Shock tube study on the reaction of Si atoms with CH3 with respect to SiC formation

Article abstract of DOI:10.1524/zpch.2001.215.6.811

The reaction kinetics of ground state Si atoms was studied behind reflected shock waves in the presence of excess CH₄ and CH₃, respectively. The very fast Si atom formation by the thermal decomposition of SiH₄ and Si₂H₆ was used as a reliable Si atom source at temperatures T > 1550K. CH₃ was formed by the thermal decomposition of different source gases like CH₃Cl and CH₃I. All test gases were highly diluted in argon. The atomic resonance absorption spectroscopy (ARAS) was applied for time-resolved measurements of Si and H atoms and it was found that the reaction of Si atoms with CH₄ (R1) is much slower than the reaction of Si atoms with CH₃ (R₂) which seems to be an important reaction leading to the formation of very reactive intermediate species on its way to form SiC. The product channels are discussed and a temperature independent mean rate coefficient for (R2) was determined: Si + CH₃ k2a ? SiCH + H₂ (R2a) k2a ? SiCH₂ + H (R2b) with k_2b/k_2a < 0.1 and a mean value for k₂ = 2.0±0.69- 10¹⁴cm³mol^-1s^-1. A reaction mechanism including the established GRI (Gas Research Institute) mechanism is presented which explains all measured Si and H concentration profiles obtained during this study. by Oldenbourg Wissenschaftsverlag, Muenchen.

Full text of DOI:10.1524/zpch.2001.215.6.811

Zeitschrift für Physikalische Chemie, 215, 6, 811–824 (2001)
by Oldenbourg Wissenschaftsverlag, München
Shock Tube Study on the Reaction of Si Atoms
with CH3 with Respect to SiC Formation
∗
By A. Kunz, K. A. Bhaskaran and R. Roth
Institut für Verbrennung und Gasdynamik, Gerhard-Mercator-Universität Duisburg,
D-47048 Duisburg, Germany
Dedicated to Prof. Dr. Dr. h.c. mult. Jürgen Troe
on the occasion of his 60th birthday
(Received July 11, 2000; accepted September 8, 2000)
Shock Tube / Chemical Kinetics / Rate Coefficient / Silicon /
Methyl Radical / SiC
The reaction kinetics of ground state Si atoms was studied behind reflected shock waves
in the presence of excess CH4 and CH3, respectively. The very fast Si atom formation
by the thermal decomposition of SiH4 and Si2H6 was used as a reliable Si atom
source at temperatures T > 1550 K. CH3 was formed by the thermal decomposition of
different source gases like CH3Cl and CH3I. All test gases were highly diluted in argon.
The atomic resonance absorption spectroscopy (ARAS) was applied for time-resolved
measurements of Si and H atoms and it was found that the reaction of Si atoms with
CH4 (R1) is much slower than the reaction of Si atoms with CH3 (R2) which seems to
be an important reaction leading to the formation of very reactive intermediate species on
its way to form SiC. The product channels are discussed and a temperature independent
mean rate coefficient for (R2) was determined:
k
2a
Si+CH3
SiCH+H2
SiCH2 +H
(R2a)
(R2b)
k
2b
with k2b/k2a < 0.1 and a mean value for k2 = 2.0 0.69·1014 cm3 mol−1 s−1. A reaction
mechanism including the established GRI (Gas Research Institute) mechanism is presented
which explains all measured Si and H concentration profiles obtained during this study.
Introduction
The significance of silicon carbide as one of the most important ceramics
is well recognised as also the persisting interest in optimizing the proper-
*
Corresponding author. E-mail: roth@ivg.uni-duisburg.de

812
A. Kunz, K. A. Bhaskaran and P. Roth
ties of the final bulk material to meet highest demands. The production of
silicon carbide nano-powder with characteristics like highest purity and qual-
ity, high specific surface area, and high sintering quality is therefore of great
importance. The most established SiC production process is still the energy
intensive Atcheson process [1], where SiO2 reacts with carbon to form SiC
and the resulting CO, which has to separated in a rather expensive process.
Different ways of Chemical Vapour Deposition (CVD) processes are avail-
able for very high purity SiC thin film production. For SiC nano-powder
production of highest purity the gas-phase synthesis in high-temperature reac-
tors became the focus of interest during the last decade. A general overview
of the synthesis of particles in high temperature gas-phase reactors is given
by Wooldridge [2] and Pratsinis [3] has presented the flame aerosol pro-
cesses. Many studies about different SiC synthesis routes were published in
the last years. Such processes are partly combined with RF, microwave or
laser energy inputs. Different source gases have been used like SiH4/CH4,
SiH4/C2H4, Si2H6/CH4, SiCl4/CH4, CH3SiCl3, (CH3)2SiHCl, (CH3)2SiCl2,
(CH3)4Si, and Si(CH3)2(C2H5O)2. To understand and to optimize such pro-
cesses, reliable kinetic and thermodyamic data of important gas-phase reac-
tions are necessary.
In many studies where silanes or chlorosilanes and methane are used as
source gases, it is mentioned that with higher temperatures higher SiC yields
can be obtained. Some of these workers have assumed homogeneous gas-
phase reactions of gaseous silicon as the favoured reaction route to form
SiC. In view of the higher stability of methane compared to those of the
silanes, these deliberations are acceptable. Tamou et al. [4] have synthe-
sized SiC powder from SiCl4/CH4 using H2 as a quenching gas in a ther-
mal plasma CVD reactor. According to equilibrium calculations they sug-
gested that the reaction Si(g) + CHx (x = 0...4) is responsible for the for-
mation of SiC. Guo et al. [5] have synthesized SiC powder by introducing
pure silicon powder into a carrier gas consisting of CH4/Ar or CH4 only,
which is injected into an induction plasma in hot and cold wall reactors, re-
spectively. They also strongly suggested the reaction of gaseous silicon with
decomposition products of methane as the predominant mechanism for SiC
formation.
A comprehensive study of the reaction of Si atoms with several hydrocar-
bons was performed by Basu and Husain [6, 7]. They have studied reactions
of ground state Si atoms with different acetylenes and olefins (e.g. C2H2,
C2H4, C3H6, C4H8) in a slow-flow reactor at room temperature. Si atoms were
generated by pulsed irradiation of SiCl4 and an atomic resonance absorption
system was used for the Si detection. Differences between the absorption de-
cay caused by the reaction of Si with SiCl4 [8, 9] and a faster decay obtained
when acetylenes and olefins are added were analysed and reported in form
of rate coefficients. They have also studied the reaction of Si with CH4 [10]
and reported an upper limit for the rate coefficient at room temperature to

Shock Tube Study on the Reaction of Si Atoms with CH3 with Respect to ...
813
be k1 ≤ 6.0·109 cm3 mol−1 s−1 which was much slower compared to reactions
with olefines or acetylenes which have rate coefficient values almost equal to
the collision frequency.
In this study the formation and consumption of silicon atom concentra-
tions by added CH4 and CH3 at high temperatures was measured behind re-
flected shock waves. An ultra-high vacuum(UHV)-shock tube reactor equipped
with highly sensitive atomic resonance absorption spectroscopy (ARAS) is
used to detect time resolved Si- and H-atom concentrations in highly diluted
SiH4/Si2H6/CH4/CH3Cl/CH3I/Ar mixtures. The paper describes the experi-
mental technique and discusses the kinetic results based on a reaction model,
which can predict all measured signals over a wide range of mixtures composi-
tions and temperatures.
Experimental
The experiments were carried out behind reflected shock waves in a stainless
steel shock tube of 80 mm internal diameter. The driver section was 3.5 m, the
driven section 6.0 m in length, which was specially prepared for ultra-high-
vacuum (UHV) requirements. The driven section can be baked out and pumped
down to pressures below 2·10−8 mbar by a turbo molecular pump. The gas
mixtures used were prepared manometrically in a stainless steel UHV cylin-
der which can be pumped down by the combination of a chemical diaphragm
pump and a turbo-molecular drag pump. The leak and/or outgassing rate was
in the order of (1 0.5)×10−7 mbar s−1 for both driven section and vessel.
The gases and liquids used were of highest commercial purity: Ar ≥ 99.9999%,
Si2H6 ≥ 99.9%, SiH4 ≥ 99.999%, CH3Cl ≥ 99.999% and CH3I > 99.99%.
A more detailed description of the experimental procedure can be found else-
where (Refs. [11] and [12]).
The ARAS technique used to observe Si-atom absorption at the Si-triplet
*
3
[4s 3P2 ← 3p2 P2] at λ = 251.6 nm consists of a pulsed Si-hollow cathode
lamp, the shock tube absorption path, which is located 15 mm ahead of the
end plate, a monochromator (0.25 m Jarrell Ash), and a photomultiplier (Burle
IP28). Perpendicular to the Si-detection system, a second ARAS diagnostic
system is arranged for simultaneously monitoring H atoms. H atoms were de-
tected at Lyman alpha λ = 121.6 nm. The system consists of a microwave
excited discharge lamp, a 0.5 m McPherson VUV monochromator, and a so-
lar blind photomultiplier (Hamamatzu Typ R1459). The microwave lamp was
operated with a flowing gas mixture of 0.1% H2 in He maintained at a con-
stant pressure of 6 mbar and a microwave power of about 50 W. The spectral
resolutions are 0.5 nm for the Si- and 0.6 nm for the H-diagnostic systems.
The spectral shape of both the Si- and H-atom resonance lines are not
known precisely due to self absorption and self reversal in the light sources.
Therefore, series of shock wave calibration experiments have been performed

814
A. Kunz, K. A. Bhaskaran and P. Roth
to relate the measured absorptions to the appropriate concentrations. The Si-
atom calibration is based on the well known high-temperature dissociation of
0.05 to 2.0 ppm SiH4 highly diluted in argon, see Mick et al. [13]. H atom cali-
bration was realized by H atom measurements during the thermal decompos-
ition of 2 ppm N2O/200 ppm H2/Ar mixtures at 1300 K ≤ T ≤ 1900 K. The
signals were compared with computer simulations, see Roth and Just [14, 15],
and Masten et al. [16]. The relation between measured absorption and cor-
responding concentration can be expressed by the modified Lambert–Beer
equation:
ꢀ
ꢀ
ꢃ ꢃ
n
ꢁ
ꢂ
σλ,X
cm3
[X]
Aλ,X = 1−exp − l ×
×
,
X = Si, H
(1)
cm−3
where l is the length of the absorption path, λ the wave length and n the
concentration exponent. The individual calibration parameters were: nSi =
0.95, σSi = 4.0·10−13 cm2; nH = 0.56, σH = 1.31·10−8 cm2. Eq. (1) fits the
calibration measurements for Si quite well in the concentration range be-
tween 2.0·1011 and 4.0·1012 atoms cm−3, and for H between 3.0·1011 and
1.0·1013 atoms cm−3. The uncertainties of Eq. (1) are about 15% for each
species.
Results and discussion
In a first experimental series Si atom concentration profiles were measured be-
hind reflected shock waves in mixtures containing SiH4 or Si2H6 and a large
excess of CH4 highly diluted in argon. The detailed experimental conditions
of all experiments of this study are listed in Table 1. Fig. 1 shows an example
for both an undisturbed Si absorption profile of a shock heated SiH4/argon
mixture and its perturbation by addition of 200 ppm CH4 at three different tem-
peratures. Time t = 0 is fixed to the shock heat-up which leads to an almost
instantaneous formation of Si atoms indicated by the absorption step from 0
to about 90%. With additon of CH4 a signal decrease was measured after the
Table 1. Experimental conditions.
Test gas mixture in Ar
T [K]
p [bar]
0.25–1.0 ppm Si2H6/25–1000 ppm CH4
0.5–1.0 ppm SiH4/200–500 ppm CH4
1550−1970
1740−2130
1.5 0.15
0.6−1.4
0.5–2 ppm Si2H6/20 ppm CH3Cl
0.25–0.5 ppm SiH4/2.5–5.0 ppm CH3Cl
1580−2320
1850−2600
1.4 0.1
1.25 0.1
0.25 ppm Si2H6/5.0 ppm CH3I
0.5–1.0 ppm SiH4/5.0–50 ppm CH3I
1550−2050
1790−2580
1.4 0.7
1.3 0.15

Shock Tube Study on the Reaction of Si Atoms with CH3 with Respect to ...
815
Fig. 1. Absorption at λ = 251.6 nm behind reflected shock waves in a SiH4/Argon mixture
without and with the addition of CH4 at three different experimental temperatures.
maximum was reached and the decay became faster with increasing tempera-
ture. The s-formed shape of the profiles indicated an induction time for the
consumption of Si atoms which is not explainable either by the uncertainty
or lower sensitivity of the Si calibration relation at higher or lower absorp-
tion ranges. It signalizes that not CH4 but a secondary reaction seems to be
responsible for the consumption of Si atoms.
According to simulation calculations using the established GRI mechan-
ism [17], CH4 decomposes slowly and forms CH3 as a major product within
our temperature and pressure ranges. Therefore, a second series of experiments
was performed in which the fast decomposition of either CH3Cl → CH3 +
Cl (R20) [18] or CH3I → CH3 + I (R29) [19–21] were used as fast CH3
sources. The main difference between the previous experiments was, that only
a few ppm of CH3Cl or CH3I, i.e. almost two orders of magnitudes less than
CH4, were sufficient to reach a similar Si signal decay compared to the first
experimental series. The measurements showed a strong dependence on the
added CH3Cl and CH3I concentrations but no dependence on the experimental
temperature. The apparent induction time found in the first series has disap-
peared, too. Fig. 2 shows measured Si absorption profiles when 5 ppm CH3Cl
and 10 ppm CH3I were added to a 0.5 ppm SiH4/Ar mixture. After a fast in-
crease of the absorption within 20 s caused by the fast formation of Si atoms,
the signal tends to decrease over the next 300 s. The ratios between the Si
precursor concentrations (SiH4, Si2H6) and the perturbing CH3 source gas
concentrations (CH3Cl and CH3I) of all investigated gas mixtures were kept
at least at 1 : 10 and above. Calculations using the GRI mechanism showed
that for almost all our experiments the amount of CH3 could be treated in
a first approximation to be constant within the period of the fast signal de-
cay (T = 1700 K: 15% loss in t = 1 ms, T = 2500 K: 20% loss in t = 0.2 ms).

816
A. Kunz, K. A. Bhaskaran and P. Roth
Fig. 2. Example of two absorption profiles by Si atoms which are measured in a SiH4/Ar
mixture with a few ppm CH3I and CH3Cl added. Insert: first-order plots of both absorp-
tion signals.
This justifies a pseudo-first-order analysis as a first step to determine a rate
coefficient k21st for the reaction of Si atoms with CH3.
k
2
Si+CH3 ꢀ products
(R2)
All measured Si absorption profiles were transfered into first-order plots,
ꢄ
ꢅ
where ln (− ln (1− ASi))(1/n) is plotted against the reaction time. The result-
ing curves of the measurements of Fig. 2 are presented in the upper right corner.
Their slope is equal to the inverse time constant τ−1
:
ꢄ
ꢅ
d
τ−1 = − ln (− ln (1− ASi))(1/n)
,
(2)
dt
where ASi is the Si-atom absorption. Rate coefficients k21st of reaction (R2) can
directly be determined from the time constants τ−1
.
τ−1
k21st
=
(3)
[CH3]0
with [CH3]0 being the initial CH3 concentration of a 100% conversion
from CH3I and CH3Cl, respectiveley. The individual results are shown as
open circles in Fig. 3. The rate coefficients are temperature-independent
within our investigated temperature range. The mean-value is k21st = 2.19
0.7·1014 cm3 mol−1 s−1 where the uncertainty is given at 1σ-level.
Possible reaction pathways are determined by comparing reaction en-
thalpies calculated using the thermodynamic data of the CHEMKIN data-
base [22] containing ab-initio calculations from Allendorf and Melius [23] for

Shock Tube Study on the Reaction of Si Atoms with CH3 with Respect to ...
817
Fig. 3. Arrhenius plot of the rate coefficient k21st and k2sim for the reaction of Si atoms
with CH3.
the Si-C-H system although they admit very high uncertainties in the data for
molecules like HSiC and SiCH.
k
2a
Si+CH3 ꢀ SiCH +H2
H2098 = −75.8 kJ/mol
H2098 = −68.2 kJ/mol
H2098 = 174.5 kJ/mol
H2098 = 177.3 kJ/mol
(R2a)
(R2b)
(R2c)
(R2d)
k
2b
ꢀ SiCH2 +H
k
2c
ꢀ SiH+CH2
k
2d
ꢀ HSiC +H2
Reaction pathways (R2a) and (R2b) are exothermic; reaction enthalpies of
other product channels than (R2c) and (R2d) are much more endothermic and
are not listed here. (R2a) and (R2b) are the favoured channels because of the
temperature independent k21st that we obtained. To distinguish the pathway of
reaction (R2), simultaneous H atom concentration measurements were per-
formed. An example is presented in Fig. 4, where both, Si and H profiles are
measured simultaneously in a SiH4/CH3Cl/Ar mixture. For comparison, H
profiles measured in a 2.5 ppm CH3Cl/Ar mixture (dots) and in a 0.25 ppm
SiH4/Ar mixture (triangles) are also shown. The H concentration shows an
explicitly slower increase compared to the Si concentration decay thereby
indicating the dominance of channel (R2a) compared to the H producing
channel (R2b).
It was observed that at higher temperatures and higher concentrations of
CH3Cl and CH3I the Si-signal decay did not end at zero concentration lev-
els as can be seen in Fig. 5. The signal continues after a fast decrease with
a second slower increase which follows then again by a slow signal decay
within our experimental measurement time. This behaviour was measured at

818
A. Kunz, K. A. Bhaskaran and P. Roth
Fig. 4. Simultaneously measured Si- and H-atom concentration profiles compared with
simulation results using reaction channel (R2a) and (R2b). H concentrations measured in
CH3Cl/Ar (dots) and SiH4/Ar (triangles) mixtures are included.
Fig. 5. Upper part: Si concentrations measured at four increasing temperatures in a 1 ppm
SiH4/50 ppm CH4I/Ar mixture. Results of calculations obtained with the reaction mech-
anism of Table 2 are included. Lower part: Appropriate sensitivity analysis for Si atoms of
the experiments at T = 2250 K with R54: CH3 +CH3(+M) C2H6(+M).

Shock Tube Study on the Reaction of Si Atoms with CH3 with Respect to ...
819
Table 2. Simplified reaction mechanism, ki = A× T n ×exp(−TA/T ) cm3 mol−1 s−1
.
Reaction
Rate Coefficient
Ref.
A
n
TA
1
2
3
Si+CH4
SiCH2 +H2
< 1.0·1012
2.0·1014
7.0·1012
1.0·1015
1.2·1014
5.3·1013
< 4.0·1012
2.0·1014
t.s.
t.s.
as.
as.
[7]
[6]
t.s.
as.
Si+CH3
products
SiCH+H → Si+CH2
1.15
11930
−3
Si+CH2 → SiCH+H
4
Si+C2H2
Si+C2H4
SiCH+CH3
Si+CH2
SiC2 +H2
SiH2 +C2H2
Si+C2H4
SiC+H2
5
6
7
8
9
10
Si2H6
SiH4
SiH4 +SiH2
5.2·1010
1.9·1010
9.1·1013
16850
22550
15100
[24]
[13]
[13]
SiH2 +H2
SiH2 +Ar Si+H2 +Ar
11
12
13
14
15
16
17
18
CH3Cl CH3 +Cl
1.4·1015
3.4·1013
2.2·1013
9.0·1013
1.3·1013
7.4·1013
2.2·1014
4.0·1013
41373
1425
4152
44010
1700
23810
48300
23840
[18]
[18]
[25]
Cl+CH3Cl HCl+CH2Cl
CH3 +Cl CH2 +HCl
HCl+Ar H+Cl+Ar
H+HCl Cl+H2
[26]
[26,27]
[28,29]
[30]
HCl+Cl H+Cl2
H2 +Ar H+H+Ar
Cl2 +Ar Cl+Cl+Ar
[28,31]
19
20
21
CH3I+Ar → CH3 +I+Ar
H2 +I → HI+H
2.3·1015
1.7·1014
8.0·1014
19845
16940
18700
[21,19]
[28]
[28]
HI+I → H+I2
22–69 GRI – Mechanism 3.0
[17]
temperatures above T = 2000 K in mixtures containing more than a tenfold
excess of CH3I and CH3Cl, respectively. It was also observed in mixtures
with an excess of CH4. The beginning of this tendency at the lower tem-
perature limit can be seen in Fig. 1 at T = 2070 K. In case of CH3Cl the
second increase was less pronounced. After the fast Si decay the signal ends
in an almost steady state and was then followed by a slow decrease. Ex-
periments with only CH4/CH3Cl/CH3I/argon mixtures were performed at
similar conditions to verify if secondary species, formed during the decompos-
ition process, absorb at λ = 251.6 nm. No absorption was measured in these
experiments.
Based on all these observations a reaction mechanism is proposed (see
Table 2) containing the GRI mechanism [17], but without oxygen and nitro-
gen containing species and reactions, and the decomposition mechanisms of
SiH4 and Si2H6 [13, 24]. Dissociation reactions for CH3I [21] and CH3Cl [18]
are added as well as reactions with Cl and HCl for completition. A reaction

820
A. Kunz, K. A. Bhaskaran and P. Roth
Si+CH4 ꢀ SiCH2 +H2 (R1a) was included based on thermodynamic consid-
erations:
k
1a
Si+CH4 ꢀ SiCH2 +H2
H2098 = −65.5 kJ/mol
H2098 = 153.6 kJ/mol
H2098 = 154.1 kJ/mol
H2098 = 282.4 kJ/mol
(R1a)
(R1b)
(R1c)
(R1d)
k
1b
ꢀ SiH+CH3
k
1c
ꢀ SiCH3 +H
k
1d
ꢀ SiH2 +CH2
Reaction (R2a) as well as reactions Si + C2H2 and Si + C2H4 suggested by
Basu and Husain [6, 7] were added using energetically favoured product
channels.
The Si decay of all CH3I and CH3Cl experiments as well as all experiments
with CH4 could be simulated with this mechanism. This confirms on one hand
the assumption of a slow reaction of Si with CH4 compared to the fast one of
Si with CH3. On the other hand it validates the GRI mechanism for the CH4
decomposition and the CH3 formation, respectiveley. The resulting k2sim are also
included in Fig. 3 and a mean rate coefficient k2 containing all k21st and k2sim for
the reaction
k
2
Si+CH3 ꢀ products
can be given as:
(R2)
k2 = 2.00 0.69·1014 cm3 mol−1 s−1
.
Fig. 6 shows the good agreement between a calculated and measured Si pro-
file in a shock heated SiH4/CH4/Ar mixture. The result of a sensitivity analysis
with respect to Si is presented in the lower part of this figure. The sensitiv-
ity coefficients SSi are plotted over time and the strong influence of reaction
(R2) compared to (R1) becomes obvious. A maximum value can be given for
the reaction of Si with CH4 to be k1 < 1·1012 cm3 mol−1 s−1 which is the high-
est value up to which the Si profiles can not be corrected by variations of k2
within the error limit. Simulated Si and H atom profiles are shown together
with the experimental results in Fig. 4. The calculations were performed by
switching on either reaction (R2a) or (R2b). Whichever channel is switched
on, it is seen that the result for the Si concentration matches the experimental
value almost perfectly, whereas the H concentration is very much overpredicted
with the (R2b) channel on. On the other hand the profile fits better with the
experimental curve, despite slight uncertainties, when the channel (R2a) alone
is allowed to proceed. This again reinforces that (R2a) is the dominant chan-
nel in the production of Si. Therefore, an upper limit can be given for the
branching ratio k2b/k2a < 0.17 by considering the uncertainties of the H-ARAS
calibration.

Shock Tube Study on the Reaction of Si Atoms with CH3 with Respect to ...
821
Fig. 6. Upper part: Comparison of a measured and calculated Si concentration profile
in a SiH4/CH4/Ar mixture. Lower part: Sensitivity analysis for Si atoms showing the
strong influence of reaction Si+CH3 (R2). R27: CH4(+M)
CH3 +H(+M), R28: H+
CH4
CH3 +H2.
The second formation of Si atoms (see. Fig. 5) can be either caused by
decomposition or bimolecular reactions of radicals formed by reaction (R2).
These possibilities were tried out and it was found that reactions SiCH +
Ar ꢀ SiC + H + Ar, H2098 = 417.2 kJ/mol, SiC + H ꢀ Si + CH, H2098
=
524.3 kJ/mol, and SiC +Ar ꢀ Si+C+Ar, H2098 = 447.7 kJ/mol could not
explain the measured profiles.
The best fit of all profiles was achieved by including the very simple overall
reaction of SiCH with H forming Si and CH2.
k
3
SiCH+H → Si+CH2
H2098 = 99.2 kJ/mol
(R3)
k3 = 7.0·1012 T1.15 exp(−11930 K/T ) cm3 mol−1 s−1
The backward reaction (−R3) is also considered and is set to a maximum
value of k = 1.0·1015 cm3 mol−1 s−1, but has no influence on the pre-
−3
sented results. A reaction Si +CH2 → SiC+H2 (R7) is included with a rate
coefficient similar to k2 but this had a negligible influence. The reaction
SiCH + CH3 → Si + C2H4 (R6) is included with a maximum rate coeffi-
cient of 4.0·1012 cm3 mol−1 s−1. Results at four different temperatures are

822
A. Kunz, K. A. Bhaskaran and P. Roth
shown in Fig. 5 which are achieved by k variations of the reactions (R2),
(R3), and (R6) within their error limits. Calculations with the above esti-
mated maximum branching ratio for (R2a) and (R2b) are found to result
in incorrect predictions for the simulated Si profiles when compared to ac-
tual measurements at higher temperatures and higher CH3 concentrations.
The additional H atoms formed by the fast reaction (R2b) varies the Si con-
centration in its shape and absolute value, which can not be balanced by
varying the value of k3. A reduction of the upper limit of the branching ratio
k2b/k2a from 0.17 to 0.10 leads to calculated profiles which fall within the
measurement uncertainties. The lower part of Fig. 5 shows the eight most
sensitive reactions for the experiment at T = 2250 K. Small uncertainties of
the H concentrations have an influence on the calculated Si concentration.
In case of those experiments with a high excess of CH3Cl such uncertain-
ties are conceivable due to reactions between Cl and HCl with other species
which are not known. The slightly different Si profile shapes in case of
CH3Cl, where the second Si increase is less strong, could easily be pre-
dicted with the mechanism by slight variations of the H concentrations by
a few percent.
Conclusions
The reaction of Si with CH3 is part of the gas phase chemistry of SiC synthe-
sis. The kinetics of this reaction was studied in a shock tube behind reflected
waves at temperatures 1500 ≤ T ≤ 2650 K and pressures 0.6 < p < 1.4 bar
using different CH3 and CH4 sources to perturb a constant Si concentration
obtained from the thermal decomposition of SiH4 or Si2H6. The measured
Si absorption decay was evaluated by a pseudo-first-order analysis. H atom
concentrations were measured simultaneously to determine the reaction path-
way. A reaction mechanism containing the GRI mechanism is proposed that
could predict all measured Si- and H-atom concentration profiles. A rate co-
k
2
efficient is given for the reaction Si+CH3 ꢀ products (R2) which was found
to be much faster than the rate coefficient for the reaction of Si + CH4 for
which an upper limit could be given. The product channel Si+CH3 ꢀ SiCH+
H2 (R2a) was found to be dominant over the channel Si + CH3 ꢀ SiCH2 +
H (R2b) and an upper limit for the branching ratio k2b/k2a < 0.1 could
be deduced.
Acknowledgement
The authors thank Mrs. N. Schlösser and Mr. L. Köppers for their help in per-
forming and evaluating the experiments. The financial support of the Deutsche
Forschungsgemeinschaft is gratefully acknowledged.

Shock Tube Study on the Reaction of Si Atoms with CH3 with Respect to ...
823
References
1. J. E. Bailey, M. Bohnet and J. Brinder, Ullmann’s Encyclopedia of Industrial Chem-
istry, Vol. 6E, Wiley, John & Son, Inc., New York, USA (1999).
2. M. S. Wooldridge, Gas-phase combustion synthesis of particles, Prog. Energy Com-
bust. Sci. 24 (1998) 63.
3. S. E. Pratsinis, Flame aerosol synthesis of ceramic powders, Prog. Energy Combust.
Sci. 24 (1998) 197.
4. Y. Tamou, T. Yoshida and K. Akashi, The synthesis of ultrafine silicon carbide in
a hybrid plasma, J. Jpn. Inst. Met. 51 (1987) 737.
5. J. Y. Guo, F. Gitzhofer and M. I. Boulos, Effects of process parameters on ultrafine
SiC synthesis using induction plasmas, Plasma Chem. Plasma Process. 17 (1997)
219.
6. S. C. Basu and D. Husain, Absolute rate data for the reaction of ground state atomic
silicon, Si (33 PJ ), with olefins determined by time-resolved atomic resonance ab-
sorption spectroscopy, Z. Phys. Chem. (Munich) 158 (1988) 1.
7. S. C. Basu and D. Husain, Measurement of absolute rate data for the reactions of
ground state atomic silicon Si (33 Pj) with acetylenes by time-resolved atomic res-
onance absorption spectroscopy, J. Photochem. Photobiol. 42 (1988) 1367.
8. D. Husain and P. E. Norris, Kinetic study of ground state silicon atoms,
Si[3p2( 3PJ )], by atomic absorption spectroscopy, J. Chem. Soc., Faraday Trans. 2
74 (1978) 93.
9. A. Kunz and P. Roth, Shock tube study of the reaction of Si atoms with SiCl4,
Ber. Bunsenges. Phys. Chem. 102 (1998) 1492.
10. D. Husain and P. E. Norris, Kinetic study of reactions of ground state silicon atoms,
Si[3p2( 3PJ )], by atomic absorption spectroscopy, J. Chem. Soc., Faraday Trans. 2
74 (1978) 106.
11. K. Thielen and P. Roth, Resonance absorption measurements of N and O atoms
in high temperature NO dissociation and formation kinetics. Symp. (Int.) Comb.
[Proc.] 20 (1984) 685.
12. D. Woiki and P. Roth, A shock tube study on the thermal decomposition of CS2
based on S( 3P) and S(1D) concentration measurements, Shock Waves 4 (1994) 95.
13. H. J. Mick, V. N. Smirnov and P. Roth, ARAS measurements on the thermal decom-
position of silane, Ber. Bunsenges. Phys. Chem. 97(6) (1993) 793.
14. P. Roth and Th. Just, Messungen zur Hochtemperaturpyrolyse von Äthan. Ber. Bun-
senges. Phys. Chem. 83 (1979) 577.
15. Th. Just, Atomic Resonance Absorption Spectrometry in Shock Tubes, Marcel
Dekker Inc., New York and Basel (1981).
16. D. A. Masten, R. K. Hanson and C. T. Bowman, Shock tube study of the reaction
H+O2 → OH+O using OH laser absorption, J. Phys. Chem 94 (1990) 7119.
17. G. P. Smith, D. M. Golden, M. Frenklach, N. W. Moriary, B. Eiteneer, M. Golden-
berg, C. T. Bowman, R. K. Hanson, S. Song, W. C. Gardiner Jr., V. V. Lissianki and
Z. Qin, GRI MECH 3.0, University of California at Berkeley, Dep. Mech. Eng.,
http://www.me.berkeley.edu/gri_mech/ (2000).
18. K. P. Lim and J. V. Michael, The thermal decomposition of CH3Cl using the
Cl-atom absorption method and the bimolecular rate constant for O +CH3
(1609–2002 K) with a pyrolysis photolysis-shock tube technique, J. Chem. Phys.
98(5) (1993) 3919.
19. S. S. Kumaran, M. C. Su and I. V. Y. Michael, Thermal decomposition of CH3I
using I-atom absorption, Int. J. Chem Kinet. 29 (1997) 535.
20. D. F. Davidson, M. D. Di Rosa, E. J. Chang and R. K. Hanson, An improved deter-
mination of the 216.615 nm absorption coefficient for methyl radicals, J. Quant.
Spectrosc., Radiat. Transfer 53 (1994) 581.

824
A. Kunz, K. A. Bhaskaran and P. Roth
21. K. Takahashi, A. Inoue and T. Inomata, Direct measurements of rate coefficients for
thermal decomposition of methyl halides using shock-tube ARAS technique, Proc.
20th Symp. on Shock Waves II (1995) 959.
22. R. J. Kee, F. M. Rupley and J. A. Miller, The chemkin thermodynamic data base,
Sandia Report SAND87-8215B, Sandia National Laboratories, April (1994).
23. M. D. Allendorf and C. F. Melius, Theoretical study of the thermochemistry of
molecules in the Si-C-H system, J. Phys. Chem 96 (1992) 428.
24. H. J. Mick, M. W. Markus, P. Roth and V. N. Smirnov, A shock tube study of the
thermal decomposition of Si2H6 based on Si and SiH2 measurements. Ber. Bunsen-
ges. Phys. Chem. 99(6) (1995) 880.
25. A. Goldbach, F. Temps and H. G. Wagner, Kinetics of the reactions of CH2(X3 B1)
with HCl and HBr. Ber. Bunsenges. Phys. Chem. 94 (1990) 141.
26. G. N. Schading and P. Roth, Thermal decomposition of HCl measured by ARAS
and IR diode laser spectroscopy, Combust. Flame 99 (1994) 467.
27. T. C. Allison, G. C. Lynch, D. G. Truhlar and M. S. Gordon, An improved potential
energy surface for the H2Cl system and its use for calculations of rate coefficients
and kinetic isotop effects, J. Phys. Chem. 100 (1996) 13575.
28. D. L. Baulch, J. Duxbury, S. J. Grant and D. C. Montague, Evaluated kinetic data for
high temperature reactions. Vol. 4, Homogeneous gas phase reactions of halogen-
and cyanide-containing species, J. Phys. Chem. Ref. Data 10, Suppl. 1 (1981).
29. B. C. Garrett and D. G. Truhlar, Generalized transition state theory. Canonical
variational calculations using the bond energy-bond order method for bimolecular
reactions of combustion products, J. Am. Chem. Soc. 101 (1976) 5207.
30. D. L. Baulch, D. D. Drysdale, D. G. Horne and A. C. Lloyd, Homogeneous gas
phase reactions of the H2-O2 system, Vol. 1 of Evaluated Kinetic Data for High
Temperature Reactions, Butterworth, London, 1. ed. (1972).
31. A. C. Lloyd, A critical review of the kinetics of the dissociation-recombination re-
actions of fluorine and chlorine, Int. J. Chem. Kinet. 3 (1971) 39.