252
J. Ma et al. / Journal of Alloys and Compounds 381 (2004) 250–253
sample with C49-type TiSi2, the C54-type TiSi2 nucleation
occurred at 800 ◦C. In our present route, the obtained TiSi2
using metallic sodium as reductant is of pure C54-type.
There is no C49-type phase in the XRD patterns. And when
the reaction temperature is 550 ◦C, the obtained TiSi2 is
also of the C54-type, though the crystallinity is not very
good. Since the reaction is carried out in an autoclave, the
pressure is very high when the reaction takes place. And
the high pressure may be in favor of the formation of the
C54-type TiSi2 phase.
In this work, we have prepared different titanium sili-
cides via a co-reduction route with different reductants. So
we can conclude that the reductant plays an important role
in the formation of different titanium silicides. Among the
four metal reductants, metallic sodium has the most reaction
activity. When forming the titanium silicide using metallic
sodium as the reductant, the reaction is more violent, gen-
erating much more heat, which can result in a phase trans-
formation of the C49-type to the C54-type. So the obtained
product is titanium disilicide. However, when using metal-
lic magnesium, zinc or aluminum as reductant, the reaction
is not very violent and not much heat is produced. So only
Ti5Si3 can be formed. In short, the different reaction activ-
ities of the reductants result in the formation of different
titanium silicides.
Fig. 2. Transmission electron microscope images of the samples obtained
via a co-reduction route using different metal reductants: (a) Na; (b) Mg;
(c) Zn; (d) Al.
4. Conclusions
In summary, different titanium silicides were successfully
synthesized via a co-reduction route by reacting silicon
tetrachloride, titanium tetrachloride with different metal re-
ductants (Na, Mg, Zn, Al) in an autoclave at 650 ◦C. The
orthorhombic TiSi2 was obtained using metallic sodium as
reductant. It is of the C54-type with an average particle
size of 60 nm in diameter. However, the hexagonal titanium
silicide (Ti5Si3) was obtained when using other metal re-
ductants. Transmission electron microscope images showed
that the average particle sizes of Ti5Si3 were 40, 30 and
25 nm in diameter, respectively. The different reaction ac-
tivities of the metal reductants plays an important role in
the formation of the different titanium silicides.
silicon tetrachloride, that is, all silicon tetrachloride cannot
be reduced at that temperature. Heating at a higher temper-
ature such as 700 ◦C will result in an obvious increase of
the crystallite sizes to about 80 nm. An optimum tempera-
ture for Ti5Si3 is about 650 ◦C. A reaction time at 650 ◦C in
the range of 6–10 h did not significantly affect the crystal-
lite size. If the time is shorter than 4 h, the reaction becomes
very incomplete and the crystallinity is very poor.
TiSi2 can occur in the form of two phases: the metastable
C49-type phase (base-centered orthorhombic) and the
C54-type phase (face-centered orthorhombic). The C49-type
phase TiSi2 has a high resistivity while the C54-type phase
has a low resistivity. A persistent challenge with the for-
C49-type phase to the low-resistivity C54-type phase [11].
phase (C49-type) to the low-resistivity phase (C54-type)
makes its update application difficult [12–14]. And the poly-
morphic transition from the C49-type phase to the C54-type
phase is still a longstanding problem [15]. Since the phase
transition is a nucleation-controlled process [16] due to the
[17], higher temperatures are therefore required to fully
transform the C49-type to the C54-type phase. It is known
that C49-type TiSi2 usually transforms into C54-type TiSi2
above 700 ◦C [18]. Ottaviani et al. [19] showed that in a
Acknowledgements
Financial support from the Chinese National Foundation
of Natural Science Research and the 973 projects of China
are gratefully acknowledged.
References
[1] J.M.M. de Nils, A. van Silfhout, Mater. Sci. Eng. B 5 (1990) 319.
[2] G.W. Rubloff, R.M. Tromp, Appl. Phys. Lett. 48 (1986) 1600.
[3] Y.U. Idzerda, E.D. Williams, R.L. Park, Surf. Sci. 177 (1986) 1028.