Tungsten Disilicide (WSi2): Synthesis, Characterization, and Prediction of New Crystal Structures

Article abstract of DOI:10.1002/zaac.201700329

Transition metal silicides have attracted great attention due to their potential applications in microelectronics, ceramics, and the aerospace industry. In this study, experimental and theoretical investigations of tungsten based silicides were performed.

Full text of DOI:10.1002/zaac.201700329

Journal of Inorganic and General Chemistry
DOI: 10.1002/zaac.201700329
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
Tungsten Disilicide (WSi2): Synthesis, Characterization, and
Prediction of New Crystal Structures
Jelena Lukovic´,*[a,b] Dejan Zagorac,*[a,b] J. Christian Schön,*[c] Jelena Zagorac,[a,b] Dragana Jordanov,[a,b]
Tatjana Volkov-Husovic´,[d] and Branko Matovic´[a,b]
Dedicated to Professor Thomas Schleid on the Occasion of his 60th Birthday
Abstract. Transition metal silicides have attracted great attention due structure of the final synthesized compound was determined by XRPD
to their potential applications in microelectronics, ceramics, and the analysis. In addition, new modifications for WSi2 were proposed and
aerospace industry. In this study, experimental and theoretical investi- investigated using first-principles calculations within density-func-
gations of tungsten based silicides were performed. Tungsten disilicide
(WSi2) was synthesized by simple thermal treatment at 1350 °C for 4
tional theory (DFT). Both LDA and PBE calculations show excellent
agreement with experimental observations and previous calculations
h in an argon atmosphere. These optimal synthesis conditions were for the existing modifications, where available.
obtained by variation of temperatures and times of heating, and the
application. Furthermore, W5Si3 coatings exhibit great abra-
1 Introduction
sive and adhesive wear resistance.[7] Tungsten based silicides
are also used as additives, where WSi2 and W5Si3 are included
as suitable materials for mixing with e.g. WC, in order to over-
come the problem of oxidation in extreme conditions, for ap-
plications in the metal and cutting tools industries.[8,9]
With the extremely fast growth of technology and industry
in the last decades has come the constant need for materials
with improved characteristics and better performance under
extreme conditions. In the metallurgical and the chemical in-
dustry, and in advanced technologies, the materials used are
frequently subjected to extreme conditions such as exposure to
chemical and mechanical stresses, and operation at high tem-
peratures.[1,2] Tungsten based silicides, and in particular tung-
sten disilicide (WSi2) with melting point above 2160 °C and
pentatungsten trisilicide (W5Si3) with melting point above
2320 °C, respectively, appear suitable for operation in extreme
conditions. In addition, WSi2 has low electrical resistivity and
good thermal stability, and it has been investigated as a protec-
tive coating on W-based alloys because of its excellent oxid-
In the past, a variety of synthesis methods have been used
for the production of tungsten silicide. Several different ap-
proaches can be distinguished: deposition methods for forma-
tion of thin films (by sputtering),[10] by ion beam bombard-
ment,[11] or ion beam mixing with rapid thermal anneal-
ing,[12,13] by low-pressure chemical vapor deposition,[14–16]
and by cold wall deposition (employing CVD).[15,17] Kim et al.
have discussed the transformation of WSi2 into W5Si3 at ele-
vated temperatures, which has been experimentally confirmed
ation resistance.[3–6] Another area of application are integrated by Kharatyan et al.[4,18]
circuits, where one goal is to replace polycrystalline silicon
with materials with lower resistivity. Thus, refractory metal
silicides have been introduced as possible candidates for this
The experimentally known modification of tungsten di-
silicide (WSi2) crystallizes in the tetragonal space group
I4/mmm (no. 139). Furthermore, the equilibrium WSi2 tetrago-
nal modification is characterized as the MoSi2 structure type
(Strukturbericht designation C11b, and Pearson symbol
tI6).[19–21] The MoSi2 crystal structure is characteristic of the
sequential stacking of the W and Si atomic layers along the c
axis, with ten-fold coordination of tungsten by Si atoms (CN
= 10, see Figure 1a). Another experimentally observed struc-
ture of WSi2 at low temperatures and/or in thin films shows
hexagonal symmetry with space group P6222 (no. 180). In the
literature, the hexagonal modification of tungsten disilicide has
been defined as metastable with CrSi2 type of structure (Struk-
turbericht designation C40, and Pearson symbol hP9).[22–24]
The hexagonal WSi2 modification keeps the ten-fold coordina-
tion of tungsten by Si atoms, but changes the coordination
polyhedron (see Figure 1b).
* Dr. J. Lukovic´
E-Mail: jelenal@vinca.rs;
* Dr. D. Zagorac
E-Mail: dzagorac@vin.bg.ac.rs;
* Prof. Dr. J. C. Schön
E-Mail: C.Schoen@fkf.mpg.de
[a] Institute of Nuclear Sciences Vinca
Materials Science Laboratory
University of Belgrade
Belgrade, Serbia
[b] Center for the synthesis, processing and characterization of materi-
als for use in extreme conditions
Belgrade, Serbia
[c] Max Planck Institute for Solid State Research
70569 Stuttgart, Germany
[d] Faculty of Technology and Metallurgy
Department for Metallurgical Engineering
University of Belgrade
Belgrade, Serbia
Z. Anorg. Allg. Chem. 2017, 643, 2088–2094
2088
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
facturer specification, and commercial silicon powder were used as
starting materials for the synthesis of tungsten silicide powders. They
were homogenized in acetone for 24 h. Starting powders were used in
the stoichiometric ratio, while the ball-to-powder ratio was 5:1. After
homogenization and air drying, acetone evaporated. The prepared mix-
ture was thermally treated at different temperatures ranging from 1250
to 1400 °C, with an increment of 50 °C. The heating rate was
10 K·min–1. The mixture was placed in a tube furnace, and the heat
treatment was conducted in an argon flow for 4 h. After the treatment,
the furnace was cooled to room temperature.
2.1.2 XRPD Analysis
Tungsten silicide powders were characterized by X-ray powder diffrac-
tion (XRPD) with a Riguku Ultima IV diffractometer using Cu-Kα
radiation and a Ni filter. In order to derive the relevant structural pa-
rameters, experimental data for Rietveld refinement were taken at a
speed of 2° per min in a range of diffraction angles 2θ of 5–80°, with
an angular resolution of 0.02°. Structural analysis was carried out using
Rietveld refinement and the program FullProf.[26]
2.2 Computational Methods
The full geometry optimizations for each proposed modification was
performed using the CRYSTAL14 code,[27,28] which is based on a Lin-
ear Combination of Atomic Orbitals (LCAO). For the ab initio calcula-
tions, a [5s4p1d] all-electron basis set (AEBS) was used for silicon,
labeled Si_86-311G**_pascale_2005.[29] For tungsten, we have used
the [4s4p2d] pseudopotential, labeled ECP60MWB.[30] The local opti-
mizations employed analytical gradients with respect to the atom posi-
tions,[31] and the cell parameters,[32] and a local optimization rou-
tine.[33] These local optimizations were performed using Density Func-
tional Theory (DFT). It is reasonable to choose at least two different
ab initio methods, in order to gain better insight in the quantitative
validity of the results.[34–37] Thus we have used the Local Density
Approximation (LDA) with Perdew-Zunger (PZ) correlation function-
als,[38] and the Generalized Gradient Approximation (GGA) with the
Perdew-Burke-Ernzerhof (PBE) functional, for comparison.[39] The
E(V) curves were calculated using constant volume optimization on a
point-by-point basis, by relaxing the crystalline structure at a sequence
of fixed cell volume values, where the shape of the cell was allowed
to be varied. Reference geometry to classify integrals was used with
the keyword FIXINDEX in order to obtain a smooth curve, finally
subjected to polynomial curve fitting.[27,28]
Figure 1. Visualization of the experimentally known WSi2 structures
(using the VESTA program): (a) equilibrium WSi2 modification in the
tetragonal MoSi2 structure type; (b) metastable WSi2 modification with
hexagonal CrSi2 structure type. Dark brown balls correspond to Si
atoms, light green balls correspond to W atoms, respectively.
However, it seems that there are still many open questions
regarding the structural behavior of WSi2, such as its depen-
dence on grain size, composition, surface morphology, time of
annealing, where even the conductivity can be dependent on
the annealing temperature.[22–25] Furthermore, since W and Si
react to form WSi2 at 1250 °C but form W5Si3 already at
1300–1400 °C, it is rather difficult to synthesize the pure sili-
cide phases, without additional impurity phases, in large
amounts.[4,18] We hope that this study, which focuses on struc-
tural properties of WSi2, and combines experimental and theo-
retical investigations, will give some answers and provide a
starting point for further research into new modifications of
tungsten silicides and their respective properties.
Since it is possible, in principle, that additional symmetries appear
during the local optimization, the symmetry of the optimized structures
was analyzed with the algorithms SFND (“Symmetry FiNDer”[40]) and
RGS (“Raum Gruppen Sucher” = space group seeker[41]), as im-
plemented in the program KPLOT.[42] The CMPZ algorithm[43] im-
plemented in the KPLOT program was used to compare the various
crystal structures. The structures investigated were visualized using
KPLOT[42] and the VESTA[44] software.
3 Results and Discussion
3.1 XRPD Analysis of Tungsten Silicides
2 Experimental Section
As mentioned earlier, tungsten and silicon react to form
tungsten disilicide (WSi2) at 1250 °C, and pentatungsten tri-
silicide (W5Si3) at higher temperatures (1300 to 1400 °C),
respectively. The temperature of 1400 °C is sufficiently high
2.1 Experimental Methods
2.1.1 Preparation of Tungsten Silicide Powder
For this study, commercial tungsten powder (Koch-Light Laboratories,
LTD, purity 99.9%) of average grain size of 1 μm according to manu- for the conversion of the starting materials to tungsten based
Z. Anorg. Allg. Chem. 2017, 2088–2094
www.zaac.wiley-vch.de 2089
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
silicides without any remains of unreacted tungsten and sili- be synthesized, requires a global exploration of its energy land-
con, but impurities like WC are present in the material, proba- scape,[45–47] in principle. The most common exploration meth-
bly due to the presence of carbon content in the furnace, lead- ods have been described in detail elsewhere.[48] However, if
ing to a reaction with tungsten to form WC. Heating for 4 h the interactions between the atoms cannot easily and robustly
at 1350 °C leads to a complete conversion of those elementary be described by simple empirical potentials (thus requiring the
powders to WSi2 and W5Si3. This is clearly confirmed in Fig- use of computationally expensive ab initio energy calcula-
ure 2, which shows the XRPD patterns of the sample after tions) – as is the case for the W/Si system –, then a useful first
heating at 1350 °C for 4 h in an argon flow.
step is to employ data mining based searches[49–54] to gain an
overview over the richness of the energy landscape. Thus, in
this study, we have performed such searches starting from the
ICSD database,[55] where we have focused on related silicide
systems, as well as related binary compounds. Besides the ex-
perimentally observed tetragonal MoSi2 and hexagonal CrSi2
modifications of WSi2, we have investigated several additional
promising modifications of tungsten disilicides.
First, we note that since transition metal silicide based re-
fractory composite alloys have great potential for applications
as, for example, the next generation of aerospace materials,
there are four crystal structures of known refractory disilicides
that are worthwhile targets of investigation in the W-Si system:
the CrSi2 (C40) type, the MoSi2 (C11b) type, the TiSi2 (C54)
type, and the ZrSi2 (C49) type.[56] Thus, we have investigated
the related structures in WSi2 on ab initio level, starting with
the TiSi2 structure type, which possesses orthorhombic sym-
metry with space group FdddZ (no. 70) (Strukturbericht desig-
nation C54, and Pearson symbol oF24). This modification can
be a good structure candidate, since the original TiSi2 structure
was preserved after the DFT optimization (confirmed using the
CMPZ algorithm in KPLOT), and it shows a ten-fold coordina-
tion of tungsten by Si atoms (CN = 10, see Figure 3a), similar
to the experimentally observed structures. In addition there
have been earlier theoretical suggestions[57] of the TiSi2 (C54)
structure type for WSi2, in agreement with our DFT calcula-
tions.
Next, we have investigated the ZrSi2 type of structure
(Strukturbericht designation C49, and Pearson symbol oC12)
in the W-Si system, another orthorhombic structure candidate.
However, the original ZrSi2 structure type was not preserved
after the DFT optimization; instead we observe a novel ZrSi2-
like type of structure in the WSi2 system (cf. Figure 3b). Al-
though the ZrSi2-like modification is showing space group
Cmcm (no. 63), like the original ZrSi2 type, changes in crystal
structure and coordination number are observed. The original
ZrSi2 type has 10 fold coordination of Zr by Si atoms, whereas
the novel ZrSi2-like modification has 11 fold (2+4+5) coordi-
nation of W by Si atoms. In particular, the coordination poly-
hedron contains two silicon atoms sitting much closer to tung-
sten (ca. 1.7 Å) than the next four Si atoms, which are located
about 2.5 Å away from the W atom (which is the average dis-
tance in other WSi2 modifications). Finally, there are five dis-
tant Si atoms, which are about 3.7–3.8 Å away from the central
tungsten atom.
Figure 2. Rietveld diagram of WSi2 and W5Si3 powders obtained after
heating at 1350 °C for 4 h in an argon atmosphere. Blue line denotes
difference between experimental (red diamonds) and theoretical (black
line) profile, while the Bragg positions are indicated by vertical green
slashes.
The X-ray analysis showed a two-phase composition for the
investigated sample, where the WSi2 phase is dominant (ca.
85 mass%) compared to W5Si3. This suggests that further fine-
tuning of the synthesis process should allow the production of
nearly phase pure WSi2. The results of the Rietveld refinement
are presented in Table 1.
Table 1. The parameters of the unit cells for the experimental phases
obtained by Rietveld refinement.
WSi2
W5Si3
Space group (no.)
I4/mmm (no. 139)
I4/mcm (no. 140)
Unit cell parameters
a /Å
b /Å
c /Å
α /°
3.2083(1)
3.2083(1)
7.8216(4)
90
9.5917(3)
9.5917(3)
4.9690(2)
90
β /°
90
90
γ /°
90
80.5075
90
457.1598
Volume V /Å3
Atomic positions
(x,y,z)
W1
0 0 0
0 0.5 0.25
W2
Si1
0.075(1) 0.222(1) 0
0 0 0.25
0 0 0.322(1)
Si2
0.168(6) 0.668(6) 0
Although it appears that WSi2 can exist in a variety of seem-
ingly complicated orthorhombic (the TiSi2 and the ZrSi2-like),
hexagonal (the CrSi2 type), and tetragonal (the MoSi2 type)
3.2 Modeling of WSi2 on ab initio Level
Predicting, which crystalline modifications can exist in a modifications, they do in fact share a common structural build-
chemical system or even, which unknown chemical system can ing block that consists of nearly hexagonal WSi2 layers. In
Z. Anorg. Allg. Chem. 2017, 2088–2094
www.zaac.wiley-vch.de 2090
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
As an additional candidate suggested by our search through
the ICSD, we have studied the SrSi2 modification of the WSi2
compound, which had not yet been investigated in the tungsten
silicide system. The SrSi2 structure type crystallizes in cubic
symmetry with space group P4332 (no. 212) (Pearson symbol
cP12). After the local DFT optimization of the SrSi2-like
modification of tungsten disilicide, the silicon atoms behave
as three-covalent anions bonded to neighbors, whereas the W
cations exhibit a sixfold coordination by silicon (CN = 6), with
average distance of ca. 2.6 Å. In the original SrSi2 compound,
Sr is eightfold coordinated by silicon, with two longer axial
bonds and six shorter pseudo-equatorial bonds. In the tungsten
disilicide the Si-atoms of the longer axial bonds reside far
away from the first coordination polyhedron (formed by the
six nearest Si-neighbors of the tungsten atom, at distances of
about 3.5 Å. This displacement effectively reduces the coordi-
nation to six; it is possibly due to a partially covalent character
(see Figure 4a).[61]
Figure 3. Visualization of two proposed modifications for the WSi2
compound: (a) the TiSi2 structure type; (b) the ZrSi2-like structure
type. Dark brown balls correspond to Si atoms, light green balls corre-
spond to W atoms, respectively.
particular, orthorhombic (the TiSi2 type), hexagonal (the CrSi2
type), and tetragonal (the MoSi2 type) modifications can be
generated by introducing simple variations in the stacking se-
quence of these nearly hexagonal WSi2 layers.[58–60] This
structural systematic seems to be related to the d-band filling
of the transition metal constituents of the structural prototypes,
e.g. the group-IV disilicides form orthorhombic TiSi2 or ZrSi2
structures, that contain four-layer (ABCD) and two-layer (AB)
stacking sequences, respectively. On the other hand, the group
V compounds consistently adopt the three-layer (ABC)
hexagonal CrSi2 type structure, while the stable phases of the
group VI materials include both the hexagonal CrSi2 structure,
and the two-layer (AB) tetragonal MoSi2 type of structure. The
transition from hexagonal to tetragonal symmetry among the
isoelectronic group-VI compounds suggests a near degeneracy
in the structural energies for these phases.[58–60] This is sup-
ported by the fact that metastable hexagonal MoSi2 and WSi2
Figure 4. Visualization of two proposed structures in the WSi2 system:
(a) the SrSi2-like structure type; (b) the CaF2 structure type. Dark
brown balls correspond to Si atoms, light green balls correspond to W
atoms, respectively.
Finally, we present another not-yet theoretically or experi-
thin films can be formed using the ion-implantation technique mentally observed structure in the W–Si system: the CaF2
and low annealing temperatures.[22–24,60]
structure (Figure 4b). The CaF2 structure type is another cubic
Z. Anorg. Allg. Chem. 2017, 2088–2094
www.zaac.wiley-vch.de 2091
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
Table 2. Cell parameters and atom positions of the WSi2 modifications observed in our experiments at 1350 °C and from earlier calculations in
the literature, compared to the calculated/predicted WSi2 modifications after local optimization on ab initio level using LDA and GGA-PBE
functionals in this study. Note that the difference between the cell parameters determined by experiment (at ca. 1600 K) and by theory (at 0 K)
is below 2%, which is very satisfactory considering the general limitations of DFT-calculations.
Modification and space group
Cell parameters /Å and fractional coordinates
a)
b,c)
a)
a)
Experiment / Theory
GGA-PBE
LDA
a)
MoSi2 type – equilibrium
I4/mmm (no. 139)
a = 3.208, c = 7.821
W (0 0 0)
a = 3.249, c = 7.948
W (0 0 0)
a = 3.214, c = 7.887
W (0 0 0)
Si [0 0 0.322(1)]
Si (0 0 0.3346)
Si (0 0 0.3347)
b)
CrSi2 type – LT phase
P6222 (no. 180)
a = 4.614, c = 6.414
W (0.5 0 0.5)
a = 4.656, c = 6.672
W (0.5 0 0)
a = 4.607, c = 6.624
W (–0.5 0 0)
Si (0.164 0.328 0.5)
Si (0.1639 0.3278 0)
Si (0.1641 0.3282 0)
c)
TiSi2 type – predicted
FdddZ (no. 70)
a = 7.838, b = 4.524, c = 8.959
a = 8.060, b = 4.630, c = 8.976
W (0.125 0.125 0.125)
a = 7.964, b = 4.590, c = 8.906
W (0.125 0.125 0.125)
Si (–0.0377 –0.375 0.125)
Si (–0.0383 –0.375 0.125)
c)
ZrSi2-like type – predicted
a = 3.180, b = 12.230, c = 4.151
a = 3.486, b = 15.524, c = 3.391 a = 3.448, b = 15.283, c = 3.361
Cmcm (no. 63)
W (0.5 –0.1276 0.25)
Si (0 0.0613 0.25)
Si (0 –0.2478 0.25)
W (0.5 –0.1266 0.25)
Si (0 0.0610 0.25)
Si (0 –0.2480 0.25)
SrSi2-like type – predicted
P4332 (no. 212)
–
–
a = 5.819
W (0.125 0.125 0.125)
Si (–0.2208 –0.2208 –0.2208)
a = 5.755
W (0.125 0.125 0.125)
Si (–0.2202 –0.2202 –0.2202)
CaF2 type – predicted
a = 5.927
W (0 0 0)
a = 5.877
W (0 0 0)
¯
Fm3m (no. 225)
Si (0.25 –0.25 –0.25)
Si (0.25 –0.25 –0.25)
a) This study. b) Reference[22]. [c] Reference[54]
.
¯
structure candidate, showing space group Fm3m (no. 225). In ergy minimum to the equilibrium WSi2 structure (see Table 3
the literature, it is known from the mineral fluorite (Pearson and Figure 5). Again, our calculations are in a good agreement
symbol cF12).[62] Similar to the previous cubic candidate with previous experiments and calculations, where the CrSi2
(SrSi2), tungsten is eightfold coordinated by silicon atoms in type modification had been found as a metastable structure,
the CaF2 modification.
experimentally observed at low temperatures and/or in thin
A complete list of the structure candidates calculated using films.[22–24]
DFT, with corresponding unit cell parameters and atomic posi-
Table 3. Calculated total energies of the WSi2 modifications after local
optimization on ab initio level, using LDA and GGA-PBE functionals.
Energy per formula unit is given in hartree (Eh).
tions, is presented in Table 2. Our DFT calculations are in very
good agreement with our experimental results and previous
experimental observations.[8,9,19,20,22–24] Similarly, the calcula-
tions presented are in very good agreement with previous theo-
retical investigations where available.[57,60,63] We note a
slightly better agreement when using the LDA functional com-
pared to the GGA-PBE results.
Regarding the kinetic stability of these predicted structures,
we note that the local minimizations were performed without
symmetry constraints and involved considerable displacements
of the atoms in the case of the newly predicted structures, sug-
gesting that the final structures are true local minima and not
saddle points.
Modification
Total energy /Eh
GGA-PBE
LDA
MoSi2 type – equilibrium
CrSi2 type – LT phase
TiSi2 type – predicted
ZrSi2-like type – predicted
CaF2 type – predicted
–646.0386
–646.0321
–646.0246
–645.9949
–645.9433
–645.8936
–643.9394
–643.9312
–643.9240
–643.8929
–643.8379
–643.7880
SrSi2-like type – predicted
Next, we consider the TiSi2 type modification, which is an
energetically favorable structure candidate, with a similar en-
Finally, a total energy ranking is presented in Table 3, and ergy as the experimentally observed modifications of WSi2
E(V) curves are shown in Figure 5. We observe that for the (see Table 3). Although, the TiSi2 type modification of WSi2
total energy ranking, the results of the calculations using the appears to be metastable, its E(V) curve is very close to the
GGA functional are consistent and comparable with those for one computed for the experimentally found low temperature
the LDA functional (Table 3). We note that the equilibrium CrSi2 modification, and the TiSi2 type is an even better struc-
WSi2 modification is the lowest energy minimum regardless ture candidate at expanded volumes than the CrSi2 type and
of DFT method used, which is in agreement with earlier exper- possibly also than the MoSi2 type (see Figure 5). This could
imental and theoretical data for a stable phase of WSi2 at nor- be an indication of a possible synthesis route for the TiSi2
mal conditions. From the calculated E(V) curves, the metasta- type as a novel WSi2 modification: perhaps the actual TiSi2
ble CrSi2 type has been observed as the closest calculated en- compound can be used as a template for the deposition of a
Z. Anorg. Allg. Chem. 2017, 2088–2094
www.zaac.wiley-vch.de 2092
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
Finally, new structures have been suggested for WSi2 and
investigated using first-principles calculations within the den-
sity-functional theory (DFT) approximation. The LDA and
PBE based calculations show excellent agreement with experi-
mental observations and previous calculations (where avail-
able). In addition, we discuss several not-yet observed or com-
putationally studied modifications of tungsten disilicide, which
might be accessible experimentally. A successful synthesis of
these new WSi2 phases would have a great impact on the sci-
entific, technological and industrial application of tungsten
based silicides
Acknowledgements
This work has been supported by the grant No. 45012 and No. 171001
from the Ministry of Education, Science and Technological Develop-
ment of the Republic of Serbia. The authors thank M. Rosic, K. Djuris,
B. Babic, and V. Maksimovic for valuable discussions. The authors
thank Prof. R. Dovesi, Prof. K. Doll, and Crystal Solutions for software
support.
Figure 5. Energy vs. volume curves, E(V), calculated for the investi-
gated structure types in the WSi2 system, using the LDA functional.
The main Figure shows the E(V) curves for the four most relevant
structure types (MoSi2, CrSi2, TiSi2, and ZrSi2-like), while those of
the SrSi2-like modification and the CaF2-type have been placed into
an inset because of their very high energies compared with the other
four structure types. Energies per formula unit are given in hartree
(Eh).
Keywords: Tungsten disilicide; Tungsten; XRPD analysis
W/Si vapor producing an amorphous WSix film from which
the predicted TiSi2 phase of WSi2 can be generated by temper-
ing at low temperatures – a type of procedure successfully
References
employed in the past to synthesize metastable com- [1] S. Oyama, Introduction to the Chemistry of Transition Metal
pounds.[64,65]
Carbides and Nitrides, in The Chemistry of Transition Metal
Carbides and Nitrides, Springer Netherlands, 1996, pp. 1–27, .
[2] H. O. Pierson, Handbook of Refractory Carbides and Nitrides:
Another interesting metastable structure candidate in the ef-
fective negative pressure region is the ZrSi2-like modification
Properties, Characteristics, Processing and Applications,
(Figure 5). However, the ZrSi2-like modification is energeti-
cally considerably less favorable than the CrSi2 and the TiSi2
phases. Nevertheless, Figure 5 indicates that at large effective
negative pressures it is likely to become the thermodynami-
cally stable phase, and thus a similar synthesis approach as the
Elsevier Inc., 1996.
[3] C. Gelain, A. Cassuto, P. Le Goff, Oxidation Met. 1971, 3, 153.
[4] S. L. Kharatyan, H. Chatilyan, A. B. Harutyunyan, Defect Dif-
fusion Forum 2001, 194–199, 1557.
[5] K. H. Lee, J. K. Yoon, J. K. Lee, J. M. Doh, K. T. Hong, W. Y.
Yoon, Surface Coatings Technol.. 2004, 187, 146.
one suggested above for the TiSi2-type phase may be feasible. [6] J. K. Yoon, K. W. Lee, S. J. Chung, I. J. Shon, J. M. Doh, G. H.
Kim, J. Alloys Compd. 2006, 420, 199.
[7] L. X. Cai, H. M. Wang, Appl. Surf. Sci. 2004, 235, 501.
On the other hand, the other predicted structures with SrSi2-
like and CaF2 type structures show much higher total energies
[8] J. Y. Suh, S. E. Shin, D. H. Bae, Powder Technol.. 2015, 289, 401.
(see Table 3 and Figure 5), and these modifications would
[9] J. Lukovic´, B. Babic´, D. Bucˇevac, M. Prekajski, J. Pantic´, Z.
probably be very difficult to synthesize in the WSi2 system.
We therefore conclude that our data-mining study has gener-
ated a variety of structure candidates some of which appear to
be promising candidates for an alternative new WSi2 modifica-
tion, and further investigations of their structural, physical and
chemical properties are planned.
Basˇcˇarevic´, B. Matovic´, Ceram. Int. 2015, 41, 1271.
[10] F. Mohammadi, K. C. Saraswat, J. Electrochem. Soc. 1980, 127,
450.
[11] M. Y. Tsai, C. S. Petersson, F. M. dЈHeurle, V. Maniscalco, Appl.
Phys. Lett. 1980, 37, 295.
[12] D. L. Kwong, D. C. Meyers, Appl. Phys. Lett. 1985, 47, 688.
[13] B. Y. Tsaur, C. K. Chen, C. H. Anderson Jr., D. L. Kwong, J.
Appl. Phys. 1998, 57, 1890.
[14] K. C. Saraswat, D. L. Brors, J. A. Fair, K. A. Monnig, R. Beyers,
IEEE Trans. Electron Devices 1983, 30, 1497.
[15] J. M. Jasinski, B. S. Meyerson, B. A. Scott, Low Pressure Chemi-
cal Vapor Deposition of Tungsten Silicide, U. S. Patent Nr. US
4684542 A, International Business Machines Corporation, United
States, 1987.
[16] M. Y. Tsai, F. M. dЈHeurle, C. S. Petersson, R. W. Johnson, J.
Appl. Phys. 1981, 52, 5350.
[17] Y. Shioya, T. Itoh, I. Kobayashi, M. Maeda, J. Electrochem. Soc.
1986, 133, 1475.
[18] H. S. Kim, J. K. Yoon, G. H. Kim, J. M. Doh, S. I. Kwun, K. T.
Hong, Intermetallics 2008, 16, 360.
[19] H. Kudielka, H. Nowotny, Monatsh. Chem. 1956, 87, 471.
[20] A. N. Christensen, J. Cryst. Growth 1993, 129, 266.
4 Conclusions
The W-Si system is one of the most important refractory
metal–silicon systems due to their great technological and in-
dustrial applications. In order to investigate tungsten based sili-
cide, we have performed combined experimental and theoreti-
cal investigations. Tungsten disilicide (WSi2) was synthesized
by simple thermal treatment at 1350 °C for 4 h in an argon
atmosphere. We found that optimizing the synthesis conditions
led to a mix of pure WSi2 (85%) and W5Si3 (15%) phases,
with negligible amounts of impurity phases present.
Z. Anorg. Allg. Chem. 2017, 2088–2094
www.zaac.wiley-vch.de 2093
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Journal of Inorganic and General Chemistry
ARTICLE
Zeitschrift für anorganische und allgemeine Chemie
[21] Y. Harada, M. Morinaga, D. Saso, M. Takata, M. Sakata, Inter-
[42] R. Hundt, KPLOT: A Program for Plotting and Investigation of
Crystal Structures, Technicum Scientific Publishing, Stuttgart,
Germany, 2017.
[43] R. Hundt, J. C. Schön, M. Jansen, J. Appl. Crystallogr. 2006, 39,
metallics 1998, 6, 523.
[22] F. M. d’Heurle, C. S. Petersson, M. Y. Tsai, J. Appl. Phys. 1980,
51, 5976.
[23] M. Y. Tsai, F. M. d’Heurle, C. S. Petersson, R. W. Johnson, J.
Appl. Phys. 1981, 52, 5350.
[24] J. H. Liang, D. S. Chao, Surface Coatings Technol. 2001, 140,
116.
[25] J. Stoetzel, T. Schneider, M. M. Mueller, H.-J. Kleebe, H. Wig-
gers, G. Schierning, R. Schmechel, Acta Materialia 2017, 125,
321.
6.
[44] K. Momma, F. Izumi, J. Appl. Crystallogr. 2008, 41, 653.
[45] J. C. Schön, M. Jansen, Angew. Chem. Int. Ed. Engl. 1996, 35,
1268.
[46] S. M. Woodley, C. R. A. Catlow, Nat. Mater. 2008, 7, 937.
[47] J. C. Schön, K. Doll, M. Jansen, Phys. Status Solidi B 2010, 247,
23.
[48] A. R. Oganov (Editor), Modern Methods of Crystal Structure Pre-
diction, Wiley VCh, Weinheim, 2011.
[49] J. C. Schön, Z. Anorg. Allg. Chem. 2014, 640, 2717.
[50] S. Curtarolo, D. Morgan, K. Persson, J. Rodgers, G. Ceder, Phys.
Rev. Lett. 2003, 91, 135503.
[26] J. Rodriguez-Carvajal, Collected Abstract of Powder Diffraction
Meeting, Toulouse, France 1990, p. 127.
[27] R. Dovesi, V. R. Saunders, C. Roetti, R. Orlando, C. M. Zicovich-
Wilson, F. Pascale, B. Civalleri, K. Doll, N. M. Harrison, I. J.
Bush, P. DЈArco, M. Llunell, CRYSTAL14 User’s Manual, Univer-
sity of Torino, Torino, Italy 2009.
[51] G. Ceder, D. Morgan, C. Fischer, K. Tibbetts, S. Curtarolo, MRS
Bull. 2006, 31, 981.
[28] R. Dovesi, R. Orlando, B. Civalleri, C. Roetti, V. R. Saunders,
C. M. Zicovich-Wilson, Z. Kristallogr. 2005, 220, 571.
[29] F. Pascale, C. M. Zicovich-Wilson, R. Orlando, C. Roetti, P. Ugli-
engo, R. Dovesi, J. Phys. Chem. B 2005, 109, 6146.
[30] D. Andrae, U. Haeussermann, M. Dolg, H. Stoll, H. Preuss,
Theor. Chim. Acta 1990, 77, 123.
[31] K. Doll, V. R. Saunders, N. M. Harrison, Int. J. Quantum Chem.
2001, 82, 1.
[32] K. Doll, Mol. Phys. 2010, 108, 223.
[52] J. Zagorac, D. Zagorac, A. Zarubica, J. C. Schön, K. Djuris, B.
Matovic, Acta Crystallogr., Sect. B 2014, 70, 809.
[53] D. Zagorac, J. C. Schön, M. Jansen, Proc. Appl. Ceram. 2013, 7,
37.
ˇ
[54] M. Cebela, D. Zagorac, K. Batalovic´, J. Radakovic´, B. Stojadi-
novic
´, V. Spasojevic´, R. Hercigonja, Ceram. Int. 2017, 43, 1256.
[55] G. Bergerhoff, I. D. Brown, Crystallographic Databases
(Ed. F. H. Allen, G. Bergerhoff, R. Sievers), Chester, International
Union of Crystallography, 1987.
[56] Pearson’s Handbook of Crystallographic Data for Intermetallic
Phases (Eds.: P. Villars, L. D. Calvert), 2nd ed., Materials Park
(OH), ASM International; 1991.
[33] B. Civalleri, P. D’Arco, R. Orlando, V. R. Saunders, R. Dovesi,
Chem. Phys. Lett. 2001, 348, 131.
[34] J. C. Schön, Z. Cancarevic, M. Jansen, J. Chem. Phys. 2004, 121,
2289.
[57] G. Shao, Acta Materialia 2005, 53, 3729.
[58] W. Jeitschko, Acta Crystallogr., Sect. B 1977, 33, 2347.
[59] H. Nowotny, in: Electronic Structure and Alloy Chemistry of the
Transition Elements (Ed.: P. A. Beck), Interscience, New York,
1963, pp. 179–220.
[35] J. C. Schön, I. V. Pentin, M. Jansen, Phys. Chem. Chem. Phys.
2006, 8, 1778.
[36] D. Zagorac, K. Doll, J. C. Schön, M. Jansen, Phys. Rev. B 2011,
84, 045206.
[37] D. Zagorac, J. C. Schön, J. Zagorac, M. Jansen, RSC Adv. 2015,
5, 25929.
[38] J. P. Perdew, A. Zunger, Phys. Rev. B 1981, 23, 5048.
[39] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865.
[40] R. Hundt, J. C. Schön, A. Hannemann, M. Jansen, J. Appl. Crys-
tallogr. 1999, 32, 413.
[60] L. F. Mattheiss, Phys. Rev. B 1992, 45, 3252.
[61] G. E. Pringle, Acta Crystallogr., Sect. B 1972, 28, 2326.
[62] D. N. Batchelder, R. O. Simmons, J. Chem. Phys. 1964, 41, 2324.
[63] K. C. Bhamu, J. Sahariya, B. L. Ahuja, J. Phys. Chem. Solids
2013, 74, 765.
[64] D. C. Johnson, Curr. Opin. Solid State Mater. Sci. 1998, 3, 159.
[65] D. Fischer, M. Jansen, Angew. Chem. Int. Ed. 2002, 41, 1755.
[41] A. Hannemann, R. Hundt, J. C. Schön, M. Jansen, J. Appl. Crys-
Received: August 31, 2017
tallogr. 1998, 31, 922.
Published online: November 23, 2017
Z. Anorg. Allg. Chem. 2017, 2088–2094
www.zaac.wiley-vch.de 2094
© 2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim