High-temperature solid-vapor and liquid-vapor transitions in binary and ternary chalcogenides La2S3, MoS2, Mo2S3 and LiInSe2

Article abstract of DOI:10.1016/j.jallcom.2007.01.175

High heating rate thermo-microscopic analysis, analytical atomic-emission technique with inductively coupled plasma and differential dissolution method have been used to investigate solid-liquid and solid-liquid-vapor transitions of refractory chalcogenides La₂S₃, MoS₂, Mo₂S₃ and LiInSe₂. The techniques were effective in determining the total pressure and vapor phase composition over the compounds in large temperature and pressure ranges. As a result, for compounds that evaporate incongruently, the melting points corresponding to their strict stoichiometry, the decomposition pressure-temperature relations, and other new results of the T-x phase diagrams for the La₃S₄-La₂S₃, Mo-MoS₂, and Li₂Se-LiInSe₂ systems were obtained.

Full text of DOI:10.1016/j.jallcom.2007.01.175

Journal of Alloys and Compounds 452 (2008) 89–93
High-temperature solid–vapor and liquid–vapor transitions in binary
and ternary chalcogenides La2S3, MoS2, Mo2S3 and LiInSe2
I.G. Vasilyeva∗, R.E. Nikolaev
Nikolaev Institute of Inorganic Chemistry, Russian Academy of Sciences, Siberian Branch, Lavrentiev Avenue 3, 630090 Novosibirsk, Russia
Received 18 September 2006; received in revised form 28 December 2006; accepted 10 January 2007
Available online 25 February 2007
Abstract
High heating rate thermo-microscopic analysis, analytical atomic-emission technique with inductively coupled plasma and differential dissolution
method have been used to investigate solid–liquid and solid–liquid–vapor transitions of refractory chalcogenides La2S3, MoS2, Mo2S3 and LiInSe2.
The techniques were effective in determining the total pressure and vapor phase composition over the compounds in large temperature and pressure
ranges. As a result, for compounds that evaporate incongruently, the melting points corresponding to their strict stoichiometry, the decomposition
pressure–temperature relations, and other new results of the T–x phase diagrams for the La3S4–La2S3, Mo–MoS2, and Li2Se–LiInSe2 systems were
obtained.
© 2007 Elsevier B.V. All rights reserved.
Keywords: High-temperature alloys; Phase transitions; Thermal analysis
1. Introduction
Mo2S3 and LiInSe2 compounds which are materials of impor-
tant technical applications. This technique allowed to observe
and indicate with high reliability the origin of thermal peaks
and to determine quantitatively the pressure and composition of
the vapor–solid phases at melting and boiling temperature. High
heating rates were not hindrance to the study of equilibrium
processes because a few minutes were shown to be sufficient
to attain vapor–liquid–solid equilibrium at high-temperatures
[1]. As a result, decomposition pressure–temperature relations
and melting points corresponding to strict stoichiometry of
these compounds evaporating incongruently were obtained as
well as some new results for the T–x phase diagrams of the
La3S4–La2S3, Mo–MoS2, and Li2Se–LiInSe2 systems.
Progress in the study of T–x phase diagrams of sulfide sys-
tems with refractory compounds decomposing before melting
was demonstrated in [1], where the melting points of compounds
were measured under conditions reducing or preventing evap-
oration losses of samples over a heating period due to special
apparatuses and procedures. However, the failure of these tech-
niques to measure vapor pressures of decomposing compounds
required the development of new methods. A special thermo-
microscopic apparatus with rapid heating, high pressures of
an inert gas and specific recording systems were proposed in
[2]. While the application of the technique to determine the
melting points of decomposing compounds was demonstrated
in [2,3], other possibilities including precise measurements of
vapor pressure still remained to be solved. In the present work,
this technique combined with a precise chemical and phase
identification of the products (vapor condensate and remain-
der) quenched from the temperature of any peak in the thermal
curves was used. It allowed both melting and evaporation pro-
cesses to be investigated in detail for refractory La2S3, MoS2,
2. Experimental
The apparatus for measuring the parameters of melting and evaporation pro-
cesses is shown in Fig. 1. The specific recording system of the apparatus was
sensitive to the detection of all types of transitions (liquid–solid, solid–vapor and
liquid–vapor) due to a suitable IR (infra red) photodiode which was mounted
on a microscope and measured the emitted thermal radiation of the sample as
a function of temperature. The transitions appeared as marked peaks on the
thermal radiation curves recorded in the dU/dτ –T coordinates, where dU/dτ
is the time derivative of sample radiation. A small Mo crucible with about
1 mg of the sample was placed inside a tungsten heater of special geometry
in direct contact with a W/W–Re (20%) thermocouple. This assembly was put
into a water-cooled chamber filled with helium as a buffer gas. The helium
∗
Corresponding author. Tel.: +7 383 330 84 65; fax: +7 383 330 94 89.
E-mail address: kamarz@che.nsk.su (I.G. Vasilyeva).
0925-8388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2007.01.175

90
I.G. Vasilyeva, R.E. Nikolaev / Journal of Alloys and Compounds 452 (2008) 89–93
linearly heated to the boiling temperature at a given helium pressure. In this
case, the boiling peak occurring first in the thermal curves showed that the
vapor pressure of the sample was equal to the helium pressure in the chamber
(PHe = Pvapor). In these experiments, helium pressures varied usually between
0.3 and 2 bar and the partial pressure of the compound tested was considered
to be the same as in the absence of helium based on the assumption that the
system with helium was ideal (low pressures and no solubility of helium in the
condensed phase). Therefore, the dependences of the boiling temperatures on
the helium pressure were conceptually identical to the pvapor–T relation of the
initial stoichiometric sample.
The DD technique as a chemical method of phase analysis was used here
for the identification of quenched melts which were amorphous as could be
expected. Themethodwasdescribedearlieranddemonstratedonmanyexamples
[7–10]. A sample is dissolved in a flow of solvent with continuously increasing
concentration and temperature. Under these conditions, different phases of a
multiphase sample dissolve sequentially but not simultaneously. The elemental
composition of the solution is monitored by an ISP AE spectrometer and the
dissolution curves of all elements of the sample are recorded. From such kinetic
curves, stoichiograms may be calculated which are the time dependences of
the molar ratios between each two elements, for example, A:B, B:C and A:C
for a three-component ABC sample. The time profile of the stoichiograms is
governed by two main rules: stoichiograms remain constant with values equal
to the stoichiometric coefficient ratios of the dissolved solid phase if only a single
phase is dissolved. The stoichiograms are variable if several phases are dissolved
at the same time. A stoichiogram indicates directly whether the mixture is really
separated in individual phases, without the necessity of using reference materials
for these phases. Therefore, the DD method is a combination of two procedures:
separation of phases and measurement of their stoichiometry. 7N HNO3 with
a temperature varying from 20 to 80 ◦C was used as a solvent to analyze the
quenched melts of LiInSe2 samples.
Fig. 1. Schematic diagram of the apparatus used for thermo-microscopic
analysis with high heating rates: (1) sample, (2) Mo crucible, (3) W/W–Re
thermocouple, (4) tungsten heater, (5) chamber, (6) IR photo diode, (7) quartz
window.
3. Results and discussion
pressure in the chamber was set for any given value in the range 0.01–3 bar.
Heating to 2500 K was performed at a constant rate in the range of 1–50 K/s,
depending on the material and the task of the investigation. The apparatus was
calibrated against the melting points of Au (1337 K), Co (1766 K), Pt (2045 K),
Rh (2236 K), Al2O3 (2323 K) and the decomposition pressure of GaAs crys-
tals equal to 1.0 0.03 bar at 1883 K (the melting point). The accuracy of the
measurements was about 1% for melting points and around 5% for the vapor
pressure.
The thermal peaks were identified by microscopic observation of partial
melting (solidus point), complete melting (liquidus point) and marked surface
change (boiling point). The peak indication was also confirmed by a special
quenching experiment: a sample heated to the temperature of the corresponding
peak was quenched at a rate of 200 K/s and the chemical and phase compositions
of the quenched products (vapor condensed on the window and the residue) were
determined by atomic-emission technique with inductively coupled plasma (ICP
AES), X-ray diffraction (XRD) or differential dissolution technique (DD).
By varying the heating rates and helium pressure in the chamber, the T–x
and pvapor–T diagrams could be obtained starting with a sample of defined com-
position such as La2S3, MoS2, Mo2S3 and LiInSe2. Taking the highest heating
rates and helium pressures (PHe ꢀ Pvapor), we can force the sample to boil at
temperatures past the melting point. The increase of the evaporation rate and the
vapor diffusion from the surface to the condenser at and above the boiling point
led to the variation of the sample composition. On the contrary, if the melting
point peak appeared first in the thermal curves, it signified that the vapor losses
were prevented (or reduced) kinetically up to the melting of the sample. The fact
that the composition of the samples quenched from the melting point remained
really unchanged, was supported by direct chemical analysis. The deviations
3.1. La–S system
The problem of correct determination of the melting point
and the saturated sulfur vapor pressure in the range 0–60 at.% of
sulfur arises only for La2S3 which decomposes before melting.
Due to the above-mentioned technique, the melting point of sto-
ichiometric La2S3 was found to be (2133 15) K (Fig. 2, curve
1). The pvapor–T relation in the temperature range 1853–2013 K
(Fig. 2, curves 2–4) could be presented by the following linear
equation:
log psulfur(bar) = (6.29 0.16) − (12660 310)T−1
(1)
Both the melting point and the pvapor–T relation are really cor-
related with the initial stoichiometry since the composition was
found by means of a special high-accurate gas-chromatographic
technique [11] to be La2S2.994(3). Following Eq. (1), congruent
melting of stoichiometric La2S3 takes place under the equi-
librium sulfur vapor pressure of 2.26 bar. The melting peaks
appearing after the boiling peaks (Fig. 2, curves 2–4) indicated
the melting of samples with compositions different from the
starting composition of La2S3 owing to the increase of the evap-
oration rate at and above the boiling point of La2S3. The shift
towards the less-volatile component resulted in the composi-
tions La2S3.0−x (x = 0.02, 0.04, and 0.1), which together with
the melting temperatures being higher than that of the stoichio-
metric La2S3, reproduce the liquidus curve of the T–x diagram
of the La2S3–La3S4 system. The melting point of La2S3 found
here as 2133 K was the lowest among those reported previ-
were always no more than 0.5 mass%. At low helium pressures (PHe < Pvapor
)
and low heating rates, the heterogeneity of the initial phases occurred for the
phases existing over little or no stoichiometry range as their composition steadily
changed (with a given step) owing to incongruent vaporization within a closed
chamber. The composition/temperature values corresponding to these variations
reflected, in this case, the form of the solidus–liquidus lines of the T–x diagram.
The boiling point method (isobaric variant), well-known from [4–6], was
realized to determine the vapor pressure of the samples. The samples were

I.G. Vasilyeva, R.E. Nikolaev / Journal of Alloys and Compounds 452 (2008) 89–93
91
ously. The discrepancy is explained by the uncertainty of the
composition/temperature correlations. In previous experiments,
evaporation losses of La2S3 were not prevented and the end
composition was not controlled.
Both the melting point and the pvapor–T relation determined
here are the key parameters providing strict stoichiometry of
La2S3 during its high-temperature preparations as large-sized
crystals and ceramics. Only stoichiometric La2S3 in the range
0.5–16 ␮ shows transparency high enough making the material
a good candidate for IR optic atmospheric windows.
3.2. MoS2–Mo system
The Mo–S system was subjected to repeated investigation
[1], but still remains not completely established. New results
were obtained by our technique in the range of Mo–MoS2 start-
ing with only two samples of stoichiometric Mo2S3 and MoS2.
Two kinds of thermal curves at high (17 and 50 K/s) and low (1
and 5 K/s) heating rates were recorded for MoS2 and Mo2S3
(Fig. 3a). At high rates, solid MoS2 began to decompose at
1853 K under pvapor = 1 bar to Mo2S3 and sulfur gas. Since this
reaction proceeded very rapidly, only one half of MoS2 decom-
posed resulting in the MoS2 + Mo2S3 mixture (1:1) with eutectic
and liquidus temperatures of 1953 and 2053 K, respectively
(Fig. 3a, curve 1). At low rates, the evaporation losses increased
markedly and MoS2 was converted into the Mo2S3 + Mo mix-
tures composed of variable amounts of these phases identified
chemically and structurally (Fig. 3a, curve 3). All the mixtures
showed an eutectic temperature of 1893 K and liquidus points
Fig. 2. ProfileofheatingcurvesofLa2S3 recordedatarateof17 K/sanddifferent
helium pressures: 3 bar (curve 1), 1 bar (curve 2), 0.6 bar (curve 3), 0.3 bar (curve
4). MP and BP symbols are melting and boiling temperatures.
Fig. 3. (a) Profile of heating curves of MoS2 and Mo2S3 recorded at a helium pressure of 1 bar and different heating rates: MoS2 (17 K/s, curve 1), Mo2S3 (30 K/s,
curve 2), MoS2 (1 K/s, curve 3), Mo2S3 (5 K/s, curve 4) and (b) the T–x diagram of the Mo–MoS2 system.

92
I.G. Vasilyeva, R.E. Nikolaev / Journal of Alloys and Compounds 452 (2008) 89–93
Table 1
Melting and evaporation parameters for MoS2 and Mo2S3
Parameter
Earlier works
This worka
Ppart = f(T) for MoS2
The flow method followed by only Mo analysis in vapor
[14]: 10−5–10−4 bar at 1523–1738 K
<1973 K, 2013 K, >2273 K [1]
Boiling points method: 1.0 0.1 bar
at (1853 15) K for solid MoS2
(2053 10) K
MP of Mo2S3
Melting behavior of Mo2S3
Incongruent melting, disproportion or dissociation [1]
Congruent melting under equilibrium
sulfur vapor pressure
Eutectic temperature of MoS2 + Mo2S3 mixtures
Eutectic temperature of Mo + Mo2S3mixtures
No data
(2226 15) K
(1888 15) K
1563 K, (1773 100) K, ∼1733 K, (1883 15) K [1]
a
Partial and full melting of samples and the appearance of solid Mo in the melts were also revealed microscopically.
between 1923 and 2053 K depending on the total composition
of the mixture (Fig. 3b). Contrary to MoS2, Mo2S3 heated at a
rate of 30 K/s melted congruently without marked loss of sulfur
(Fig. 3a, curve 2) which was not attained in previous experi-
ments [1]. Heating at low rates varied the composition Mo2S3
to a Mo2S3 + Mo mixture as verified by XRD and chemical
examination of the quenched products. The different mixtures
showed eutectic temperature of 1893 K and liquidus tempera-
turesbetween1900and1980 K, asinthecaseofmixturesformed
by decomposition of MoS2 (Fig. 3a, curve 4). All the results
are summarized in the T–x phase diagram shown in Fig. 3b,
where the dashed lines relate to data taken from [1]. In Table 1,
our results are compared with literature data for the MoS2–Mo
system. They demonstrate clearly the effect of evaporation on
stoichiometry and the heterogeneity of heated samples of Mo2S3
and MoS2 which is important for a correct representation of the
T–x diagram.
To support the thermal experiment that revealed the coexis-
tence of two liquids in the Li2Se-rich range, a DD analysis was
performed for melts quenched from temperatures slightly higher
than the melting, monotectic and liquidus points, respectively
(Fig. 5). The changes in composition follow immediately from
the profiles of both the dissolution curves of Li, In, Se elements
and the molar Li/In ratio. The single-phase three-component
species dissolved as shown in Fig. 5a and c: here the kinetic
3.3. Li2Se–LiInSe2 system
LiInSe2 was included in this study as an object for which
all transitions appeared in the thermal curves above the melt-
ing point. Since both its vapor and melt were very aggressive
toward any container material, the high heating rate technique
was useful to study the melting and evaporation processes due
to the lower time of melt-crucible contact. A typical thermal
curve (profile depended only slightly on helium pressure and
heating rate) is depicted in Fig. 4a. The origin of the peaks lying
after the melting peak at 1169 K was clarified by the quenching
procedure and chemical/phase identification of the quenched
products. This experiment showed the preferential transition of
the In2Se3 component into the vapor phase and indicated the
incongruent evaporation of LiInSe2 melts. The shift of the ini-
tial composition to the Li2Se side led to the occurrence of the
liquid immiscibility characterized by a monotectic temperature
of 1213 10 K. The peak at 1316 K (at 1323 K on another curve
not shown here) occurred when the inhomogeneous melt turned
back into the homogeneous state (Fig. 4b). The boiling peaks
being the last in the thermal curve appeared at temperatures
between 1397 and 1446 K when the helium pressure in the cham-
ber changed from 0.1 to 0.3 bar. If the results are presented as
log pvapor−1/T and the line is extrapolated to the melting point
of LiInSe2 at 1169 K, the pressure of saturated vapor over the
melt would be about 10−5 bar.
Fig. 4. (a) Profile of heating curves of LiInSe2 recorded at a rate of 17 K/s and
helium pressure 0.3 bar and (b) The T–x diagram of the Li2Se–LiInSe2 system.

I.G. Vasilyeva, R.E. Nikolaev / Journal of Alloys and Compounds 452 (2008) 89–93
93
here may lead to the occurrence of non-stoichiometry and unde-
sirable Li2Se precipitates which are the scattering centers of
LiInSe2 crystals grown from the overheated melts [12]. This
seems not to fulfill the basic requirements providing practical
applications of LiInSe2 as a promising nonlinear optic material
[13].
4. Conclusions
A rapid heating thermal technique combined with analyti-
cal methods was applied to La2S3, Mo2S3, MoS2 and LiInSe2
which were evaporated incongruently to study the details of
the melting and evaporation processes. The unique ability of
the technique to measure high vapor pressures (up to 3 bar) at
high-temperatures (up to 2500 K), to determine correctly the
high melting points of decomposing compounds while pre-
serving their initial stoichiometry, and to study transitions in
overheated melts is demonstrated. The rapid heating technique
combined with quenching experiments followed by analyses of
the quenched products is a very promising method for investi-
gating solid–liquid–vapor and liquid–liquid–vapor equilibria of
compounds with a rather high own vapor pressure. This moti-
vates further development of the technique for application to
other materials.
Acknowledgement
This work was supported partly by RBFR grant nos. 05-03-
32761 and 05-03-32236.
References
[1] G. Moh, Top. Curr. Chem. 76 (1978) 108–125.
Fig. 5. Kinetic curves of Li (1), In (2), Se (3) dissolution and molar Li/In ratios
(4) as functions of dissolution time for LiInSe2 melts quenched from melting
(a), monotectic (b), and (c) boiling temperature.
[2] Ya. Gibner, I. Vasilyeva, J. Therm. Anal. 53 (1998) 151–160.
[3] I. Vasilyeva, E. Belyaeva, Ja. Gibner, J. Therm. Anal. 52 (1998) 403–412.
[4] T. Sato, H. Kaneko, Technol. Repts. Tohoku Univ. 16 (1952) 18–33.
[5] D.T. Ulrich, Naturwiss-phil. FFH, Diss. Braunschweig, 1954.
[6] V. Glasov, A. Pashinkin, A. Malkova, M. Pasternak, Neorg. Mater. 15
(1979) 408–411.
[7] V. Malakhov, J. Mol. Catal. A 158 (2000) 143–148.
[8] I. Vasilyeva, A. Vlasov, V. Malakhov, M. Predtechensky, Thin Solid Films
292 (1997) 85–90.
[9] I. Vasilyeva, J. Alloy Compd. 323/324 (2001) 34–38.
[10] A. Prokofiev, I. Vasilyeva, V. Ikorskii, V. Malakhov, I. Asanov, W. Assmus,
J. Solid State Chem. 177 (2004) 3131–3139.
[11] L. Chuchalina, I. Vasilyeva, A. Kamarzin, V. Sokolov, Zh. Analyt. Khimii,
(Russ.) 33 (1978) 190–192.
curves of all the elements go synchronically and the Li/In ratio
is almost constant and close to unity similar to stoichiometric
LiInSe2. Dissolution of two phases is demonstrated in Fig. 5b:
the small peak appearing at 0.5 min corresponds to dissolution
of a binary Li–Se species. Another three-component species
with synchronic kinetic curves dissolves later and its averaged
Li/In ratio is also close to unity. This means that two liquids
with compositions Li2Se and close to LiInSe2 are present in
melts quenched from the monotectic temperature. This finding
agrees well with the form of solidus–liquidus lines obtained in
the thermal experiment and shown in Fig. 4b. A more detailed
study of the behavior of overheated LiInSe2 melts will be pub-
lished later. However, it is clear even now that the incongruent
evaporation of LiInSe2 and the immiscibility region detected
[12] L. Isaenko, I. Vasilyeva, A. Merkulov, A. Yelisseyev, S. Lobanov, J. Cryst.
Growth 275 (2005) 217–223.
[13] D. Nikogosyan, Newly Developed and Perspective Crystals, Springer Sci-
ences/Business Media, Inc., Printed in USA, 2005, p. 267.
[14] R. Isakova, Izv. Acad. Nauk Kaz. SSR (Russ.) 3 (1961) 3–7.