". ind'''''"!1
I
~
'
e
. ~
:orporatlon \
m grant no.
i
¥
~.
I;
~
~
If
~
S. J Luszez,i
[, R ZWICKER!
>:
'ioN.C R KENJ
'er Sci 24 (1989)~
~
. WHITE, C R.t
T. D CLAAR ~
f
~
n Proceedings oi ~
Mulricomponem S.
:r 19kk. edited bvW
.lI1d R. W. Sie g
;1 f
v.t
nnsylvania. 1989)L
HAJANIAN and~
)988)975.
1
7
11. P. BtH. in
i
d AIN IAI AlIovs. )
Meeti;lg, Dall~s.I
.
!
iety, Westerville. ~
~
I. P. BIEL and
,
'.
'
.
.,
'
13) 1865. ...
G. SMITH and
194)293. ~,
~.
I
I
t
JOURNAL OF MATERIALS SCIENCE LETTERS 14 (J9951 31-32
Preparation and nitridation of silicon whiskers
p.ls. GOPALAKRISHNAN, P. S. LAKSHMINARASIMHAM
MJterialsScience Division, National Aerospace Laboratories, Bangalore 560017, India
,
,
.
I
t
f
Silicon nitride is or great interest as a high tempera-
ture ceramic material for use in engine and turbine
parts, cutting tools and bearings [1]. To fabricate
reliable components out of silicon nitride for these
applications it is essential that the starting powder of
silicon nitride should have high purity, uniform
particle size and high percentage of a-phase [2].
Various methods of preparing silicon nitride powder
have been discussed by many authors [3-5]. The
major problem encountered in the direct nitridation
of silicon is that the product is always coarse,
necessitating prolonged milling, which causes conta-
mination. In the carbothermal reduction of silica, a
very large excess of carbon has to be used, which is
removed later by heating in air, but a small amount
of carbon always remains as an impurity. In the
ammonolysis of silicon halides the intermediate
product, silicon imide, is highly sensitive to air and
moisture, and complete removal of the halide is also
difficult. We have been looking for a process in
which these problems are avoided. In this connec-
tion one of our interests has been to see whether
high purity silicon powder can be prepared from
silicontetrachloride and whether this powder can be
nitrided in situ. However, our experiments have not
been successful in this regard, but resulted in silicon
whiskers and whisker-like silicon nitride. These
experiments are reported here.
Metals like magnesium, iron, zinc and cadmium
can reduce silicon tetrachloride to silicon. This
principle had been used to prepare semiconductor
grade silicon. Whiskers of silicon have been pre-
pared by the disproportionation of silicon dibromide
[6], gas phase reduction of silicon tetrachloride with
hydrogen [7] and thermal decomposition of silane
[8]. Our attempts to prepare silicon powder by the
reduction of silicon tetrachloride with iron filings
resulted in ferrosilicon globules. It was realized that
to get silicon powder, a reaction between the metal
vapour and silicon tetrachloride vapour has to be
employed. Zinc, being a metal with low boiling
point, was considered suitable for this purpose.
Recently, magnesium vapour has been used to
reduce silicon tetrachloride for the preparation of
siliconpowder [9].
The experiment basically involves production of
zinc vapour by heating the metal to its boiling point
and reacting this vapour with silicon tetrachloride
vapour diluted with nitrogen. A schematic of the
experimental set-up is shown in Fig. 1. It consists of
a tWo-zonehorizontal furnace; one zone, of length
200mm, is for generating the zinc vapour and the
other, of length 400 mm, is for reacting the zinc
026l-8028 @ 1995Chapman&Hall
Zone 2 Zone1
-~
T
1
j:
L
rr-N2
I
-
Zinc
chloride
SiCI4+N2
Granulated
zinc
Silicon
whiskers
Figure / A schematic diagram of the experimental
seHlp.
vapour with silicon tetrachloride vapour in the
presence of nitrogen. The reaction tube is of fused
quartz with diameter 25 mm and length 1000 mill.
About 10g of granulated zinc (Analar, SOH) was
put in a quartz tube of diameter IS mm and length
ISO mm with one end closed and the other end
tapered. This was enclosed in a jacket made out of
fused quartz with a diameter of 20 mm and length of
200 mm. This jacket had an inlet through which dry
nitrogen could be passed. This was kept in zone 1 of
the furnace. The temperature of this zone was
increased gradually to above the boiling point of zinc
while passing nitrogen at a rate of 11 min-l, through
the jacket so that the zinc vapour was carried into
the reaction zone (zone 2 of the furnace). The
temperature of this zone was maintained between
900 and 1200°c. Dry nitrogen was bubbled at a rate
of 0.5 Imin-I through distilled silicon tetrachloride
liquid in a bottle, and the gaseous mixture of silicon
tetrachloride and nitrogen was passed over the
jacket to the reaction zone. The reaction took place
according to the equation:
SiCI4(g)+ 2Zn(g) ~ ZnCI2(g) + Si(C)
White fumesof zinc chloride were observed from the
outlet of the reaction tube as long as the reaction was
taking place. These fumes also condensed at the
cooler end of the reacton tube. The temperature of
both the zones was reduced after the reaction was
over. The product was taken out of the reaction tube
after it had cooled to the room temperature. It was
observed that the products of the reaction con-
densed in two distinct regions of the reaction tube.
The product that condensed in the cooler region was
pure zinc chloride, and the product that condensed
in the hot zone consisted of only whiskers. These
whiskers were characterized by X-ray diffraction
(XRD) and electron microscopy. The XRO pattern
confirmed that these whiskers were only crystalline
silicon but not silicon nitride. When the temperature
of the reaction was 1200°e, the whiskers were thick
31
Figure 2 Scanning electron micrograph of the silicon whiskers.
and had the characteristic lustre of silicon. At lower
temperature the whiskers were finer and the colour
was light brown. When the reaction temperature was
950°C, the whiskers were very thin and looked like
wool, creamy white in colour. AD.electron micro-
graph of these whiskers is shown in Fig. 2. The
thickness of these whiskers was about 0.5 ,urn and
their length was more than 1 em.
These experiments were repeated with increased
nitrogen flow to see whether the whiskers could be
nitrided in situ. However, these experiments were
not successful in this regard, but the whiskers
became still finer, down to a thickness of 0.1 ,urn.
Therefore, the silicon whiskers were nitrided separ-
ately using dry ammonia gas (10] in the temperature
range 1200-1300 dc. The XRD and infrared absorp-
tion data showed that the product of nitridation was
a-silicon nitride. An electron micrograph of this
silicon nitride is shown in Fig. 3. By comparing this
micrograph with that of silicon whiskers, it can be
inferred that during the nitridation the silicon
whiskers do not become powder, but become
shorter, probably pseudomorphs. The thickness. of
these pseudomorphs was about 0.5 ,urn.
The length is fairly uniform and is about 12 ,urn.
Such whisker-like particles with uniform aspect ratio
may also be useful in ceramic composites.
"
32
Figure 3 Scanning electron micrograph of the whisker-like silicon
nitride.
Acknowledgement
We thank Mr M. A. Venkataswamy for the electron
micrographs.
References
l. L. 1\1 SHEPPARD. Amer. Ceram. Soc. Hull. 69 (llJ90)
1012.
2. \V ENGEL. Powder Mer. 1m. 10 (1978) 124.
3. H. LANGE. G. WOTTING and G. WINTER. Angew.
Chon. lilt Ed Eng!. 30 ( 1991) 1579.
4. G. M. CROSBIE. R. L. PREDMESK Y. J M NICHOL-
SON and E. D ST I L E S. Amer. Ceral/!. soc.
.
Bull 6!'!(1989)
\
"
.,,".,
1010 :
5. K S. MAZDIYASNI and C. tv! COOKE. 1. Amer.
Ceram Soc. 56 (1973) 628. .
6. A. V SANDULOVA. I. P BOGOYAVLENSKAYA and
L.IPYRSKO.NeorgMafer.2(1966)1684(lnRussian).. .
7. T YOSHIDA. A. SERIKYAKU and M. YOSHIMATSU.
[
Fukuoka Daigaku Rigaku Shuho 16 (1986) 81 (in Japanese). '
8. Y. OSADA. H. NAKA YAMA. M. SHIN DO. T. ODAKA "
and Y. OGATA. 1. Electrochem. Soc. 126 (1979)3l. J
9. B. w JONG. G. J. SLAVENS and D E. TRAUT. ~
1. 'jHaler. Sci. 27 (1992) 6086. ,
10. P. S GOPALAKRISHNAN and P. S. LAKSHMINARA- i,
S I M H A~l. J. Marer. Sci. Leu. 12 (1993) 1422.
Received 7April
and accepted 20 July 1994
JOUI
:.--
TOI
tar
WA
Insti
WA
ApP
Rep
The
spu
De'
hay
for
ato
do~
exl
mo
lar:
n-i
col
eXI
Ui
pn
bu
Isr
ha
ff(
su
ca
wi
[3
ar
h(
pi
tl
al
0
Ii
d
1
C
I]
5
~