Chemical route for formation of intermetallic Zn4Sb3 phase

Article abstract of DOI:10.1016/j.jssc.2010.03.024

Synthesis of intermetallic zinc antimonide phases via low temperature solution route was investigated. Trial experiments were carried out under inert atmosphere at 70 °C using metallic Zn, SbCl₃ and NaBH₄ as reactants and tetrahydrofuran (THF), dimethylsulfoxide (DMSO) as organic media. Powder X-ray analysis confirmed the nucleation and growth of ZnSb phases in presence of excess Zn. SEM analysis revealed the existence of core-shell structure comprising of Zn core and Sb shell. Such particles get transformed into Zn₄Sb₃ crystalline phases upon thermal treatment at 300 °C/6 h in a silica tube closed under high secondary vacuum.

Full text of DOI:10.1016/j.jssc.2010.03.024

ARTICLE IN PRESS
Journal of Solid State Chemistry 183 (2010) 1090–1094
Contents lists available at ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Chemical route for formation of intermetallic Zn Sb phase
4
3
1
Ã
A. Denoix, A. Solaiappan , R.M. Ayral , F. Rouessac, J.C. Tedenac
Institut Charles Gerhardt Montpellier UMR 5253 (ICGM), C2M-UM2-CNRS, Pl. E. Bataillon, 34095 Montpellier Cedex 5, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Synthesis of intermetallic zinc antimonide phases via low temperature solution route was investigated.
Trial experiments were carried out under inert atmosphere at 70 1C using metallic Zn, SbCl3 and NaBH4
as reactants and tetrahydrofuran (THF), dimethylsulfoxide (DMSO) as organic media. Powder X-ray
analysis confirmed the nucleation and growth of ZnSb phases in presence of excess Zn. SEM analysis
revealed the existence of core–shell structure comprising of Zn core and Sb shell. Such particles get
Received 3 December 2009
Received in revised form
1
1 March 2010
Accepted 14 March 2010
Available online 18 March 2010
transformed into Zn
4 3
Sb crystalline phases upon thermal treatment at 300 1C/6 h in a silica tube closed
Keywords:
under high secondary vacuum.
Solution synthesis
Zinc antimonide
Intermetallic
&
2010 Elsevier Inc. All rights reserved.
Solvent media
1
. Introduction
Conversion of waste heat directly into useful electrical energy
4 3
attention is given for growing b-Zn Sb due to its attractive low
ꢁ
3
ꢁ1 ꢁ1
thermal conductivity (k200ꢁ400 1C¼ ꢀ6 ꢂ 10 W cm
K
) and
density (6.07).
and vice versa is achieved by thermoelectric generators (TEG).
They make use of ‘Seebeck effect’ in semiconducting materials.
Development of reliable, low temperature operating TEG is
becoming necessary today for converting the waste heat arising
out of huge numbers of personal computers, high speed trains and
automobiles and even from the domestic cooking ovens and hot
Zinc antimonide phases are synthesized, usually by melting
the high purity metallic Zn and Sb powders taken in the mole
ratio of 53:47 with 2–5 at% excess Zn, in an evacuated quartz
ampoule followed by granulation and hot pressing [4,5].
Evaporation of molten zinc and antimony, in homogeneous
microstructure, enlarged grain size, formation of undesirable
plates [1]. In a typical TEG device, the conversion efficiency (
Z
TE) is
phases (orthorhombic
Zn Sb and cracking due to thermal expansion mismatch are the
problems realized in fabrication of Zn Sb materials. Processing
methods such as direct hot pressing [6], mechanical alloying [7],
pulse discharge sintering [8], sinter forging [9], spark plasma
sintering [10], gradient freeze technique [11] and conventional
solid-state ceramic route [12] followed by repeated cold rolling
and annealing have been reported over the years.
4 3
b-ZnSb and d-Zn Sb ), decomposition of
proportional to a quantity called ‘figure of merit’ (ZT). For all the
4
3
2
thermoelectric materials, ZT is theoretically defined as ZT¼(a s/k)T,
4
3
where
and
s
and k are respective electrical and thermal conductivities
is the Seebeck coefficient. It is mandatory for TEGs to have low
a
thermal and high electrical conductivities for achieving high heat
conversion efficiency. A target value of ZT¼4 is recommended for
low temperature operating TEGs [2].
Zinc antimonides are classical thermoelectric materials with
4 3
Preparation of nano-Zn Sb is emerging because it provides large
the figure of merit value ZT¼ ꢀ1.3 in the intermediate
numbers of active grain boundaries for phonon scattering and hence,
a decrease in the bulk thermal conductivity [13]. Wet chemical
routes are better known for their efficiency and easiness for
controlling the particle size and morphology. To our knowledge,
bulk synthesis of Zn–Sb intermetallic particles through chemical
methods has not been reported. Only one study can be found in the
literature: in that one, synthesis of nanostructured zinc antimonide
was attempted by solution route [14]. First, the activated Zn and Sb
nano particles were separately prepared from the respective
chloride salts through redox reactions. Later, the chemically derived
Zn, Sb nano particles were mixed and then heated to 200 1C under Ar
atmosphere for prolonged hours to obtain the intermetallic Zn–Sb
phases. We have recently initiated wet chemical synthesis of zinc
o
temperatures (o400 C). The ZnSb and Zn
4 3
Sb are the potential
zinc antimonide phases present in the Zn–Sb binary alloys system
3]. Three distinguished phases are identified in Zn–Sb system:
[
1
ZnSb decomposing peritectically at 546 C and Zn
congruently at 563 C. The stable ZnSb and Zn Sb phases have
been formed in between 150 and 494 C. Presently, more
4
3
Sb melting
1
4 3
o
Ã
Corresponding author.
E-mail address: rose-marie.ayral@univ-montp2.fr (R.M. Ayral).
1
Permanent address: National Institute for Interdisciplinary Science and
Technology [NIIST], CSIR, Trivandrum-695019, India.
0022-4596/$ - see front matter & 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2010.03.024

ARTICLE IN PRESS
A. Denoix et al. / Journal of Solid State Chemistry 183 (2010) 1090–1094
1091
antimonides with an aim to produce colloidal zinc antimonide
dispersions. The present paper will study the synthesis of nano-
Zn and Sb metals. The given thermal energy, either externally
supplied or generated in-situ as in the case of mechanical alloying,
acts as main driving force for forming Zn, Sb solid solutions and
inter-diffusion of atoms to achieve homogeneous intermetallic phase.
Earlier studies report that once the Zn and Sb solid solutions are
occurred, the nucleation and growth of ZnSb take place at a faster
4 3
Zn Sb colloid and will show the different characterizations of this
powder, including phase stability as a function of temperature using
Raman spectroscopy.
rate. Upon continuous aging at high temperatures, the Zn
achieved. Though the ZnSb phase is formed quickly, the transforma-
tion of ZnSb into Zn Sb reaction is more sluggish even at elevated
4 3
Sb phase is
2
. Material and methods
4
3
Metallic Zn powder (99.99%, Good Fellow, USA), antimony
temperatures. The reaction kinetics is very critical and a slight
variation even in the stoichiometry as well as in thermal gradients
4
trichloride (ACROS ORGANICS, USA), NaBH (93%, TCI Europe,
Belgium), Tetrahydrofuran (THF, 99.9%, Sigma-Aldrich, GmbH),
dimethylsulfoxyde (DMSO) (Carlo erba reagents 99.5%) were used
as reagents. Acetone (99%, Sigma-Aldrich, GmbH) was used for
washing the precipitate. The Zn metal powder was stored in air
tight container and preserved under argon for preventing the
surface oxidation.
4 3
caused decomposition of Zn Sb into ZnSb and also formation of
metastable zinc antimonide phases [15].
The solution route is entirely different from the conventional
thermal processes and, therefore the involved reaction
mechanism is also quite different. In solution route, factors such
as initial molar concentration, polarity and pH of the reaction
medium, enthalpy of the reactants, and diffusivity of the solute
ions at the given reaction temperatures and time are deciding the
reactivity, nucleation and growth of the desired phases.
In our case, the complete dissolution of antimony precursor in
the given reaction medium THF, is first wetting the metallic zinc
4
In a typical synthesis, Zn powder and NaBH were first taken in a
reflux flask fitted with the condenser. The flask was filled with high
purity argon gas. The 80 mL THF were added and stirred under
3
magnetic stirring. 10.5 mmol SbCl was prepared separately in 20 mL
THF. Zn:Sb molar ratio was fixed as 65:35 with Zn in excess.
o
The reaction temperature was maintained at 75 C. After 30 min of
3 4
surface. The SbCl is attacked by both Zn metal and NaBH
stirring at this temperature, SbCl
3
solution was quickly injected. The
reductant. At this condition, the completely dissolved antimony
chloride ions interact with Zn at molecular scale and form reactive
Sb-nuclei and get deposited over the metallic Zn powders finally
forming a coating. The Zn powder acts as a substrate for the
growth of Sb structures. Such nano-Sb has high surface energy
and therefore reacts readily with Zn for forming ZnSb nuclei. In
reaction mixture suddenly turned black in color indicating the
3
+
reduction of Sb into metallic Sb according to the reaction (1):
9
NaBH
4
(sol)+3SbCl
3
(sol)-3Sb (s)+9NaCl (s)
(g)
+
9BH
3
(g)+Cl+9/2H
2
3
fact, we tested the reduction of SbCl using Zn powder without
o
The refluxing was continued at 75 C/24 h. At the end, a black
any NaBH . We have seen from the powder-X-ray analysis that
4
residue was collected and washed with excess THF followed by
acetone. The powder was then dried under primary vacuum at
there is only Sb and Zn peaks. It confirms that for accelerating the
reaction kinetics primary reductant such as NaBH is required and
4
8
0 1C for 24 h.
the excess Zn can act only as secondary reducing agent. The
The following treatments were done: first, using an other solvent
DMSO). This type of solvent allows increasing the temperature of the
solution up to 190 1C, giving the opportunity to increase the diffusion
rate inside the particles. Secondly, the obtained powder after THF
synthesis was introduced into a silica tube sealed under secondary
vacuum and heated at a temperature of 300 1C for 6h.
reaction temperature was purposefully controlled below 70 1C. In
these conditions, the obtained product was analyzed by X-ray
diffraction and only peaks belonging to Zn and Sb phases were
evidenced (Fig. 1). The first step of our procedure consisting of
reduction of SbCl3 into Sb was successful. In order to enhance
reactivity, DMSO was added to the solution which was heated to
180 1C. Fig. 2 shows XRD pattern of the resultant products using
THF medium plus DMSO at 180 1C. All the diffraction peaks can be
(
In each case, the products were characterized by X-ray
1
diffraction at 2
y
¼20–60 (XPert pro, CuK
a
radiation, scanning
step size 0.0161 and step time 40 s). The morphology and
composition of the precipitate were examined by transmission
electron microscopy (TEM, JEOL 1200EXII 120 kV) and scanning
electron microscopy (SEM, FEI Quanta 200) equipped with an
energy dispersive X-ray spectroscopy (EDS). Raman spectroscopy
indexed and correspond to metallic Sb and Zn phases and Zn Sb3
4
phase. The peaks at 2y¼28.81, 40.241 and 42.11 represent
(LabRAM-Isa – Dilor) was used to confirm results obtained by
TEM and SEM. A Linkam TS1500 heating device designed to work
from 50 to 1500 1C, adapted on the Raman spectrometer, was used
for high temperature measurements in order to study the stability
of the obtained nanopowder. The temperature was controlled by a
thermocouple at the bottom of the crucible. Measurements were
performed in argon through the silica windows from room
temperature up to 350 1C. Each time a new temperature was
reached, the sample was stabilized for a few minutes before
acquisition. The spectra at high temperature were corrected for
the thermal radiation background.
Finally, the product was characterized by Seebeck coefficient
as a function of temperature.
3
. Results and discussion
4 3
The preparation of ZnSb and Zn Sb phases through conventional
Fig. 1. X-ray diffraction pattern of the product directly obtained by the chemical
processes usually involves high reaction temperatures for melting the
route in THF medium.

ARTICLE IN PRESS
1
092
A. Denoix et al. / Journal of Solid State Chemistry 183 (2010) 1090–1094
Fig. 2. X-ray diffraction pattern of the product directly obtained by the chemical
route in THF+DMSO media.
rhombohedral Sb phase. The major Zn
300), (113) and (122) crystal planes are clearly noticed. These
results show that an increase of temperature inside the reactor by
means of the use of DMSO is determinant in formation of Zn Sb
4 3
Sb peaks corresponding to
(
4
3
phase. In order to understand the mechanism of formation
obtained by chemical route, the powders were polished and
then, observed by scanning electron microscopy. The distributions
of Sb and Zn at% from core to the shell were also examined.
The chemical compositions from surface to core are plotted
(
Figs. 3a and b). The composition of the outer shell is Sb phase rich
and when we approach towards core only a rich Zn is present.
Formation of the binary phase can occur at the interface between
zinc core and antimony shell.
The analysis of the powder by Raman spectroscopy was made
in different zones corresponding to the interface between the core
and the shell and one corresponding to the shell (Figs. 4a and b).
Three peaks are easily observable in the Raman spectrum of the
ꢁ
1
powder: the peak at 175 cm
-Zn Sb phase. Whereas, the one at 115 cm to Sb phase and
4 3
the peak at 152 cm to both Sb and Zn Sb phases. Zn phase
undoubtedly corresponds to
ꢁ
1
b
4
3
ꢁ
1
cannot be evidenced by Raman spectroscopy using this device
because the characteristic line of Zn phase is observable at
o50 cm . The powder particle is composed of a core–shell
structure in which the core is constituted by intermetallic phase
and the shell by antimony.
Fig. 3. EDX analysis and chemical composition of the Zn core and Sb shell (a curve
indicating the distribution of atomic wt% of Zn and Sb in core and shell b:
Microstructure and ED scanning positions of the particle from core to shell. Data
taken for plotting the curve).
ꢁ
1
Ageing at 300 1C for 6 h in an evacuated quartz ampoule was
carried out on the THF treated powder. XRD examination (Fig. 5)
shows the absence of Sb and/or Zn peaks as well as ZnSb ones. All
peaks in the pattern correspond to the reflections of
composed of particles not higher than a nanometer. Fig. 7
shows a typical TEM image of the powder de-agglomerated by
ultrasonicator. It can be seen from this figure that the powder is
uniform and consists of nanoparticles with average particle size of
20 nm.
The analysis of the powder by Raman spectroscopy was made
in different random zones (Fig. 8) at 25 1C. Three peaks (*) are
easily observable in the Raman spectrum of the powder: 155,
rhombohedral phase Zn
4
Sb
˚
3
which are in agreement with the
˚
cell parameters, a¼12.233 A, c¼12.428 A. The ageing at 300 1C for
6
h was good enough to give rise to an inter-diffusion of antimony
and zinc in order to elaborate the pure phase Zn Sb . XRD can be
used to evaluate peak broadening with crystallite size. Scherer
formula [16] D¼k , where D is the crystallite size, k is
the shape factor (0.9), is the wavelength of CuK radiation
is the instrumental corrected integral breadth of the reflection (in
radians) located at 2 and is the angle of reflection (in degrees)
was utilized to relate the crystallite size to the line broadening.
The instrumental corrected broadening h k l corresponding to the
4
3
ꢁ
1
ꢁ1
l
/b
h k l cos
y
175 cm
correspond to
the XRD analysis. In this spectrum, the peak at 115 cm
undoubtedly corresponds to Sb phase does not appear. We obtain
a pure intermetallic phase of -Zn Sb due to complete diffusion
and one between 320 and 330 cm , all of them
l
a
1
b
h k l
b
-Zn Sb phase. This result is in agreement with
4
3
ꢁ
1
which
y
y
b
4
3
b
of Zn and Sb during the ageing at 300 1C.
diffraction peak of Zn
4
Sb
3
was estimated by using the relation
. The average crystallite size of the
Zn Sb formation mechanism can be divided into three steps
such as wetting of Zn metal surface by molecular scale SbCl3
4
3
2
2
instrumental
1/2
b
h k l¼[(
b
h k l) ꢁ
b
]
Zn
4
Sb
3
nanoparticles was found to be 60 nm.
precursor solution, reduction of SbCl into Sb nuclei and growth of
3
SEM observations (Fig. 6) of the powder show an
agglomeration of grains of approximately 20 m, which are
Sb nuclei over the surface of Zn to form a core–shell Zn–Sb
m
particles and finally transformation into Zn Sb phase when
4
3

ARTICLE IN PRESS
A. Denoix et al. / Journal of Solid State Chemistry 183 (2010) 1090–1094
1093
Fig. 6. SEM observation of the synthesized powder.
Fig. 4. (a) Raman spectroscopy of the shell structure (Sb evidenced) and (b) Raman
spectroscopy of a particle Zn Sb phase evidenced.
4
3
Fig. 7. TEM observation of the powder.
Fig. 5. X-ray diffraction pattern of the product obtained by the chemical route in
THF media and after ageing in an evacuated quartz ampoule at 300 1C for 6 h.
DMSO used at 180 1C. We have clearly seen the formation of Zn
core and Sb shell morphology from the SEM analysis. Such
core–shell precursor’s nature is highly favorable for the
accelerated diffusion. In core–shell particles, the diffusion path
is less and therefore the mass transfer takes place even at low
reaction temperatures. In our case, annealing at 300 1C for 6 h
Fig. 8. Raman spectroscopy of the product obtained by chemical route in THF
media and after ageing in an evacuated quartz ampoule at 300 1C for 6 h at
different temperatures.
temperature 25 1C. For each temperature, the sample was
stabilized for 10 min before recording the spectrum. Very few
changes are induced by temperature up to 350 1C apart from a
progressive broadening of the bands and a weak peaks shift to left
as temperature increased. It is interesting to note that, at
4 3
gave rise to the formation of Zn Sb phase into the powder.
Fig. 8 shows Raman spectra obtained in situ at four different
temperatures: 25, 200, 350 1C and finally back to room

ARTICLE IN PRESS
1
094
A. Denoix et al. / Journal of Solid State Chemistry 183 (2010) 1090–1094
3 4
reaction using metallic zinc and SbCl as reactants and NaBH as
reducing agent in organic solvents, THF and DMSO. The process
initially yielded core–shell ZnSb particles that get transformed
4 3
into mixture Zn Sb phase upon thermal treatment at 300 1C/6 h
in a silica tube under secondary vacuum. The obtained powder
was then characterized by SEM and TEM and was composed of
nanosized-grains. Now the study in progress is to characterize it
from a thermoelectric point of view.
Acknowledgments
The authors would like to thank ANR-07-Nano-021-02’Nano-
microstil’for its financial support.
Fig. 9. Temperature dependence of Seebeck coefficient as
a function of
temperature.
References
T
4
¼25 1C, after thermal treatment, only vibrations corresponding
to Zn Sb phase are evidenced. This result evidences the thermal
stability of nano-Zn Sb phase until 350 1C.
[
1] A. Weidenkaff, R. Robert, M. Aguirre, L. Bocher, T. Lippert, S. Canules,
Renewable Energy 33 (2008) 342–347.
4
3
4
3
[2] T.M. Tritt, M.A. Subramanian, MRS Bull. 31 (2006) 188.
[
[
3] J.B. Li, M.C. Record, J.C. Tedenac, J. Alloys Compd. 438 (2007) 171–177.
4] T. Caillat, J.-P. Fleurial, A. Borshchevsky, J. Phys. Chem. Solids 58 (1997)
1
Seebeck coefficient measurements as a function of tempera-
ture in the range 25–425 1C were realized on these powders
compacted under a pressure of 55.5 MPa without any sintering to
avoid grains growth with temperature. The sample is placed in a
furnace heated at 50 1C/h up to 400 1C under primary vacuum.
Fig. 9 shows the temperature dependence of Seebeck coefficient:
119–1125.
[5] T.J. Zhu, X.B. Zhao, M. Yan, S.H. Hu, T. Li, B.C. Zhou, Mater. Lett. 46 (2000)
4–48.
[
[
[8] T. Itoh, J. Shan, K. Kitagawa, J. Prop. Pow. 24 (2008) 353.
[9] S.C. Ur, P. Nash, I.H. Kima, Mater. Lett. 58 (2004) 2937–2941.
10] T. Souma, G. Nakamoto, M. Kurisu, J. Alloys Compd. 340 (2002) 275–280.
11] G. Nakamoto, N. Akai, M. Kurisu, J. Alloys Compd. 437 (2007) 151–156.
12] S.C. Ur, P. Nash, I.H. Kim, J. Alloys Compd. 361 (2003) 84–91.
13] A. Weidenkaff, R. Robert, M.H. Aguirre, L. Bocher, L. Schlapbach, Phys. Status
Solidi 6 (2007) 247–249.
4
6] S.C. Ur, I.H. Kim, P. Nash, Mater. Lett. 58 (2004) 2132–2136.
7] V. Izard, M.C. Record, J.C. Tedenac, J. Alloys Compd. 345 (2002) 257.
it increased from 60 to 170
mV/K at 325 1C. This value is similar to
[
[
[
[
the result of Zn Sb reported by Caillat et al. [4].
4
3
4
. Conclusions
[
[
14] S. Schlecht, C. Erk, M. Yosef, Inorg. Chem. 45 (2006) 1693–1697.
15] K. Ueno, A. Yamamoto, T. Noguchi, T. Inoue, S. Sodeoka, H. Takazawa, C.H. Lee,
H. Obara, J. Alloys Compd. 384 (2004) 254–260.
Synthesis of zinc antimonide intermetallic particles has been
attempted for the first time through low temperature reduction
[16] J.I. Langford, A.J.C. Wilson, J. Appl. Crystallogr. 11 (1978) 102.