Formation of impurity Si2OH6 in silane synthesized from silicon tetrafluoride

Article abstract of DOI:10.1134/S0036023611040061

The possibility of the reduction of hexafluorodisiloxane by calcium hydride in the synthesis of silane from silicon tetrafluoride has been studied. This reaction is shown to be not decisive for oxygen contamination of silane. The most likely reason for the appearance of impurity Si₂OH₆ in "fluoride" silane is the Ca(OH)₂-catalyzed reaction of silane with trace water. The concentration of impurity Si₂OH ₆ in silane at the stage of synthesis may be efficiently decreased by the preliminary purging of calcium hydride with a hydrogen (grade A) flow.

Full text of DOI:10.1134/S0036023611040061

ISSN 0036ꢀ0236, Russian Journal of Inorganic Chemistry, 2011, Vol. 56, No. 4, pp. 510–512. © Pleiades Publishing, Ltd., 2011.
Original Russian Text © A.D. Bulanov, P.G. Sennikov, A.Yu. Sozin, A.Yu. Lashkov, O.Yu. Troshin, 2011, published in Zhurnal Neorganicheskoi Khimii, 2011, Vol. 56, No. 4,
pp. 550–552.
SYNTHESIS AND PROPERTIES
OF INORGANIC COMPOUNDS
Formation of Impurity Si2OH6 in Silane Synthesized
from Silicon Tetrafluoride
A. D. Bulanov, P. G. Sennikov, A. Yu. Sozin, A. Yu. Lashkov, and O. Yu. Troshin
Institute of Chemistry of High Purity Substances, Russian Academy of Sciences,
ul. Tropinina 49, Nizhni Novgorod, 603950 Russia
eꢀmail: bulanov@ihps.nnov.ru
Received December 22, 2009
Abstract—The possibility of the reduction of hexafluorodisiloxane by calcium hydride in the synthesis of
silane from silicon tetrafluoride has been studied. This reaction is shown to be not decisive for oxygen conꢀ
tamination of silane. The most likely reason for the appearance of impurity Si2OH6 in “fluoride” silane is the
Ca(OH)2ꢀcatalyzed reaction of silane with trace water. The concentration of impurity Si2OH6 in silane at the
stage of synthesis may be efficiently decreased by the preliminary purging of calcium hydride with a hydrogen
(grade A) flow.
DOI: 10.1134/S0036023611040061
Oxygen is a limited admixture in silicon used for
A specific feature of “fluoride” silane is the presꢀ
the production of semiconductor devices, along with ence of impurity fluorosilanes SiHxF4 – x resulting from
boron, phosphorus, and carbon. Its concentration the incomplete reduction of SiF4 according to reaction
must not exceed 1015 at/cm3. Oxygen is also one of the (4) at a level of 10–3 mol %. By analogy with chlorosiꢀ
limited impurities in the synthesis of a highꢀpurity lanes, these admixtures can be hydrolized by trace
28
structurally perfect monoisotopic Si single crystal water to give disiloxane, for example, by the reaction
used as the basis in the manufacturing of the standard
mass. Interstitial oxygen is known to reduce the denꢀ
sity of a silicon crystal [1].
2SiH3F + H2O
Si2OH6 + 2HF.
(5)
On the other hand, hexafluorodisiloxane Si2F6O
,
whose concentration may reach several percents, is the
major molecular admixture in silicon tetrafluoride [9].
It may be supposed that the reduction of hexafluoroꢀ
disiloxane proceeds simultaneously with reactions (4)
and (5) as follows:
Impurity oxygen in polycrystalline silicon and films
manufactured from silane may appear from various
sources. It is known [2] that silane reacts with oxygen
to form suspended submicronꢀsized silicon dioxide
particles [3, 4] according to the reaction
Si2OF6 + 3CaH2
Si2OH6 + 3CaF2,
(6)
SiH4 + 2O2
SiO2 + 2H2O.
(1)
thus making a certain contribution to the oxygen conꢀ
tamination of “fluoride” silane.
The objective of our work is to verify this supposiꢀ
tion.
Thermodynamic and quantumꢀchemical calculaꢀ
tions [5] indicate the possibility of a reaction between
silane and water at least in two stages as follows:
SiH4 + Н2O
SiН3ОН + H2,
(2)
(3)
SiН3ОН + SiH4
SiН3–О–SiН3 + H2.
EXPERIMENTAL
However, the kinetic restrictions of reactions (2)
and (3), which are poorly studied in gas phase [6],
should be taken into account in practice. The hydrolꢀ
ysis of monosilane is actively catalyzed by alkalies and
wellꢀdeveloped surfaces (for example, of storage cylinꢀ
ders) [7]. Reactions (1)–(3) may be considered as uniꢀ
versal sources of impurity oxygen in silane indepenꢀ
dently of the method of its synthesis.
The reaction of silicon tetrafluoride with calcium
hydride was studied at room temperature. Frionally
distilled silicon tetrafluoride of natural isotopic comꢀ
position was used. According to IR Fourier spectrosꢀ
copy, it contained (2.0 0.6)
10–1 mol % of impurity
×
Si2F6O. Calcium hydride was synthesized by the
hydrogenation of distilled metal calcium swarf with
hydrogen purified on a palladium filter. Mechanically
dispersed calcium hydride inside a box purged by dry
nitrogen was placed into a 200ꢀcm3 stainless steel reacꢀ
tor. Then, gaseous silicon tetrafluoride was admitted
to the reactor, the reactor was attached directly to the
In this work we consider silane synthesized from
silicon tetrafluoride by its reduction with calcium
hydride [8] by the reaction
SiF4 + 2CaH2 = SiH4 + 2CaF2.
(4)
510

FORMATION OF IMPURITY Si2OH6
511
Table
с
,
mol %
Sample no.
concentration in SiH4 (CMS)
concentration in SiF4 (IR)
Si2OН6
Si2OН5F
Si3O2H7F
–
1
2
3
4
5
6
0.20 0.06
0.01 0.005
0.10 0.05
1.50 0.75
0.30 0.15
8.0 4.0
1.1 0.4
1.2 0.3
–
(8 3)
(3 1)
(3 1)
×
×
×
10–2
10–4
10–4
(1.5 0.5)
×
×
×
10–2
10–5
10–5
(1.5 0.6)
×
×
×
10–2
10–1
10–4
(4 2)
(6 2)
(1.3 0.4)
(9.7 1.0)
–
–
0.9 0.3
(5 2)
×
10–2
(8 3)
×
10–3
inflow system of an MIꢀ1309 mass spectrometer
equipped with an electron impact ion source, and
vapor phase samples were taken directly into the ionꢀ
ization chamber. The concentration of impurity
hexafluorodisiloxane was measured in arbitrary units
RESULTS AND DISCUSSION
As determined by massꢀspectrometric analysis, the
concentration of impurity hexafluorodisiloxane
remains constant when fractionally distilled silicon
tetrafluoride is in contact with calcium hydride at
room temperature for one month. Hence, it has been
established that hexafluorodisiloxane does not react
with calcium hydride at room temperature and is staꢀ
ble in gas phase in a mixture with SiF4, in agreement
with [15].
28
+
as the percent ratio of the intensities of base (
Si F4 ,
28 16
+
m
/
e
= 104) and impurity (
Si2 О F6 ,
m/e
= 186)
molecular ions. The ion source had an ionizing voltage
of 50 V, collector current of 1.5 mA, and accelerating
voltage of 1.5 kV. Silicon tetrafluoride samples from
the reactor were periodically analyzed during one
month.
Both the IR spectroscopy data [16] and CMS studꢀ
ies performed in this work show that hexafluorodisiꢀ
The silicon tetrafluoride samples used to synthesize loxane, along with disilane, is the major molecular
silane had different impurity hexafluorodisiloxane admixture in silane synthesized by the reduction of
concentrations: 99.99% monoisotopic 28SiF4 (sample 1, SiF4 (table). It should be noted that impurity Si2H5FO
NTTs Tsentrotekh, St. Petersburg), 99.99% SiF4 of and Si3H7FO2 were also revealed by CMS in the synꢀ
natural isotopic composition (sample 2, ZAO Astor, thesized silane samples, but their concentrations were
St. Petersburg), 99.99% monoisotopic 28SiF4 (sample 3), one order of magnitude lower than that of Si2Н6O . In
99.9% monoisotopic 29SiF4 (sample 4), 99.9% principle, the presence of these admixtures in silane
monoisotopic 30SiF4 (sample 5, OAO PO EKhZ, Zeleꢀ may result from both reaction (6) and hydrolysis of fluꢀ
nogorsk).
orosilanes.
The heavy dump fraction (sample 6) obtained after
The table shows that the hexafluorodisiloxane conꢀ
recovering 28Si isotopes from a natural mixture by cenꢀ
trifugation was also used for the synthesis of silane.
According to the data of massꢀspectrometry, this fracꢀ
tion had the following isotopic composition, %: Si28,
58.8; Si29, 24.7; Si30, 16.5. The impurity hexafluoroꢀ
disiloxane concentrations determined for these samꢀ
ples by IR Fourierꢀtransform spectroscopy are given in
the table.
centrations in samples 1, 2, and 6 differ by three orders
of magnitude. However, the same table shows that the
Si2Н6O concentration in silane synthesized from these
samples according to the method [10] is at a level of
1.0%. Therefore, it might be thought that, even if
hexafluorodisiloxane enters reaction (6) in the considꢀ
ered system at 250–300 С, this reaction is not decisive
°
for the oxygen contamination of “fluoride” silane.
Silane synthesis from initial samples 1, 2 and 6 was
carried out according to the method described in [10,
11] by passing a flow of silicon tetrafluoride mixed
with hydrogen (grade A) through a calcium hydride
bed. Before synthesis of silane from samples 3, 4 and 5,
calcium hydride and the entire laboratory unit were
purged by a hydrogen (grade A) flow for 16, 8, and 24 h,
respectively.
Reactions (2) and (3) are the most probable source
of impurity Si2OH6 in “fluoride” silane. The process
may be catalyzed by Ca(OH)2 formed by the reaction
of residual water with calcium hydride [17, 18]. The
possibility of reaction (5) or analogous reactions with
other fluorosilanes likewise cannot be ruled out comꢀ
pletely. The thermodynamics and kinetics of the first
two stages of hydrolysis of fluorosilanes are close to
The impurity composition of synthesized silane those for SiF4 owing to the presence of the Si–F bond
was studied by IR Fourierꢀtransform spectroscopy in their molecules. The hydrolysis of an SiF4 molecule
[12] and chromatography/mass spectrometry (CMS) is thermodynamically forbidden [5], but at the same
[13, 14].
time, the activation barrier to the formation of Si2F6O
RUSSIAN JOURNAL OF INORGANIC CHEMISTRY Vol. 56 No. 4 2011

512
BULANOV et al.
2. E. P. Belov, E. N. Lebedev, Yu. P. Grigorash, et al.,
Monosilane and Semiconductor Materials Technologies
(NIITEKhim, Moscow, 1989) [in Russian].
3. G. G. Devyatykh, V. V. Balabanov, V. I. Zvereva, et al.,
Dokl. Akad. Nauk SSSR 230 (2), 342 (1976).
4. T. Ohmi, M. Nakamura, A. Ohki, et al., J. Electroꢀ
is much lower than for reaction (3) [19]. On the whole,
this matter requires deeper studies.
From the analysis of samples 3–5, it can be seen
that the preliminary purging of calcium hydride and
the entire laboratory unit with a hydrogen (grade A)
flow efficiently decreases the content of residual water,
and the more prolonged is the purging, the lower the
chem. Soc. 139 (9), 2654 (1992).
5. P. G. Sennikov, S. K. Ignatov, A. E. Sadov, et al.,
Zh. Neorg. Khim. 54 (2), 287 (2009) [Russ. J. Inorg.
Chem. 54 (2), 252 (2009)].
6. L. A. Chuprov, S. K. Ignatov, P. G. Sennikov, et al., Proꢀ
ceedings of XII Conference “High Purity Chemicals and
Materials. Preparation, Analysis, Applications" (Nizhni
Novgorod, 2004), p. 223 [in Russian].
concentration of impurity Si2OH6
.
In summary, the possibility of the reduction of
hexafluorodisiloxane with calcium hydride in the synꢀ
thesis of silane from SiF4 has been studied. This reacꢀ
tion is shown to be not determinative for the oxygen
contamination of silane. The reaction of SiF4 with
trace water, probably catalyzed by Ca(OH)2, is most
likely to cause the appearance of impurity Si2OH6 in
“fluoride” silane. However, the hydrolysis of impurity
fluorosilanes cannot be ruled out completely. It has
been established that the preliminary purging of calꢀ
cium hydride and the entire laboratory unit with a
hydrogen (grade A) flow efficiently decreases the conꢀ
centration of impurity Si2OH6 in silane at the stage of
synthesis and excludes the need for deep purification
from this lowꢀvolatile admixture by cryofiltration.
7. H. J. Klockner and M. Eschwey, Chem. Ing. Tech. 60
(11), 815 (1988).
8. A. D. Bulanov, V. V. Balabanov, D. A. Pryakhin, and
O. Yu. Troshin, Neorg. Mater. 38 (3), 356 (2002).
9. P. G. Sennikov, M. A. Ikrin, S. K. Ignatov, et al., Izv.
Akad. Nauk, Ser. Khim., No. 48, 93 (1999).
10. G. G. Devyatykh, E. M. Dianov, A. D. Bulanov, et al.,
Dokl. Akad. Nauk 391 (5), 638 (2003).
11. A. D. Bulanov, V. S. Mikheev, O. Yu. Troshin, and
A. Yu. Lashkov, Zh. Neorg. Khim. 53 (1), 11 (2008)
[Russ. J. Inorg. Chem. 53 (1), 6 (2008)].
12. L. A. Chuprov, P. G. Sennikov, K. G. Tokhadze, et al.,
ACKNOWLEDGMENTS
Neorg. Mater. 42 (8), 1017 (2006).
We are grateful to V.A. Krylov, D.A. Pryakhin, and
L.A. Chuprov for performing chromatographic/mass
spectrometric, mass spectrometric, and IR spectroꢀ
scopic analyses.
13. V. A. Krylov, O. Yu. Chernova, and A. Yu. Sozin, Massꢀ
Spektrometriya
14. V. A. Krylov, A. Yu. Sozin, A. D. Zorin, et al., Massꢀ
Spektrometriya (4), 281 (2008).
4 (2), 125 (2007).
5
This work was supported by the Presidential Grant
for the State Support of Leading Scientific Schools in
the Russian Federation, grant no. NShꢀ4701.2008.3,
and by the Presidium of the Russian Academy of Sciꢀ
ences through program no. 20, para. 2.
15. J. L. Margrave, K. G. Sharp, and P. W. Wilson, J. Am.
Chem. Soc. 92 (6), 1530 (1970).
16. P. G. Sennikov, A. P. Kotkov, S. A. Adamchik, et al.,
Neorg. Mater. 46 (4), 415 (2010).
17. J.ꢀP. Tessier, P. Palau, J. Huot, et al., J. Alloys Compd.
376, 180 (2004).
18. M. Hirscher, M. Becher, M. Haluska, et al., J. Alloys
REFERENCES
Compd. 356–357, 433 (2003).
1. A Report to the Avogadro Working Group of the CCM 19. S. K. Ignatov, P. G. Sennikov, B. S. Ault, et al., J. Phys.
(Kenichi Fujii, Washington, DC, 1998). Chem. 103, 8328 (1999).
RUSSIAN JOURNAL OF INORGANIC CHEMISTRY Vol. 56 No. 4 2011