Production of SiO and Si(3P) atom in the reaction of silane with O(1D)

Article abstract of DOI:10.1021/jp9910237

The mechanism of SiO production in the reaction of O(¹D) with SiH₄ is reinvestigated, where O(₁D) is produced by the 193 nm photolysis of N₂O and nascent distribution of SiO (v -0-8) is determined by using a laser-induced fluorescence (LIF) detection system. Improvement of the reliability of the data is attained by using a simple tunable optical parametric oscillator system, since this system can cover the progression of the LIF spectrum over a wide range of wavelengths within a single scan. Nascent vibrational distribution of SiO is approximately expressed by a Boltzmann distribution of about T_v = 5200 K. Also, the Si(³P) atom is identified as a direct product of this reaction, although the branching fraction is estimated to be only 3 x 10^-4 of the total products. The mechanism of the formation of SiO is discussed by comparing the observed nascent vibrational distribution with that calculated by a quasi-static theory for the multiple-step unimolecular decomposition of chemically activated silanol.

Full text of DOI:10.1021/jp9910237

J. Phys. Chem. A 1999, 103, 11315-11320
11315
Production of SiO and Si(3P) Atom in the Reaction of Silane with O(1D)
Atsuko Takahara,† Atsumu Tezaki,* and Hiroyuki Matsui
Department of Mechanical Engineering, The UniVersity of Tokyo, 7-3-1 Hongo,
Bunkyo-ku, Tokyo 113-8656, Japan
ReceiVed: March 24, 1999; In Final Form: October 11, 1999
The mechanism of SiO production in the reaction of O(1D) with SiH4 is reinvestigated, where O(1D) is produced
by the 193 nm photolysis of N2O and nascent distribution of SiO (V ) 0-8) is determined by using a laser-
induced fluorescence (LIF) detection system. Improvement of the reliability of the data is attained by using
a simple tunable optical parametric oscillator system, since this system can cover the progression of the LIF
spectrum over a wide range of wavelengths within a single scan. Nascent vibrational distribution of SiO is
approximately expressed by a Boltzmann distribution of about Tv ) 5200 K. Also, the Si(3P) atom is identified
as a direct product of this reaction, although the branching fraction is estimated to be only 3 × 10-4 of the
total products. The mechanism of the formation of SiO is discussed by comparing the observed nascent
vibrational distribution with that calculated by a quasi-static theory for the multiple-step unimolecular
decomposition of chemically activated silanol.
Introduction
angular distribution of the products 2H2 + SiO is close to
isotropic; thus, the lifetime is suggested to be longer than the
The chemical kinetics of silane and its derivatives have long
been studied extensively because of their importance in the
microelectronic fabrication processes and the safety problems
related to their hazardous nature.1-15 However, the accumulated
information on the rates and mechanisms for the reactions of
silicon-containing species still seems insufficient. In particular,
direct experimental studies of the reactions of species composed
of Si-H-O atomic systems have been scarce, since most of
the stable molecules of such a system are not easily available
at ambient conditions. The deficiency may be partly compen-
sated by using a chemical activation method; for example, the
unimolecular decomposition reaction of highly excited silanol
(SiH3OH) was examined in our previous study by using the
insertion reaction of O(1D) into a Si-H bond of SiH4.14 The
reaction of O(1D) with SiH4
rotational period. The experimental result supports that SiO is
produced by the unimolecular decomposition of SiH3OH, where
the chemical energy may be randomly partitioned in SiH3OH,
as well as in other reaction intermediates involved with SiO
formation. Chemically activated silanol produced in the reaction
1 may decompose consecutively to a variety of products such
as
O(1D) + SiH4
f (SiH3OH)*
f SiH3 + OH
f H3SiO + H
f H2SiOH + H
f SiH2 + H2O
f H2SiO + H2
f HSiOH + H2
f SiO + 2H2
∆H298K ) -236 kJ/mol
∆H298K ) -238 kJ/mol
∆H298K ) -365 kJ/mol
∆H298K ) -437 kJ/mol
∆H298K ) -571 kJ/mol
∆H298K ) -565 kJ/mol
∆H298K ) -575 kJ/mol
∆H298K ) -264 kJ/mol
(1a)
(1b)
(1c)
(1d)
(1e)
(1f)
(1g)
(1h)
O(1D) + SiH4 f products
(1)
was shown in our previous study to have many product channels,
and the branching fractions of H, OH, and SiO were determined
as 0.36, 0.24, and 0.07-0.13, respectively. Recently, a crossed
molecular beam study was conducted on the O(1D) + SiH4
reaction and it is demonstrated that H3SiO is observed with
backward angular distribution in the H + H3SiO reaction
channel and that the production of OH has an angular distribu-
tion of the forward scattering for the reaction channel of OH +
SiH3.15 Also, the nascent vibrational energy distribution mea-
sured in our previous work by using a laser-induced fluorescence
(LIF) method is found to be highly inverted, i.e., [OH(V′′ )
1)]/[OH(V′′ ) 0)] ) 1.52.14 Therefore, OH and H may be
produced by direct mechanisms such as abstraction and substitu-
tion reactions rather than the unimolecular decomposition of
silanol.
f Si + H2 + H2O
+M8 SiH3OH
∆H298K ) -755 kJ/mol
(1i)
where the heats of formation are mainly taken from the ab initio
calculations conducted by Zachariah et al.8 Other theoretical
studies also support the possibilities of the reactions 1a-1f.8-13
In this paper, the nascent vibrational distribution of SiO
produced in the reaction of SiH4 + O(1D) is reinvestigated by
using an improved system of the LIF measurement in order to
examine the dynamics of this unique reaction because previous
results on the vibrational energy distribution of SiO by using a
dye laser system had a large scatter due to the difficulty of
The production of SiO has been also confirmed in the crossed
molecular beam experiment. In contrast to H and OH, the
† Present address: Institut fu¨r Physicalische Chemie, Universita¨t Go¨t-
tingen, Tammannstr. 6, D-37077 Go¨ttingen, Germany.
10.1021/jp9910237 CCC: $18.00 © 1999 American Chemical Society
Published on Web 12/03/1999

11316 J. Phys. Chem. A, Vol. 103, No. 51, 1999
Takahara, et al.
conducting successive scans over a wide range of wavelengths.
An accurate measurement for the nascent distribution is essential
for the examination of the reaction mechanism. A comparison
with the prior distribution based on the conventional statistical
theory is also performed for the multiple-step unimolecular
decomposition of energized silanol as a standard procedure. In
addition, another unexpected product, Si(3P) atom by reaction
channel 1h is also confirmed in this study. The product
branching fractions and the reaction pathways for Si production
are discussed.
Experimental Section
An experiment using a laser-photolysis pump and an LIF
probe technique was conducted in a quasi-static flow cell. An
optical parametric oscillator (OPO; Continuum, Surelite OPO)
was used as the probe light source. This OPO laser was pumped
by a frequency-tripled Nd:YAG laser (Continuum, Surelite-II
YAG) to generate tunable outputs of “signal” at 420-700 nm
and “idler” at 720-2300 nm at the same time. The visible OPO
“signal” was further frequency-doubled with a BBO crystal to
observe vibrational distributions of SiO in the 230-285 nm
region. For scanning the wavelength, the doubling crystal and
the OPO crystal were synchronously rotated by a pair of
homemade PC-controlled mechanical devices. The OPO system
used in this study had no dispersing optics such as a grating;
therefore, the line width of the visible output was rather wide
(typically 50 cm-1 according to the manufacturer’s specifica-
tion). However, the line width of the UV probe was reduced (4
cm-1 at 230 nm to 7 cm-1 at 300 nm) because of nonlinearity
and the phase-matching condition for the doubling crystal. This
width is still quite wide compared to that of typical dye lasers
or narrowband OPOs. Nevertheless, it has sufficient resolution
to discriminate most of the vibrational bands of SiO. This setup
has a merit that one can quickly scan the output over a wide
range of wavelengths without replacing the optical parts, or
without adjusting the optical alignment, so that spectra of
reproducible intensities are easily obtained within a relatively
short time.
Figure 1. LIF spectra observed in the ArF laser photolysis of 5 mTorr
of SiH4 and 10 mTorr of N2O mixture diluted in 10 Torr of He. A
frequency-doubled OPO laser is used for the upper trace, showing
(A1Π-X1Σ+) (V′-V′′) bands of vibrationally hot SiO. A frequency-
doubled dye laser of relatively high resolution is used for the lower
trace in the frame, showing Si(3PJ) lines superimposed over the (1′-
3′′) band of SiO.
Mixtures of N2O and SiH4 diluted in He were flowed in a
cell (23 mm i.d.) and were irradiated by 193 nm ArF excimer
laser (Lambda Physik, Compex 102) pulses coaxially with the
flow, where N2O is photolyzed to produce O(1D). A typical
fluence of the photolysis pulse was 10 mJ/cm2 at the observation
point. Fluorescence from SiO was transmitted through a selected
UV band-pass filter, detected by a photomultiplier tube, and
then acquired by a boxcar integrator. Assuming a plug-flow
condition, the gas residence time between the inlet and the
observation point was estimated to be 80 ms at a typical overall
flow rate of 200 sccm and 10 Torr total pressure, while the
lasers were operated at 10 Hz repetition. All the experiments
were carried out at room temperature (295 ( 3 K).
Figure 2. Time profiles of SiO in the vibrational states 0-5 from the
photolysis of 5 mTorr of SiH4/10 mTorr of N2O mixture in 10 Torr of
He. Signal intensities are scaled as observed and only roughly represent
the vibrational populations.
concentration of SiH4 and then gives the overall rate constant
for reaction 1 by using a pseudo-first-order analysis. It is found
that the rate constants determined in this way lie in the range
(2.6-3.0) × 10-10 cm3 molecule-1 s-1 for all the observed
vibrational levels. Consequently, it is essentially independent
of the vibrational level and agrees very well with the consump-
tion rate of O(1D) measured by using a vacuum ultraviolet
(VUV) LIF method in our previous study. It is thus reconfirmed
that SiO is a kinetically direct product of reaction 1. From the
time dependence measurements, vibrational relaxation of SiO
is found to be much slower. Nevertheless, the coupling of the
production of SiO and successive relaxation is taken into account
in the present analyses for determination of the vibrational
distribution at t ) 0. The results for the state-specific vibrational
relaxation rate of SiO will be reported in a separate publication.16
Integrated intensities of the observed SiO vibrational bands
are converted into the relative vibrational population by using
Results and Discussion
A. Nascent Vibrational Distribution of SiO. An example
of the LIF spectrum demonstrating production of vibrationally
excited SiO in the reaction of SiH4 + O(1D) is shown in Figure
1, where A-X bands of v′′ ) 0-8 are assigned unambiguously.
In this example, the probe delay was fixed at 20 µs after the
photolysis; 75% of O(1D) is consumed at this time by reactions
with SiH4 and N2O under the employed reactant concentrations.
Time-dependent traces of SiO at the vibrational levels (0-
5) are demonstrated in Figure 2. As shown in this figure, the
rapid rise of the signal intensity corresponds to the production
process of SiO. The initial rise rate is plotted against the

Production of SiO and Si(3P)
J. Phys. Chem. A, Vol. 103, No. 51, 1999 11317
Figure 3. Observed and calculated nascent vibrational distributions
of SiO produced in the SiH4 + O(1D) reaction. Solid circles are
experimental data, and a solid line is a linear regression fit representing
a vibrational temperature of 5200 K. Open symbols and dashed lines
are results of model calculations described in the text.
reported SiO (A-X) Franck-Condon factors,17 with corrections
for the wavelength dependence of the photomultiplier sensitivity
and the transmittance of the filters. Since some bands are
overlapped by weaker ones of different transitions, small
contributions from the minor peaks are corrected using an
iteration procedure until the observed intensity ratios are
reproduced by the nascent vibrational distribution given as a
guess. As shown in Figure 3, the evaluated nascent vibrational
distribution analyzed in this way is approximately given by the
Boltzmann distribution with a vibrational temperature Tv ) 5200
( 660 K (2 × standard deviation). This result is substantially
higher than the estimates of the previous study at 3900 K14
obtained by using a dye laser source for the LIF system. It is
clear that the experimental error in estimating the vibrational
population has been reduced in the present study.
Figure 4. Time profiles of Si atom LIF at J ) 2′-1′′ transition in the
photolysis of SiH4/N2O mixtures: (a) at fixed [SiH4] (3 mTorr) and
indicated [N2O]; (b) at fixed [N2O] (10 mTorr) and indicated [SiH4].
Smooth lines are calculated with eq 5.
If Si(3P) is a direct product of SiH4 + O(1D), generation and
consumption of Si initiated by the photolysis of N2O can be
described as
N2O + hν f O(1D) + N2
(0)
(1)
B. Detection of Triplet and Singlet Silicon Atom. The
ground-state Si(3PJ) was observed by 4s(3PJ′)-3p2(3PJ′′) transi-
tions lying at 250-253 nm for ArF laser irradiation into N2O/
SiH4 mixture as shown in the inset in Figure 1. Here, the Si
atom was detected by using a dye laser probe system having
relatively higher resolution than the OPO system.
O(1D) + SiH4 f φSi + other products
O(1D) + N2O f products
Si + SiH4 f products
Si + N2O f products
(2)
(3)
(4)
To confirm that the observed Si(3P) atom is a direct product
of the SiH4 + O(1D) reaction, the time dependence of the Si-
(3P) atom was monitored and kinetic analyses for the dependence
on the concentrations of SiH4 and N2O were performed. The
difference of the time dependence between spin-orbit angular
quanta J ) 0-2 was not observed because of the rapid energy-
transfer process among these energy levels. The most intense
(2′-1′′) line at 250.7 nm was mainly used for the kinetic
examination. It is also noted that this line has no interference
from SiO(V′-V′′) bands.
Since O(1D) is a minor species relative to SiH4 and N2O, a
pseudo-first-order analysis gives the time dependence of the
concentration of Si atom as follows:
exp(-kOt) - exp(-kSt)
[Si] ) φk1[SiH4]σ[N2O]
(5)
kS - kO
The profile of the Si(3P) atom has rise and decay components
and is dependent on both SiH4 and N2O concentrations as shown
in parts a and b of Figure 4, respectively. The profile is fitted
to the following double-exponential function:
where kO is the rate of O(1D) consumption, kO ) k1[SiH4] +
k2[N2O], kS is the rate of Si consumption, kS ) k3[SiH4] + k4-
[N2O], and σ is the photolysis yield of O(1D) from N2O. Because
of the symmetric nature of eq 5 in terms of kO and kS, the
corresponding order with kr and kd can be altered according to
the sign of (kS - kO). The present observations are accounted
for when (kS - kO) is assumed to be positive; that is, kS and kO
correlate to the rise and the decay parts, respectively. The slope
of kd against [SiH4], i.e., 3.0 × 10-10 cm3 molecule-1 s-1, is
X ) X0[exp(-kdt) - exp(-krt)]
where the rise rate kr is inevitably larger than the decay rate kd.
The rates thus obtained are plotted against the concentrations
of SiH4 and N2O in parts a and b of Figure 5, respectively.

11318 J. Phys. Chem. A, Vol. 103, No. 51, 1999
Takahara, et al.
(3P) is also estimated by comparing the integrated fluorescence
intensity of Si with that of the SiO(1′-3′′) band monitored in
the same run as shown in Figure 1. The relative production yield
of Si(3P) against SiO is evaluated to be 2.3 × 10-3, where
consumption of Si by the reactions with SiH4 and N2O, as well
as the measured vibrational distribution of SiO(V), is taken into
account in the analysis of the integrated intensity of the spectra.
It should be noted that the magnitude of the oscillator strength
of the A-X transition of SiO has not been completely fixed
yet. Here, fSiO ) 0.103 from a theoretical evaluation of
Oddershede and Smith24 is employed to evaluate the branching
fraction for SiO, since this is consistent with the fluorescence
lifetime of SiO(A) measured directly by Suzuki.25 This choice
gives the product branching fraction of SiO as 0.13, which is
in good agreement with the estimation in the crossed beam
experiment.15 Consequently, the branching fraction of Si
production in the reaction of O(1D) + SiH4 is given as 3 ×
10-4
.
Si(1D) can be also a direct product of the reaction 1, since it
does not violate the total spin conservation during the reaction.
In this case, the observed Si(3P) atom can be produced as a
result of the quenching of the initial product Si(1D). As a
consequence, Si(1D) was detected by the 4s(1PO1)-3p2(1D2)
transition at 288.16 nm in the N2O/SiH4 mixture after the
irradiation by the 193 nm pulse. However, this was not ascribed
to a product of the title reaction because the rise of Si(1D) was
found to be within 1 µs, i.e., much faster than predicted from
the collisional process and the experimental time profile, and
is much inconsistent with the calculated one for the reactions
1-4. Si(1D) was supposed to be generated by the photolysis of
some residual species including Si in the cell. The amount of
Si(1D) is also estimated to be an order of magnitude less than
that of the Si(3P) atom by using the same procedure as above.
Therefore, the quenching of observed Si(1D) is concluded to
give no influence on the time dependence of Si(3P).
Figure 5. Pseudo-first-order rise and decay rates in the double-
exponential fit of the Si time profiles: (a) as a function of N2O
concentrations and (b) as a function of SiH4 concentrations. Also shown
with open squares in (a) are the decay rates of O(1D) observed in the
former study.14
TABLE 1: Reaction Rate Constants Relevant to Production
and Consumption of Si Atom in the Photolysis of SiH4/N2O
Mixture
rate constant
temperature
(K)
reaction
(10-10 cm3 molecule-1 s-1
)
reference
The branching fractions of the products of reaction 1 are
summarized in Table 2, together with our previous experimental
results.
1 O(1D) + SiH4
3.0 ( 1.0
3.0 ( 0.3
1.1 ( 0.2
1.16 ( 0.35
1.16
4.4 ( 1.0
3.5 ( 1.0
2.0 ( 0.75
0.82 ( 0.41
1.9 ( 0.2
295 ( 3
295 ( 3
295 ( 3
200-300
200-350
295 ( 3
320
295 ( 3
350
300
this work
Okuda7
2 O(1D) + N2O
this work
DeMore19
Atkinson20
this work
Tanaka18
this work
Swearengen21
Husain22
C. Analysis of Multiple-Step Decomposition Process of
Activated Silanol. From the above discussion, it may be
reasonable to assume that SiO and Si(3P) are produced through
the decomposition of silanol formed by the insertion of O(1D)
into a Si-H bond of SiH4. The insertion mechanism of O(1D)
has been supported by former studies26,27 for its reactions with
hydrocarbons and other compounds. Owing to the large excess
energy, various product channels are energetically allowed, as
is already mentioned. However, the reaction mechanism of SiO
formation is probably not straightforward, but multiple-step
decomposition channels such as
3 Si + SiH4
4 Si + N2O
thus considered to be the rate constant k1, since it is exactly the
same as the consumption rate of O(1D) by the reaction with
SiH4, as shown in Figure 5a and in Table 1. Similarly, the
dependence of kr against [SiH4] gives k3 ) 4.3 × 10-10 cm3
molecule-1 s-1, which is in good agreement with the value that
was measured in a plasma reactor.18
(H3SiOH)* f TS1 f HSiOH + H2 f TS2 f SiO + 2H2
(A)
In addition, from the slopes of kd and kr against [N2O], k2
and k4 are determined as 1.2 × 10-10 and 2.0 × 10-10 cm3
(H3SiOH)* f TS3 f H2SiO + H2 f TS4 f SiO + 2H2
molecule-1 s-1, respectively. They are in reasonable agreement
(B)
19,20
with literature values for k2
and k4.21,22 Also, as shown by
(H3SiOH)* f TS3 f H2SiO + H2
the solid curves in parts a and b of Figure 4, the observed
dependence of the relative signal intensity on the initial
concentrations of SiH4 and N2O is completely reproduced with
eq 5. It seems reasonable, therefore, to consider that Si(3P) is a
direct product of reaction 1. The rate constants of reactions 1-4
determined in this study are summarized in Table 1, together
with those of previous studies for comparison.
f TS5f HSiOH + H2 f TS2 f SiO + 2H2
(C)
may be responsible. These reaction pathways are schematically
shown in the energy diagram in Figure 6. If the excited
intermediates, i.e., H3SiOH, H2SiO, and HSiOH have suf-
ficiently long lifetimes to randomize the excess energy, the prior
energy distributions of fragments are calculated by using the
By use of the oscillator strengths of Si 23 and SiO 24 and the
Franck-Condon factor for SiO,17 the fractional yield for Si-

Production of SiO and Si(3P)
J. Phys. Chem. A, Vol. 103, No. 51, 1999 11319
TABLE 2: Product Branching Fractions in Reaction O(1D) + SiH4
product
fraction
methodology
reference
SiO
Si(3P)
OH
H
0.07-0.13a
3 × 10-4
0.36
spectral intensity, SiO(2′-0”) vs NO(0′-0′′)
spectral intensity, Si(3P) vs SiO(1′-3′′)
Okuda7
this work
Okuda7
Okuda7
relative yield against reference reaction O(1D) + H2
0.24
a 0.13 is recommended in this paper.
Figure 6. Energy level diagram showing product branching pathways of the SiH4 + O(1D) reaction. Energies of species and transition states are
mainly referenced to the ab initio study of Zachariah and Tsang.8
TABLE 3: Reaction Rates and Branching Fractions for Channels Forming SiO + 2H2 by the BIM/RRKM Model for
Multiple-Step Unimolecular Decomposition of Activated Silanol
channel A
channel B
channel C
first step decomposition
kr a (s-1
H3SiOH f TS1 f HSiOH + H2
1.0 × 1012
H3SiOH f TS3 f H2SiO + H2
)
3.7 × 1011
second step
HSiOH f TS2 f SiO + H2
5.4 × 109
H2SiO f TS4 f SiO + H2
2.2 × 109
H2SiO f TS5 f HSiOH
kr a (s-1
)
5.1 × 109
third step
HSiOH f TS2 f SiO + H2
Tv(SiO)b (K)
5700
0.72
2700
0.06
5600
0.22
fraction in overall SiO
a kr is the microcanonical RRKM reaction rate. b Calculated by BIM.
conventional statistical theory. In this study, the internal energies
of the reaction intermediates and products are calculated by
using the barrier impulsive model (BIM).28 In this model, it is
assumed that the reactant internal energy is statistically parti-
tioned at the transition state (TS) for all the degrees of freedom;
however, the potential energy released after passing the TS is
simply given to the intermolecular translational motion and
products rotations. When enough vibrational energy is parti-
tioned into the second intermediate, HSiOH in the reaction
channel A for example, the successive decomposition forming
SiO + H2 is energetically allowed. The energy distribution in
SiO and H2 is evaluated by using the first-step product energy
distribution. The statistical calculations can be conducted in a
standard way using given vibrational frequencies and moments
of inertia for the intermediates, products, and transition states.8
RRKM calculations are also conducted to estimate the rates
and the branching ratios for these possible reaction channels
by using the same information for the molecular structures and
the energies of the transition states.
The calculated results of the product energy distributions and
branching fractions for the assumed channels are summarized
in Table 3. Calculated vibrational distributions of SiO assuming
channels A, B, and C are compared with the measured nascent
distribution in Figure 2. It is found that both channels A and C
give slightly larger vibrational energies, but they are in
reasonable agreement with that of present experimental result,
since the difference is almost within the limit of uncertainty of
the experiment. On the other hand, channel B gives much lower
vibrational energy because of the higher exit barrier. It seems
impossible to discriminate channels A and C from only the
observed vibrational distribution; however, channel A (HSiOH
intermediate) is more favorable for the production of SiO by
the consideration of relative rates based on the RRKM calcula-
tions.
For the production of the Si(3P) atom, H2 and H2O are
supposed to be the counterparts of the product. Hence, two
distinctly different channels, i.e., H2O is produced prior to H2
or vice versa, can be expressed as follows:
(H3SiOH)* f SiH2 + H2O f Si + H2 + H2O (D)
(H3SiOH)* f HSiOH (or H2SiO) + H2 f Si + H2O + H2
(E)
As is indicated, Si(3P) is a product of the reaction 1; then
crossing from singlet to triplet surface should be included in
these reaction pathways. Shock tube studies have already shown
that Si(3P) + H2 is a product channel in the thermal decomposi-
tion of SiH2 via intersystem crossing without having too high
a barrier compared to the endothermicity of the reaction.29-31
The spin conservation rule is generally moderated for systems
including a heavy atom like Si.
By application of the same BIM procedure, the product
branching fractions for Si(3P) is estimated for both channels D

11320 J. Phys. Chem. A, Vol. 103, No. 51, 1999
Takahara, et al.
and E as 1.6 × 10-2 and 5 × 10-4, respectively, where the
second-step decomposition of HSiOH and SiH2 is treated as
barrierless and the effect of the intersystem crossing is disre-
garded.
Although it is impossible to judge which channel is dominant
for production of the Si atom, it implies that the production of
Si at this small yield is by no means a mystery for the
unimolecular decomposition of highly excited SiH3OH.
sity: Germany, 1996.
(7) Koda, S.; Suga, S.; Tsuchiya, S.; Suzuki, T.; Yamada, C.; Hirota,
E. Chem. Phys. Lett. 1989, 161, 35.
(8) Zachariah, E. R.; Tsang, W. J. Phys. Chem. 1995, 99, 5308.
(9) Allendorf, M. D.; Melius, C. F.; Ho, P.; Zachariah, E. R. J. Phys.
Chem. 1995, 99, 15285-15293.
(10) Gordon, M. S.; Pederson, L. A. J. Phys. Chem. 1990, 94, 5527.
(11) Darling, C. L.; Schlegel, H. B. J. Phys. Chem. 1993, 97, 8207.
(12) Kudo, T.; Nagase, S. J. Phys. Chem. 1984, 88, 2833.
(13) Tachibana, A.; Sakata, K. Appl. Surf. Sci. 1997, 117, 151.
(14) Okuda, K.; Yunoki, K.; Oguchi, T.; Murakami, Y.; Tezaki, A.;
Koshi, M.; Matsui, H. J. Phys. Chem. A 1997, 101, 2365.
(15) (a) Lin, J. J.; Lee, Y. T.; Yang, X. Abstract of Symposium on
Chemical Kinetics and Dynamics, Okazaki, Japan, 1999; p 34. (b) Lin, J.
J.; Lee, Y. T.; Yang, X. Private communication.
(16) Takahara, A.; Tezaki, A.; Matsui, H. Bull. Chem. Soc. Jpn.,
submitted.
(17) Liszt, H. S.; Smith, W. H. J. Quant. Spectrosc. Radiat. Transfer
1972, 12, 947.
(18) Tanaka, T.; Hiramatsu, M.; Nawaka, M.; Kono, A.; Goto, T. J.
Phys. D: Appl. Phys. 1994, 27, 1660.
(19) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Molina, M. J.;
Hampson, R. F.; Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R. Chemical
kinetics and photochemical data for use in stratospheric modeling.
EValuation number 9; JPL Publication 90-1; Jet Propulsion Laboratory,
Pasadena, CA, 1990.
(20) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Troe, J. J. Phys. Chem. Ref. Data 1992, 21, 1125.
(21) Swearengen, P. M.; Davis, S. J.; Niemczyk, T. M. Chem. Phys.
Lett. 1978, 55, 274.
(22) Husain, D.; Norris, P. E. J. Chem. Soc., Faraday Trans. 2 1978,
74, 93.
Conclusions
Through our previous and current studies, the reaction of SiH4
+ O(1D) has been carefully examined. Chemically activated
SiH3OH decomposes into various fragment species, and then
some fraction of the first fragment of silanol further proceeds
to the second decomposition channels. It may be reasonable to
conclude that by this mechanism SiO is finally formed through
a multiple-step unimolecular decomposition of the intermediates
such as HSiOH, where the reaction intermediates associated with
this reaction channel have a sufficiently long lifetime to
distribute their internal energy statistically.
The reaction mechanism described in the current study is
useful also in understanding the oxidation mechanism of silanes.
Within our search, no experimental study of the reaction kinetics
of SiH3OH has been performed. Direct examination of the
reaction mechanism for SiH3OH thus will provide a very useful
contribution to the improvement of the structure of integrated
mechanisms,5,6 which have been constructed on the basis of
indirect information or theoretical calculations.
(23) CRC Handbook of Chemistry and Physics, 76th ed.; CRC Press:
Boca Raton, FL, 1995.
(24) Oddershede, J.; Elander, N. J. Chem. Phys. 1976, 65, 3495.
(25) Suzuki, G. Unpublished results in National Institute for Environ-
mental Studies, Tsukuba, Japan. (The zero-pressure lifetime of SiO(A; V )
0) has been determined to be 30 ns by a direct observation of the
fluorescence decay.)
(26) Wiesenfeld, J. R. Acc. Chem. Res. 1982, 15, 110.
(27) Hack, W.; Tiesemann, H. J. Phys. Chem. 1995, 99, 17364.
(28) North, S. W.; Blank, D. A.; Gezelter, J. D.; Longfellow, C. A.;
Lee, Y. T. J. Chem. Phys. 1995, 102, 4447.
(29) Francisco, J. S.; Barnes, R.; Thoman, J. W., Jr. J. Chem. Phys.
1988, 88, 2334.
(30) NoorBatcha, I.; Raff, L. M.; Thompson, D. L.; Viswanathan, R. J.
Chem. Phys. 1986, 84, 4341.
References and Notes
(1) Jasinski, J. M.; Gates, S. M. Acc. Chem. Res. 1991, 24, 9.
(2) Jasinski, J. M.; Meyerson, B. S.; Scott, B. A. Annu. ReV. Phys.
Chem. 1987, 28, 109.
(3) Kondo, S.; Tokuhashi, K.; Nagai, H.; Iwasaka, M.; Kaise, M.
Combust. Flame 1995, 101, 170.
(4) Hartman, J. R.; Famil-Ghiriha, J.; Ring, M. A.; O’Neal, H. E.
Combust. Flame 1987, 68, 43.
(5) Babushok, V. I.; Tsang, W.; Burgess, D. R., Jr.; Zachariah, M. R.
Proceedings of the Twenty-SeVenth Symposium (International) on Combus-
tion; The Combustion Institute, Pittsburgh, 1998; p 2431.
(6) Koshi, M.; Murakami, Y.; Matsui, H. Modeling of Chemical
Reaction Systems, CD-ROM Workshop Proceedings; Heidelberg Univer-
(31) Votintsev, V. N.; Zaslonko, I. S.; Mikheev, V. S.; Smirnov, V. N.
Kinet. Catal. 1987, 27, 843.