of dis-
ꢀ
ꢇ ꢀ
ꢇ
ammonium tetrathiotungstate, ammonium tetrathiomolybdate, and the
ꢇ ꢀ
ꢇ
product of cocrystallization of these two thiosalts.
solving at room temperature commercial ammonium consequently of the existence of a solid solution. We also
heptamolybdate (NH ) Mo O , 4H O (40 g) in 400 mL of tried to identify the cocrystallization products by EXAFS
ꢇ ꢈ
ꢉ ꢀꢇ
ꢇ ꢀ
ꢀ
an aqueous solution of (NH ) S (20%). After one week, red measurements: in the case of a two-phase system, the
crystals formed and grew. They were isolated by "ltration, cationic neighbors of Mo would be uniquely Mo and those
rinsed with acetone, and dried under vacuum. Chemical of W uniquely W; in the case of a solid solution, Mo and
analysis agreed with the expected (NH ) MoS formula. An W would have both Mo and W neighbors. Unfortunately,
ꢇ ꢀ
ꢇ
UV-Vis spectrum of diluted solution of this compound only due to the large interatomic distance ('6 A
exhibited intense absorption bands at 470, 318, and 242 nm. sponding peak could not be detected. Whatever their na-
According to Diemann and Muller (15), this is the signature ture, solid solution or intimate mixture of (NH ) WS and
>
), the corre-
K
ꢇ ꢀ
of pure (NH ) MoS since the presence of residual (NH ) MoS , the solids resulting from this cocrystallization
ꢇ ꢀ ꢇ ꢀ
ꢇ
ꢇ
ꢇ
oxothiomolybdate species would give rise to several other were then heated under a H /H S (15%) #ow (6 L/h), at
ꢀ
ꢀ
bands. Ammonium tetrathiotungstate was prepared using
the method described in the literature (16, 17). Forty grams
of ammonium dodecatungstate (NH ) W O , 14H O
ꢇ ꢂꢄ ꢂꢀ ꢇꢂ
ꢀ
were dissolved in 200 mL of a 1.75 M aqueous solution of
NH . The solution was then kept for 4 h under H S bubbl-
ꢃ
ꢀ
ing at 333 K and then cooled and maintained at 273 K
overnight. A precipitate of orange}yellow (NH ) WS crys-
ꢇ ꢀ
ꢇ
tals was isolated by "ltration, rinsed with acetone, and dried
under vacuum. As for the Mo precursor, chemical analysis
and UV-Vis spectroscopy con"rm the synthesis of pure
(
NH ) WS .
ꢇ ꢀ
NH ) WS and (NH ) MoS , taken in various propor-
ꢇ
(
ꢇ ꢀ
ꢇ
ꢇ ꢀ
ꢇ
tions, were then dissolved in water, and then the solution
dried under vacuum. X-ray di!raction patterns did not
make it possible to identify unambiguously the resulting
solids as (NH ) W Mo
system due to the similarity of the cell parameters of
S solid solution or a two-phase
ꢇ ꢀ V
(ꢂ\V) ꢇ
(
NH ) WS and (NH ) MoS (18, 19). However, for the
ꢇ ꢀ
ꢇ
ꢇ ꢀ
ꢇ
equimolar composition, a very weak line attributable to
the 001 re#ection could be detected. This line, forbidden
by the space group of (NH ) WS and (NH ) MoS (Pnma
ꢇ ꢀ
ꢇ
ꢇ ꢀ
ꢇ
(18, 19)), could be the signature of a solid solution
with ordered arrangement of the W and Mo cations.
Nevertheless, considering its weakness, this line alone
cannot be considered as a proof of the ordering, and decomposition of the corresponding cocrystallized thiosalt (G"400,000).
FIG. 3. TEM picture of W
Mo
S sample prepared at 673 K by
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ

1
50
THOMAZEAU ET AL.
XRD patterns of the samples obtained from thiosalts
decomposition were recorded at the IRC using a Bruker
D5005 di!ractometer (CuKa radiation).
ꢂ
III.3. TEM-EDX Analysis
TEM analysis were performed with a HF 2000 Hitachi
microscope (point-to-point resolution: 2.3A
>
) working at
2
00 kV and equipped with a Kevex EDX detector with
a Si}Li diode.
III.4. EXAFS Study
FIG. 4. Fourier transform obtained at the Mo K-edge for MoS and
ꢀ
EXAFS spectra were recorded at the Laboratoire d'Util-
isation du Rayonnement Electromagnetique (LURE), the
French Synchrotron Laboratory, on the XAS 13 spectrom-
eter, using a Si (111) monochromator for the tungsten
W Mo
S
compounds prepared at 673 K by treatment, under H S
V
(ꢂ\V) ꢀ
ꢀ
H
#
ow, of mixed thiosalts resulting from cocrystallization of ammonium
tetrathiotungstate and ammonium tetrathiomolybdate.
¸
edge (10,100}11,200 eV, 2 eV step, 2 s per point, sum of
S compounds three spectra) and a Ge (400) one for the molybdenum
'
''
6
73 K, for 2 h, in order to obtain W Mo
V
(ꢂ\V) ꢀ
under the same conditions that make it possible to obtain K-edge (19,800}21,000 eV, 4 eV step, 2 s per point, sum of
WS or MoS from the corresponding thiosalts.
three spectra). The storage ring (DCI) used 1.85-GeV posit-
A sample with x"0.50 was prepared in a similar rons with an average intensity of 250 mA. For the sample
way, but with a H /H S treatment carried out at higher preparation, pellets were pressed using quantities of powder
ꢀ
ꢀ
ꢀ
ꢀ
temperature (1073 K instead of 673 K); it will be referred to calculated in order to achieve suitable absorption coe$cient
as W Mo S -1073. Another sample (x"0.50) was and edge jump. Data were collected in the transmission
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
prepared at 673 K using a "nely ground mixture of mode by measurement of the beam intensities I and I,
ꢄ
(
NH ) WS and (NH ) MoS instead of the cocrystalliz- respectively, before and after passing through the sample,
ꢇ ꢀ ꢇ ꢀ
ation product obtained from the solution; it will be referred using ionization chambers.
ꢇ
ꢇ
to as W
Syntheses of the W Mo
Mo
S -MM (MM"mechanical mixture).
Standard analysis of the EXAFS spectra (normalization,
S compounds were also at- background removal, Fourier transformation, and curve
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
V
(ꢂ\V) ꢀ
tempted from the two disul"des. WS and MoS were "rst "tting) were carried out using the SEDEM software (20)
ꢀ
ꢀ
obtained by a three-day reaction of sulfur with metal wires, with FEFF (21) theoretical phase and amplitude functions.
in evacuated Pyrex or silica tubes, at 793 and 923 K, respec- The curve-"tting procedure was performed in R-space.
tively. Mixtures of the two disul"des taken in the x/(1!x) Fourier transformation of the normalized kꢃ-weighted
molar ratio were then carefully ground before "ring at EXAFS signal was performed over the k-range 2.5}16 A
\ꢂ
>
8
23 K for one day and at either 1073 or 1273 K for one with Kaiser window functions. Coordination numbers (N),
week and "nally air-quenched or slowly cooled (4 K/h). interatomic distances (R), Debye}Waller parameters (pꢀ),
These products will be referred to as W Mo
S -DM and energy shifts (*E ), were used as variables in the "tting
V
(ꢂ\V) ꢀ
ꢄ
(
DM"disul"de mixture).
procedure. Scale factors Sꢀ were "xed at 0.8.
ꢄ
We also prepared a W
Mo
S
sample by direct
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
synthesis, i.e., reaction of sulfur on a mixture of "ne Mo and
W powders. The Pyrex tube was heated at 773 K for one
week. After grinding, the resulting product was annealed
in a silica tube at 1273 K for another week and "nally
slowly cooled. This compound will be referred to as
IV. RESULTS AND DISCUSSION
IV.1. The W Mo
S Compounds Obtained at 673 K
V
(ꢂ\V) ꢀ
from the Cocrystallized ¹hiosalts
The low temperature (673 K) used for the synthesis of the
W
Mo
S -DS (DS"direct synthesis).
W Mo
S (x"0.25, 0.50, 0.75) compounds, and of the
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
V
(ꢂ\V) ꢀ
pure MoS and WS (x"0 and 1), leads to poorly organ-
ized samples as demonstrated by the broadness of the X-ray
di!raction lines (Fig. 2) and by the image of the small slabs
ꢀ
ꢀ
III.2. X-Ray Diwraction Study
X-ray di!raction patterns of the samples prepared from with lattice fringes spacing of about 0.6 nm, characteristic of
the disul"des and by direct synthesis were recorded at the the MoS (or WS ) (002) basal plane (Fig. 3). It can never-
ꢀ
ꢀ
IMJR with an INEL CPS120-equipped powder dif- theless be seen that all the di!ractograms correspond to the
fractometer set up in a horizontal Debye}Scherrer hexagonal cell of MoS or WS . However, taking into
ꢀ
ꢀ
geometry, using CuKa radiation (quartz monochromator). account the important linewidth and the similarity of the
ꢂ

MOLYBDENUM}TUNGSTEN DISULFIDE LAMELLAR SOLID SOLUTION
151
since Mo}W and Mo}Mo peaks have opposite phases:
because of the resulting negative interference, the apparent
peak positions are further apart than expected on the basis
of the di!erences in distance and phase shift. In fact, this
Mo}W contribution is already present for lower x values
even if it does not clearly appear in the Fourier transform.
E!ectively (Table 1, Fig. 6), re"nements of the EXAFS
spectra show that the cationic in-plane environment of
molybdenum contains both molybdenum and tungsten,
with N(W)/N(Mo) ratios compatible, taking into account
the estimated error on the re"ned N values, with those
involved in the synthesis. It thus appears that the [MS ]
ꢀ
layers are heterocationic, according to the [W Mo
formula.
S ]
V
(ꢂ\V) ꢀ
A similar observation can be done at the W ¸ -edge, as
FIG. 5. Fourier transform obtained at the W ¸ -edge for WS and
''
'
ꢀ
'''
W
Mo
S prepared at 673 K by treatment, under H S #ow, of the illustrated by Fig. 5, which allows the Fourier transforms
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
ꢀ
mixed thiosalt resulting from cocrystallization of equimolar proportions of
ammonium tetrathiotungstate and ammonium tetrathiomolybdate.
obtained for WS and W
Mo
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
Mo S , the cationic environ-
ꢄꢅꢆꢄ ꢀ
S (given as an example)
ꢀ
to be compared. For W
ꢄꢅꢆꢄ
ment signal is split into two peaks, due to the presence of
W}Mo and W}W contributions (with opposite phases). It
MoS and WS cell parameters, these data cannot make it means again that every [MS ] layer contains molybdenum
ꢀ
ꢀ
ꢀ
possible to determine if the di!raction pattern of each of the and tungsten. This is con"rmed by the re"nement of the
mixed W Mo
S
compounds must be interpreted EXAFS spectrum (Table 2, Fig. 7). It can thus be concluded
V
(ꢂ\V) ꢀ
as a unique set of di!raction lines corresponding to the that these compounds belong to a solid solution that
expected solid solution, or as the addition of the MoS and involves heterocationic [W Mo
S ] layers, i.e., an in-
ꢀ
WS diagrams corresponding to a two-phase system. It will tralayer solid solution.
ꢀ
V
(ꢂ\V) ꢀ
be seen below that EXAFS results unambiguously answer
this question.
IV.2. W
Mo
S -1073, the Compound Obtained
The "rst peak of the Fourier transform of the EXAFS
spectra recorded either at the Mo K-edge (Fig. 4) or at the
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
at 1073 K from the Cocrystallized ¹hiosalts
W ¸' edge (Fig. 5) is fairly reproduced with +6 sulfur
EXAFS data make it possible to conclude that this com-
>
, which corresponds to pound is perfectly identical with those obtained at 673 K.
''
neighbors at a distance d+2.40 A
the expected anionic surrounding of the cations. Concerning The Fourier transforms are very similar and the "ttings also
now the cationic environment, it can be seen at the make it possible to conclude that the cationic surrounding
Mo K-edge (Fig. 4) that the Mo}Mo contribution (second of each of the two cations is constituted by both molyb-
peak) progressively decreases when the tungsten amount denum and tungsten neighbors (Table 3), in agreement with
increases. For a high tungsten amount (x"0.75), the the existence of an intralayer solid solution. Therefore this
cationic neighborhood signal is clearly split. This splitting new solid solid solution is thermally stable at least up to
can be related to the presence of a Mo}W contribution, 1073 K.
TABLE 1
EXAFS Re5nement Results (Mo K-edge): Characteristics of the Cationic In-plane Environment of Molybdenum in MoS
2
and
W
x
Mo(1!x)
S
2
Compounds Prepared at 673 K by Treatment, under H S Flow, of Mixed Thiosalts Resulting from Cocrystallization
2
of Ammonium Tetrathiotungstate and Ammonium Tetrathiomolybdate
Mo shell
W shell
N(W) pꢀ;10ꢃ (A
R(Mo) (A
>
)
N(Mo)
pꢀ;10ꢃ (A
>
ꢀ)
*Eꢄ (eV)
R(W) (A
>
)
>
ꢀ) *Eꢄ (eV)
MoS
3.14
3.15
3.16
3.16
5.9
4.5
3.2
2.3
4.2
4.0
4.5
3.5
!1.6
2.8
*
*
1.9
2.1
4.8
*
8.8
4.7
4.2
*
5.8
6.9
!1.4
ꢀ
Mo
Mo
Mo
W
W
W
S
3.15
3.17
3.15
ꢄꢅꢉꢆ ꢄꢅꢀꢆ ꢀ
S
2.4
2.0
ꢄꢅꢆ ꢄꢅꢆ ꢀ
S
ꢄꢅꢀꢆ ꢄꢅꢉꢆ ꢀ
> >
Note. The maximum errors estimated by the SEDEM software (20) are *R"$0.02 A, *N"$1, *(pꢀ)"$10\ꢃ Aꢀ, *(*E )"$1 eV.
ꢄ

1
52
THOMAZEAU ET AL.
FIG. 6. Fourier-"ltered EXAFS kꢃs(k) spectrum (diamonds) of the WꢄꢅꢆꢄMoꢄꢅꢆꢄ ꢀ
S sample (cocrystallized thiosalt synthesis) "tted with a model
Mo!"ꢁ}S}(Mo, W)}S (solid curve).
IV.3. W Mo
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
at 673 K from Mechanical Mixture of Ammonium
S -MM, the Compound Obtained
signi"cant amount of tungsten could be detected, and for
crystallites where tungsten is present, no signi"cant amount
of molybdenum could be detected. It means that, in contrast
to the cocrystallized thiosalts, mechanical mixture of
¹etrathiomolybdate and ¹etrathiotungstate
EXAFS study of this compound gives results signi"cantly ammonium thiomolybdate and thiotungstate leads to
di!erent from the previous ones. As shown by Fig. 8 (Mo a mixture of the MoS and WS phases and is not a precur-
ꢀ
ꢀ
K-edge), the second peak of the Fourier transform is perfect- sor that allows a solid solution to be obtained. This di!er-
ly the same for the mixed W Mo S -MM compound ence can be related to a higher segregation degree of the Mo
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
as for MoS . It means that molybdenum has only molyb- and W cations in the mechanical mixture than in the cocrys-
ꢀ
denum in-plane cationic neighbors, and that tungsten has tallized thiosalts, whatever their precise nature, solid solu-
only tungsten in-plane cationic neighbors. This observation tion or intimate mixture.
is con"rmed by the re"nement of the EXAFS spectrum
(
Table 3). It can thus be thought that this compound corres-
IV.4. The W Mo
S -DM Samples Obtained at 1073
ponds to an interlayer solid solution or to a two-phase
system. EDX analyses performed with a 15}20 nm probe-
size gave, as expected, a sulfur/metal atomic ratio close
V
(ꢂ\V) ꢀ
and 1273 K from the Disulxdes
No notable di!erence could be detected in this series
to 2, but for crystallites where molybdenum is present, no between quenched and slowly cooled samples of the same
TABLE 2
EXAFS Re5nement Results (W LIII-edge): Characteristics of the Cationic In-plane Environment of Tungsten in WS
2
and
W0.50Mo0.50
S
2
Prepared at 673 K by Treatment, under H S Flow, of the Mixed Thiosalts Resulting from Cocrystallization of
2
Equimolar Proportions of Ammonium Tetrathiotungstate and Ammonium Tetrathiomolybdate
Mo shell
W shell
pꢀ;10ꢃ (A
R(Mo) (A
>
)
N(Mo)
pꢀ;10ꢃ (A
>
ꢀ)
*Eꢄ (eV)
R(W) (A
>
)
N(W)
>
ꢀ) *Eꢄ (eV)
WS
Mo
*
3.18
*
2.8
*
1.66
*
15.4
3.15
3.16
3.9
3.5
1.7
2.0
8.8
9.9
ꢀ
W
S
ꢄꢅꢆ ꢄꢅꢆ ꢀ
Note. The maximum errors estimated by the SEDEM software (20) are *R"$0.01 A
>
, *N"$0.6, *(pꢀ)"$1.2;10\ꢃA
>
ꢀ, *(*E )"$1.4 eV.
ꢄ

MOLYBDENUM}TUNGSTEN DISULFIDE LAMELLAR SOLID SOLUTION
153
FIG. 8. Fourier transform obtained at the Mo K-edge for MoS and
ꢀ
W
Mo S -MM prepared at 673 K by treatment, under H S #ow, of
FIG. 7. Fourier-"ltered EXAFS kꢃs(k) spectrum (diamonds) of the
Mo sample (cocrystallized thiosalt synthesis) "tted with
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ ꢀ
a mechanical mixture of ammonium tetrathiotungstate and ammonium
tetrathiomolybdate taken in equimolar proportions.
W
S
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
a model W!"ꢁ}S}(Mo, W)}S (solid curve).
composition. These products are better crystallized than splitting could be observed, even for the high angle 00l lines.
those obtained from the thiosalts, as evidenced by the TEM Under these conditions, the c parameter evolution could
picture (Fig. 9), and cell parameters could be re"ned from result not from the existence of a solid solution, but from the
X-ray di!raction data. No signi"cant evolution of the use of a single-phase model for the re"nement of the di!rac-
a parameter vs. composition can be detected, which is not tion diagram of a two-phase system. At the MoK-edge, no
surprising since the a
and a
51
identical. As for the c parameter, it is found to increase with the W
parameters are nearly signi"cant di!erence can be observed between the FTs of
+
ꢀ1
ꢀ
ꢀ
Mo
S samples treated at 1073 and 1273 K,
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
the tungsten content, which could indicate the occurence of which are both similar to that of a well-crystallized MoS
a solid solution. However, despite the high synthesis tem- sample (Fig. 10), as con"rmed by the re"ned parameters (see
perature, the di!raction lines are broadened with respect to Table 3). EDX analysis gives the same results as those
ꢀ
the instrumental linewidth and c
and c
are not su$- obtained for the compound prepared from mechanical mix-
+
ꢀ1
ꢀ
51
ꢀ
ciently di!erent for the MoS and WS lines to be resolved ture of the thiosalts. It clearly means that, despite the high-
ꢀ
ꢀ
in the di!raction pattern of a two-phase sample. This was temperature treatments, these samples contain distinct
veri"ed using weighted additions of the experimental MoS and WS phases: even at temperatures as high as
ꢀ
ꢀ
di!ractograms of MoS and WS in order to simulate 1273 K, mixed [W
Mo
S ] layers cannot be obtained
ꢀ
ꢀ
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
the di!ractograms of two-phase W Mo S systems: no by reaction of preformed [MoS ] and [WS ] layers.
V
ꢂ\V ꢀ
ꢀ
ꢀ
TABLE 3
EXAFS Re5nement Results (Mo K-edge): Characteristics of the Cationic In-plane Environment of Molybdenum
Samples Prepared According to Di4erent Methods
in Several W0.50Mo0.50
S
2
Mo shell
pꢀ;10ꢃ (A
W shell
pꢀ;10ꢃ (A
R(Mo) (A
>
)
N(Mo)
>
ꢀ)
*Eꢄ (eV)
R(W) (A
>
)
N(W)
>
ꢀ) *Eꢄ (eV)
Mo
Mo
Mo
Mo
Mo
Mo
W
W
W
W
W
W
S -673
ꢄꢅꢆ ꢄꢅꢆ ꢀ
S -1073
ꢄꢅꢆ ꢄꢅꢆ ꢀ
S -MM
ꢄꢅꢆ ꢄꢅꢆ ꢀ
S -DM-1073
ꢄꢅꢆ ꢄꢅꢆ ꢀ
S -DM-1273
ꢄꢅꢆ ꢄꢅꢆ ꢀ
S -DS
ꢄꢅꢆ ꢄꢅꢆ ꢀ
3.16
3.14
3.16
3.15
3.13
3.14
3.2
2.9
5.4
6.4
5.6
3.3
4.5
1.2
3.1
2.3
3.2
1.0
2.4
!0.3
2.5
2.5
!5.0
!7.2
3.17
3.13
*
*
*
2.1
3.4
*
*
*
4.7
3.0
*
*
*
6.9
!1.6
*
*
*
3.14
2.7
0.4
!6.4
Note. W
Mo
S -673 and W
Mo
S -1073 were prepared by treatment of the mixed thiosalt resulting from cocrystallization of equimolar
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
proportions of ammonium tetrathiotungstate and ammonium tetrathiomolybdate, under H S #ow, at 673 and 1073 K respectively. W
ꢄꢅꢆꢄ
was prepared in a similar way at 673 K, but starting from a mechanical mixture of ammonium tetrathiotungstate and ammonium tetrathiomolybdate.
Mo
S -MM
ꢀ
ꢄꢅꢆꢄ ꢀ
W
W
Mo
Mo
S -DM-1073 and
W
Mo
S -DM-1273 resulted from heating WS #MoSꢀ mixtures at 1073 and 1273 K, respectively.
ꢄꢅꢆꢄ ꢀ
S -DS was obtained from reaction (1273 K) of sulfur on Mo#W powders.
ꢄꢅꢆꢄ ꢀ
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
ꢄꢅꢆꢄ
ꢀ
The maximum errors estimated by the SEDEM software (20) are *R"$0.01 A
>
, *N"$0.6, *(pꢀ)"$0.4;10\ꢃ A
>
ꢀ, *(*E )"$1 eV.
ꢄ

1
54
THOMAZEAU ET AL.
FIG. 10. Fourier transform obtained at the Mo K-edge for MoS
ꢀ
(
crystallized at 1073 K) and two W
ꢄꢅꢆꢄ
heating WS #MoSꢀ mixtures at 1073 and 1273 K.
Mo
S samples resulting from
ꢄꢅꢆꢄ ꢀ
ꢀ
two-cation layers of the intralayer solid solution (Table 3). In
the case of this synthesis route, it can then be thought that in
the presence of sulfur, molybdenum, and tungsten powders
FIG. 9. TEM picture of the W
ꢄꢅꢆꢄ
disul"des at 1273 K (G"300,000).
Mo
S sample prepared from the
ꢄꢅꢆꢄ ꢀ
directly react to form the mixed [W Mo S ] layers.
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
V. CONCLUSION
IV.5. The W
ꢄꢅꢆꢄ ꢄꢅꢆꢄ ꢀ
by Direct Synthesis (=#Mo#S)
Mo
S -DS Sample Obtained at 1273 K
It must "rst be underlined that, using various synthesis
conditions, we never observed formation of an interlayer
As for the W Mo S -DM products obtained from solid solution. Even if the [MoS ] and [WS ] layers present
V
ꢂ\V ꢀ
ꢀ
ꢀ
MoS #WS , X-ray di!raction cannot determine the na- quasi-identical in-plane periodicities, the intralayer solid
ꢀ
ꢀ
ture of the W
Mo
S -DS sample, whether it is a solid solution appears more stable, as in the case of the Nb}Mo
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ
solution or mixture of WS and MoS . EXAFS measure- disul"de system. The geometrical factor cannot thus be
ments at the Mo K-edge make it possible to answer this considered as responsible for the obtention of the two-ca-
ꢀ
ꢀ
question since the cationic peak of the Fourier transform tion [M?M@
S ] layers. Band structure calculations will
Fig. 11) is clearly split, which is the signature of the be performed in order to identify the origin of the better
V
(ꢂ\V) ꢀ
(
FIG. 11. Fourier transform obtained at the Mo K-edge for the W
Mo
S sample prepared by direct synthesis.
ꢄꢅꢆꢄ
ꢄꢅꢆꢄ ꢀ

MOLYBDENUM}TUNGSTEN DISULFIDE LAMELLAR SOLID SOLUTION
155
stability of the intralayer model. Concerning now the in-
tralayer solid solution, it appears that the synthesis temper-
ature is not a decisive factor for its obtention. At
temperatures as low as 673 K and starting from the cocrys-
3. I. L. Singer, in &&Fundamentals in Friction: Macroscopic and
Microscopic Processes'' (I. L. Singer and H. M. Pollock, Eds.). Kluwer
Academic, Dordrecht, 1992.
4
. A. S. Golub, G. A. Protzenko, I. B. Shumilova, Ya. V. Zubavichus,
C. Payen, Yu. N. Novikov, and M. Danot, Mol. Cryst. ¸iq. Cryst., 311,
tallized thiosalts, the two-cation [W Mo
S ] layers of
3
77 (1998).
V
(ꢂ\V) ꢀ
the intralayer solid solution are formed and are still stable at
least up to 1073 K. However, up to 1273 K, no reaction
between MoS and WS , i.e., between already-existing
5. F. Kadijk, Thesis, University of Groningen, 1969.
6
7
. V. Gaborit, Thesis, University of Nantes, 1998.
. F. Wypych, K. Sollmann, and R. Schollhorn, Mater. Res. Bull. 27,
545 (1992).
K
ꢀ
ꢀ
[
MoS ] and [WS ] layers, occurs. Starting from mechan-
8
9
. B. K. Miremadi and S. R. Morrison, J. Appl. Phys. 67, 1515 (1990).
. C. Calais, N. Matsubayashi, C. Geantet, Y. Yoshimura, H. Shimada,
A. Nishijima, M. Lacroix, and M. Breysse, J. Catal. 174, 130 (1998).
ꢀ
ꢀ
ical mixtures of the thiosalts does not lead to a solid solution
either. These results show that the nature of the precursor is
the decisive factor, and that the solid solution can be pre-
pared only with a precursor in which the two cations are
closely associated.
1
0. W. J. Schutte, J. L. de Boer, and F. Jellinek, J. Solid State Chem. 70,
07 (1987).
2
1
1
1. R. G. Dickinson and L. Pauling, J. Am. Chem. Soc. 45, 1466 (1923).
2. K. D. Bronsema, J. L. de Boer, and F. Jellinek, Z. Anorg. Allgem. Chem.
Lastly, this work demonstrates the power of EXAFS to
5
40, 15 (1986).
characterize the structure of random 2-D compounds, 13. JCPDS "les, 08-0237 and 37-1492.
1
4. M. Breysse, R. Fre
ett. 26, 97 (1984).
5. E. Diemann and A. MuK ller, Coord. Chem. Rev. 10, 79 (1973).
H ty, M. Lacroix, and M. Vrinat, React. Kinet. Catal.
something which would be practically impossible to do by
di!raction methods.
¸
1
1
1
6. K. Ramanathan and S. W. Weller, J. Catal. 95, 249 (1985).
7. G. Alonso, M. D. Valle, J. Cruz, V. Petranovskii, A. Licea-Claverie,
and S. Fuentes, Catal. ¹oday 43, 117 (1998).
REFERENCES
1
8. P. Belougne, N. Che
32, 3087 (1976).
19. K. Sasvari, Acta Crystallogr. 16, 719}724 (1963).
H zeau, and J. Lapasset, Acta Crystallogr. Sect. B
1
. F. H uK lliger, in &&Structural Chemistry of Layered-Type Phases''
(
F. Levy, Ed.). D. Reidel, Dordrecht, 1976.
H
2
. H. Topsoe, B. S. Clausen, and F. E. Massoth, in &&Catalysis Science and 20. D. Aberdam, J. Synchrotron Rad. 5, 1287}1297 (1998).
Technology, Vol. 11, Hydrotreating Catalysis'' (J. R. Anderson and 21. J. J. Rehr, S. I. Zabinsky, and R. C. Albers, Phys. Rev. ¸ett. 69, 3397
M. Boudart, Eds.). Springer-Verlag, Berlin, 1996.
(1992).