Kinetic study of the chemiluminescence spectrum of rarified flame in dichlorosilane oxidation in the near IR region

Article abstract of DOI:10.1023/A:1012351026408

Electronically excited HO₂^. radicals (A²A′-X²A″), OH^. radicals (v = 2 - 0), and HCl molecules (v = 3 - 0) are identified using the emission spectra at 0.8-1.6 μm in the rarefied flame in dichlorosilane combustion at 293 K and low pressures. The spectrum also contains the composite bands of the H₂O (0.823 μm) and H₂O₂ (0.854 μm) molecule vibrations. The maximum intensity of emission of these species is attained behind the front of the active chemical transformation, and the equilibrium between the vibrational and translational degrees of freedom is established in the region of the regular thermal regime of cooling. SF₆ additives act as a reservoir that accumulates the vibrational energy in developed ignition. The processes responsible for the inhibition of dichlorosilane oxidation by SF₆ additives are considered.

Full text of DOI:10.1023/A:1012351026408

Kinetics and Catalysis, Vol. 42, No. 5, 2001, pp. 594–600. Translated from Kinetika i Kataliz, Vol. 42, No. 5, 2001, pp. 657–663.
Original Russian Text Copyright © 2001 by Chernysh, Rubtsov, Tsvetkov.
Kinetic Study of the Chemiluminescence Spectrum of Rarified
Flame in Dichlorosilane Oxidation in the Near IR Region
V. I. Chernysh, N. M. Rubtsov, and G. I. Tsvetkov
Institute of Structural Macrokinetics, Russian Academy of Sciences, Chernogolovka, Moscow oblast, 142432 Russia
Received May 12, 2000
.
.
2
2
Abstract—Electronically excited HO radicals (A A'–X A''), OH radicals (v= 2 – 0), and HCl molecules
2
(
v= 3 – 0) are identified using the emission spectra at 0.8–1.6 µm in the rarefied flame in dichlorosilane com-
bustion at 293 K and low pressures. The spectrum also contains the composite bands of the H O (0.823 µm)
2
and H O (0.854 µm) molecule vibrations. The maximum intensity of emission of these species is attained
2
2
behind the front of the active chemical transformation, and the equilibrium between the vibrational and trans-
lational degrees of freedom is established in the region of the regular thermal regime of cooling. SF additives
6
act as a reservoir that accumulates the vibrational energy in developed ignition. The processes responsible for
the inhibition of dichlorosilane oxidation by SF additives are considered.
6
INTRODUCTION
O mixtures [9]. The SF additives inhibit self-ignition
2 6
and retard flame propagation in the DCS + O mixtures
2
Investigation of the mechanism of the combustion
of silanes and their chlorine derivatives and the kinetics
of the oxidation of these compounds under various con-
ditions and at different gas-mixture compositions is of
great scientific and practical interest. These processes
are widely used in microelectronics to apply isolating
and passivating layers of silicon dioxide in the inte-
grated circuit of semiconductor instruments [1, 2].
Reacting with halogens, monosilane and its chlorine
derivatives form vibrationally excited hydrogen halides
[
16]. At the same time, several main steps of this pro-
cess, including chain propagation reactions, have not
been identified yet. The retardation of ignition by water
vapors and molecular nitrogen has not been explained
even at such low pressures either, when termolecular
reactions can be neglected [17].
The aim of this work was to determine the chemical
nature of vibrationally excited species in the rarefied
flame during DCS oxidation and to establish the mech-
anism of the inhibition of this reaction by the SF addi-
[
3, 4], which may be used as active media in infrared
6
tives.
chemical and electron-discharge lasers [5, 6]. The
development of the methods of synthesis of nanopow-
ders based on the processes involving inorganic
hydrides is of considerable interest [7, 8].
EXPERIMENTAL
The experiments were conducted in a vacuum setup
described elsewhere [14]. The reaction vessel was a
quartz cylinder (diameter, 120 mm; height, 120 mm).
The reactor was equipped with inlets for electrodes and
gas supply and optical quartz windows. A fuel mixture
prepared beforehand was fed from a by-pass volume
into the reactor through a vacuum valve. The reactor
Studies of the combustion of silanes and their deriv-
atives are also important for the theory of chemical
transformation. The oxidation of these compounds
occurs via the branched-chain mechanism [9, 10]. The
.
2
+
electronically excited radicals OH (A Σ ) [11],
.
.
1
1
1
1
3
1
SiO (A Π – X Σ), SiCl (A B – X A , a B – X A ) [12,
2
1
1
1
1
–4
was evacuated to 4 × 10 torr using 3NVR-1D forevac-
1
3], and chlorine atoms [10] were registered in the rari-
uum and N-01 oil-diffusion pumps. The residual pres-
sure was measured with VIT-2 thermocouple–ioniza-
tion andVDG-1 gas-discharge vacuum gages. The mix-
tures were prepared beforehand by admitting a required
fied flame during dichlorosilane (DCS) oxidation. The
IR spectrophotometric study [14] showed that the pro-
cess in oxygen at [DCS] < 50% and low pressures may
be described by the following overall equation:
amount of DCS into oxygen (containing SF additives
6
if necessary) through a narrow capillary. Both oxygen
SiH Cl + O
SiO + 2HCl.
2
2
2
2
and SF were of reagent grade. The DCS purity was
6
The final products of DCS combustion in air under ~98% as determined by IR spectroscopy [14]. The reac-
ordinary conditions over a wide range of initial DCS tion was initiated from outside and conducted under
concentrations also include H O, chlorosilanes, and steady-state conditions at room temperature and 2–5 torr.
2
molecular hydrogen [15]. Certain hydrocarbons (e.g., Mixtures of the following composition were used:
C H and C H ) inhibit flame propagation in the DCS + 23.5% DCS + O , 26.5% DCS + 7% SF + O , and
2
2
2
4
2
6
2
0
023-1584/01/4205-0594$25.00 © 2001 MAIK “Nauka /Interperiodica”

KINETIC STUDY OF THE CHEMILUMINESCENCE SPECTRUM
595
2
2% DCS + 18% SF + O . A coil of Nichrome wire
6 2
9
2
4
(
diameter, 0.3 mm) was placed in the middle of the
8
8
a
a
reactor for heating. Ignition was initiated by pulse heat-
ing of this coil with a condenser battery (capacitance,
3
7
1
3
*
000 µF). Chemiluminescence in the visible region
registered with a FEU-39 photomultiplier (spectral sen-
sitivity, 0.2–0.6 µm) was used to start the registration
system. An FEU signal was transmitted to the synchro-
nization inlet and one beam of an S9-8 two-beam mem-
ory oscilloscope starting in a leading-phase mode.
Emission during combustion was also registered at
11
5
6
1
0
Fig. 1. Optical scheme of the experimental setup: (1) photo-
diode, (2) outlet slit of monochromator, (3) monochromator,
4) inlet slit of monochromator, (5) light filter, (6) con-
0
.7−1.9 µm using an MDR-3 monochromator (diffrac-
(
tion lattice, 300 stroke/mm) and silicon and quartz light
filters. The spectral slit width was ~0.005 µm. The
kinetic curves of the integral emission in the visible
region and spectral emission upon single ignition were
registered simultaneously at a fixed wavelength every
denser, (7) collimator, (8) optical windows, (9) reaction ves-
sel, (10) spherical flame front, (11) reflecting mirror (Au),
and (a–a) visual field of the registration system.
RESULTS AND DISCUSSION
∆
λ = 0.005 µm. The monochromator was calibrated by
the characteristic absorption bands of chloroform at
Figure 2a presents the emission spectra in the near
IR region for the initiated self-ignition of the 23.5%
DCS + O (spectrum 1) and 22% DCS + 18% SF + O
1
.15, 1.21, 1.41, and 1.69 µm [18].
2
6
2
The system of emission registration included an FD-10
(
spectrum 2) mixtures. The vibrational–rotational
photodiode and a preamplifier equipped with field-con-
trolled transistors and U3-29, V6-9, and U2-8 amplifi-
ers connected in a series (the signal from the third one
was transmitted to the S9-8 oscilloscope). The registra-
tion was performed by the alternating current with
emission modulation by a mechanical interrupter oper-
ating at a frequency of 3300 Hz sufficient for ignition
registration (the characteristic time of combustion was
1
bands of HCl in the X Σ state (v= 3–0, where vis the
vibrational quantum number; band center, 1.198 µm; R
and P branches [19, 20]) are observed in the near IR
region (Fig. 2). A band system corresponding to both
.
2
2
the electron transition of HO (A A' – X A'', transitions
2
v = 0–0 at 1.43 mm, v= 1–1 at 1.48 µm [21]) and the
.
2
vibrational transition of OH (X Π , v= 2–0 transition;
i
>3 ms, see below). The interrupter was placed in front
band center, 1.437 µm; R, Q, and P branches [22, 23])
is registered at 1.34–1.51 µm.
of the inlet slit of the monochromator. The alternating
signal was thus registered and its maximal value was
determined as the arithmetic mean for 5–7 ignitions.
Using these data and the maximal signal intensity on
the kinetic curve for each wavelength λ, the emission
intensity vs. λ curve was plotted. The reproducibility of
the maximal emission intensity in each separate igni-
tion was ~15%. Figure 1 presents the optical scheme of
The intense composite emission bands of the H O*
2
(
λ = 0.823 µm) and H O * (λ = 0.854 µm) molecule
2 2
vibrations are observed in the far visible region at
0
.7−0.9 µm using an FD-10 detector without silicon
light filter [24, 25]. These bands can be attributed to the
following recombination reactions between H and
.
the setup. Under the operating conditions, the spectral HO :
2
–
1
slit width was ~35 cm at λ = 1.2 µm. This value is
comparable with the distance between separate vibra-
tional–rotational lines of HCl (3–0) and HF (2–0). For
HF (2–0) near the P and R branches, this distance is ~50
.
H + O + H O
HO + H O* + 47 kcal/mol,
2 2
2
2
.
.
2
HO + HO
H O* + O + 33.5 kcal/mol.
2 2 2
–
1
2
and 30 cm , respectively. For HCl (3–0), this distance
is still shorter because of the lower rotational constant During the combustion of mixtures with the SF addi-
6
B [19, 20]. The Doppler line width under these condi-
e
tives (Fig. 2a, spectrum 2), an additional HF X'Σ band
–
1
tions (~2000 K) is ~0.045 cm . Therefore, the experi- (v = 2–0 transition; band center, 1.29 µm; R and
mental spectra are represented as solid lines (like the P branches) is observed [19, 20]. The emission inten-
spectra of unresolved rotational structure).
.
.
2
2
2
sity of HO (A A' – X A'') and OH X Π (v= 2–0) in the
2
i
Before each run, the reactor was evacuated for 5 min band system at 1.34–1.51 µm dramatically decreases.
–
4
at a residual pressure of 4 × 10 torr. To remove the This suggests that the SF additives efficiently quench
6
.
deposits of solid products from the surface of the opti-
cal windows, the latter were polished after every 15
ignitions. This was a routine procedure because the
deposition of the SiO aerosol on the windows had to the translational temperature T , one can estimate the
virtually no effect on the signal registered in the near IR average temperature of the combustion products during
region. emission registration from the distribution of the emis-
electronically excited HO radicals.
2
Assuming that the rotational temperature T is equal
r
2
t
KINETICS AND CATALYSIS Vol. 42 No. 5 2001

5
96
CHERNYSH et al.
1
1
HCl X Σ (3–0)
HF X Σ (2–0)
zone of the propagating spherical flame (the regions
•
2
of the intensive chemical transformation, see Fig. 1).
In this case, the temperature of the combustion prod-
ucts is close to the adiabatic one, and the reaction heat
is Q ~ 170 kcal/mol [14, 27]. To estimate the contribu-
tion of each component to the overall emission, we reg-
istered the kinetics of the integral emission J in the vis-
ible (J ) and near IR spectrum regions (J ) (Fig. 3).
2
2
HO (A A' – X A'')(0–0)
(
b) P
R
R
P
Q
•
2
OH X Π (2–0)
i
R
P
3
3
(
a)
H O *
vis
IR
2
2
HF
•
•
HO + OH
2
H O*
The kinetic curves of the emission intensity for the
mixtures of the same composition in the visible (Fig. 3,
curve 4) and near IR (Fig. 3, curve 2) regions differ
2
1
2
HCl
max
4
4
widely from each other: the Jvis value is attained ear-
max
1
.1
1.2
1.3
1.4
1.5 0.80 0.85
0.90
λ, µm
Visible region
lier (time τ) than JIR . The observed rate of flame prop-
max
Near IR region
agation U before the attainment of the Jvis value (U ~
R/τ, where R is the reactor radius) was estimated from
Fig. 3 to be U ~ 20 m/s, which is close to its experimen-
tal value measured by rapid schlieren photography
Fig. 2. (a) Emission spectra of the rarefied flame during
DCS oxidation: (1) 23.5% DCS + O ; (2) 22% DCS +
2
1
8% SF + O ; (3) spectral sensitivity of an FD-10 photo-
6 2
[16]. This indicates that the Jvis value (Fig. 3, curve 4)
diode [45]; (4) aerosol background; (T = 300 K; P = 3 torr;
reflects the active chemical transformations in the
slit, 1 mm; and (b) emission spectra in the near IR region of
.
.
branched-chain reaction [12, 24], whereas the J value
IR
the HCl and HF molecules and HO and OH radicals [20,
2
(
curves 1–3) characterizes further emission in the
2
1, 23, 26, 29]. The intensity distribution for the R and P
resulting stable and inert reaction products. The shape
of the kinetic curves of the spectral ignition in the IR
branches is qualitative.
region is virtually independent of emitter (HCl, HF,
H O, or HO ) to which a kinetic curve refers. There-
.
sion intensities of the vibrational–rotational bands of
HCl (v= 3–0) and HF (v= 2–0). Taking into account
the data of [19, 20] and the experimental error in the
spectral intensity of the vibrational–rotational transi-
tions of HCl and HF, we found from Fig. 2 that the rota-
tional quantum numbers jmax ~ 10 ± 2 and (v = 2–0)
jmax ~ 7 ± 1 correspond to the maximum band emission
for HCl (v= 3−0) and HF (v= 2–0). From the relation
2
2
fore, Fig. 2 illustrates the ignition spectrum of the com-
bustion products in the bulk of the reactor. Comparison
of the kinetic curves (Fig. 3, curves 1–3) shows that the
SF additives cause an increase in the lifetime of the
6
vibrationally excited states (the existence of the rela-
max
tively constant JIR values). This can be attributed to
1
/2
j
~ (kT/2B ) [19], where k is the Boltzmann number
the transition of the excess internal energy of the SF6
max
e
–1
and B is the rotational constant (B
= 9.365 cm for molecules to the vibrational degrees of freedom of the
e
v= 3
–
1
HCl and Bv= 2 = 19.028 cm for HF [26]), we obtained combustion products. In other words, the addition of
that T ~ T = (1700–3900) K. This suggests that com- SF to the fuel mixture results in the formation of an
r
t
6
bustion occurs under nearly adiabatic conditions.
additional reservoir for the vibrational energy. That is
Taking into account that HCl is the main stable final why the energy amount from the vibrational levels of
product [27] and using the available probabilities of the the SF molecules is close for some time to the energy
6
vibrational transitions during ignition (the Einstein of heat loss during emission and heat exchange with the
–
1
coefficients) equal to 0.0379 s (HCl, v = 3–0) and reactor walls.
–
1
2
9.31 s (HF, v= 2–0) [28–30] and the estimates for
The Mache effect [26, 33] also influences the region
–
1
H O* and H O* equal to ~10 s [31], we assessed the
2
2
2
max
of stable JIR values due to heat release in the bulk of
number of these species (with respect to HCl) equal to
0.2% for HF and ~4.5% for H O and H O*. The esti-
the reactor during the adiabatic compression of the
combustion products upon the complete combustion of
~
2
2
2
mated concentration of water vapors is close to its cal-
culated equilibrium value [27, 32] in the hot products
of DCS + O + 3.76 N mixture combustion equal to
the fuel mixture. An increase in the SF concentration
6
in the mixture causes a noticeable retardation of flame
propagation [16] and, hence, a decrease in the contribu-
tion of the Mache effect because of an increase in the
heat loss through the wall as combustion continues.
2
2
~
1.72 mol %.
Using rapid schlieren photography, we observed
earlier [16] spherical flame propagation if emission was
initiated in the middle of the reaction vessel. Thus, the
max
However, the region of the relative stability of the JIR
spectrum registered (see Fig. 2) may be related to the values enlarges with an increase in the SF concentra-
6
emission of the final combustion products from the tion in the fuel mixture (Fig. 3, curves 1–3). This sug-
bulk of the sphere and the emission from the narrow gests that the contribution of the Maucher effect to the
KINETICS AND CATALYSIS Vol. 42 No. 5 2001

KINETIC STUDY OF THE CHEMILUMINESCENCE SPECTRUM
597
ln(J/Jmax) [arb. units]
0
heat capacity of the combustion products is negligible
and can be neglected.
The fact that the kinetic curves of the spectral emis-
.
2
sion for HCl, HF, H O, and HO in the IR region
–1
–2
2
2
exhibit the same characteristic times of the intensity
3
decrease and increase upon SF addition to the reaction
1
6
mixture suggests that the energy is transferred to the
vibrational modes of these species from the same
4
–
3
energy reservoir supported by the SF additives. The
6
heat loss for emission can also be neglected because of
the low probabilities of emission in the IR region. For
example, these probabilities are 33.9, 2.32, and
0
5
10
15
20 t, ms
–1
1
0
.0379 s for HCl in the ground state X Σ and the 1–0,
Fig. 3. Kinetics of emission during initiation ignition:
2
−0, and 3–0 transitions, respectively [28, 30] and 26.4
(
(
1) 23.5% DCS + O ; (2, 4) 26.5% DCS + 7% SF + O ; and
2 6 2
–
1
and 2.15 s for the composite transitions of the vibra-
tionally excited water molecules (101–000) and
3) 22% DCS + 18% SF + O . P = 3 torr, 293 K. (1, 3) Inte-
gral emission in the near IR region, and (4) integral emission
6
2
(111−000), respectively [31].
in the visible region.
The heat loss may also be due to the emission of
condensed species (for example, silane dioxide). How-
thermal regime of the cooling a spherical gas volume
ever, upon SF addition, the continuous emission disap-
6
from T to T (T is the wall temperature) in the absence
0
0
pears (see Fig. 2 and [12]), whereas the time interval,
during which IR emission is registered, is prolonged.
Additional energy can also be attributed to the presence
of overheated condensed species of silicon dioxide and
the energy liberated in the recombination of the active
of heat release in the volume, Frank-Kamenetskii
derived the following relation [36]:
2
1
(
T – T ) ∼ exp(–k α t),
(1)
0
sites of the chain reaction on the surface of the forming where k is the minimal eigenvalue of the thermal con-
1
new phase. Although SF consumes some part of ductivity linear equation, k = π/R , R is the reactor
6
1
0
0
energy released in phase formation [32], its additives radius, and α is the coefficient of temperature conduc-
cause a dramatic decrease in the overall amount of tivity. Assuming that the α value is equal to the diffusion
aerosol produced by ignition [34] via the reaction
SiO + 4HF SiF + 2H O.
coefficient D [36] and taking into account Eq. (1), one can
determine the characteristic time of cooling τ as
2
4
2
–
1
2
1
1.5
τ
= k D (T/T ) (P/P ),
(2)
Therefore, the vibrational energy of the sulfur
0
0
0
hexafluoride molecules is higher than the equilibrium
where D is the diffusion coefficient [37] under ordi-
0
max
value until the JIR values remain virtually unchanged. nary conditions (300 K, 760 torr) for the combustion
products, and P and P are the pressures corresponding
Then, a regular cooling regime is observed (see below)
characterized by the energy equilibrium between the to the temperatures T and T
vibrational and translational degrees of freedom of the that T = 1500 K for the beginning of the regular thermal
0
, respectively. Assuming
0
2
reaction products.
Let us estimate the characteristic times of deactivation 6 cm, we obtain the characteristic time of cooling
regime of cooling, D = 1.6 cm /s, P = 3 torr, and R =
0 0 0
−1
–1
due to the vibrational relaxation in the gas phase. For the τ ~ 210 s . This value is close to the experimental
–
1
HCl and O molecules, which are the main gaseous reac- one equal to ~200 s (Fig. 3, curves 1–3). Our finding
2
tion products, the rate constants for the vibrational deacti- suggests that the equilibrium between the vibrational
vation (the V–V exchange) from the v = 3, 2, 1 to the and translational degrees of freedom, that is, T = T , is
v
t
–
14
–12
3
v= 2, 1, 0 levels, respectively, are 10 –10 cm /s at
established in the region of the regular regime of cool-
ing. Therefore, the time of the establishment of the V−V
and V–T equilibria is shorter than the time of cooling
the reaction products. This also indicates that a delay in
cooling (the time interval corresponding to the region
3
00–2000 K [35]. Hence, the characteristic times of the
vibrational–vibrational relaxation for 3 torr τ are
V–V
–3
–5
nearly 10 –10 s. The relaxation of the vibrational
energy to the translational energy (the V–T exchange)
3
5
–5
requires 10 –10 collisions per second and τ = 10 –
max
V–T
–3
of the relatively constant JIR values (Fig. 3)) is due to
1
0 s at 2000 K. This suggests that the τ and τV–T val-
V–V
the relaxation of the energy of the vibrational reservoir
ues are comparable. In this case, if the rate of the
of added SF . After combustion, the concentration of
V−I exchange is higher than the rate of cooling the com-
6
bustion products during heat exchange with the reactor the vibrationally excited molecules increases, and its
walls, the characteristic time of the dip of the kinetic relative increase is proportional to the SF concentra-
6
curve J from the data of Fig. 3 should coincide with tion. The delay in cooling also increases with an
IR
time of cooling the reaction products. For the regular increase in the concentration of sulfur hexafluoride
KINETICS AND CATALYSIS Vol. 42 No. 5 2001

5
98
CHERNYSH et al.
(
Fig. 3). Therefore, the SF molecules, passed the lane oxidation and that the rate of flame propagation
6
chemical reaction front during spherical flame propa- depends on their chemical nature. The latter is due to a
gation, accumulate an overeqilibrium content of the change in the rate of termolecular termination of the
vibrational energy, which is further transformed into reaction chains via the reaction
the thermal energy of combustion products without
.
noticeable loss. This is also confirmed by the fact that
the final degree of expansion of the combustion prod-
ucts is close to the adiabatic one [16, 38] at various
H + O
2
+ M
HO
2
+ M
(III)
under the conditions of developed combustion. Note
also that the SF molecules can participate in chain ter-
mination involving charged species during DCS com-
bustion [17].
6
SF concentrations with due regard to the overall heat
capacity of the combustion products C . At the same
time, a dramatic decrease in the apparent rate of flame
propagation U in the presence of SF [16, 38] (Fig. 3,
6
p
Reaction (III) also plays a critical role in hydrocar-
bon combustion in mixtures that are poor or insuffi-
ciently enriched in fuel [43]. Note that the thermal dis-
sociation of fuel during the combustion of hydrocar-
6
curves 1 and 4) can barely be explained by the thermal
–
1/2
nature of flame propagation because U ~ Cp and vari-
ations in the diffusion coefficient of the active sites in bons (and inorganic hydrides) is a relatively rapid
the mixtures with SF additives [33] (see Appendix).
process resulting in the formation of hydrogen atoms
and other chain carriers [43]. The thermal decomposi-
6
Let us consider some possible processes responsible
tion of DCS also results in H formation under certain
for a decrease in the U value in the presence of sulfur
hexafluoride. It is known that SF additives cause a
decrease in the second self-ignition limit in the H + O
2
conditions [15, 27, 44], that is, the main elementary
reactions of hydrogen oxidation should play a signifi-
cant role in developed combustion at temperatures
6
2
2
.
reaction [39], thus stabilizing the inert HO radicals approaching 2000 K.
2
formed in the process
In this work, we found experimentally that the addi-
tion of SF to the DCS + O fuel mixture results in both
.
6
2
V – V
6
H + O + SF
HO + SF
,
(I)
2
6
2
the formation of the HF molecules and the deactivation
.
2
2
V – V
where SF6
of electronically excited HO radicals (A A' – X A'').
2
is a vibrationally excited species.
These findings indicate that the inhibition of flame
propagation in this reaction system by sulfur hexafluo-
ride additives should be due to reactions (I) and (II)
resulting in the replacement of the active centers of
reaction chains by inactive centers. In this case, reac-
tion (I) is accompanied by energy transfer to the vibra-
tional degrees of freedom of the sulfur hexafluoride
molecules.
The SF molecules added to the H + O + Ar flames
6
2
2
react with hydrogen atoms at 1300–1940 K with the
15
3
–1 –1
rate constant k = 2 × 10 exp(–(30 ± 5)/RT) cm mol s
the activation energy is expressed in kcal/mol) [40]. In
(
this case, the reaction
.
H + SF6
HF + SF
(II)
5
with the above rate constant is a chain termination reac-
tion. The SF molecules also dramatically inhibit com-
6
ACKNOWLEDGMENTS
bustion in the flames that are rich in hydrogen and air
[
41]. SF virtually does not dissociate under these con-
We thank V.V. Azatyan for fruitful discussions. The
work was supported by the Russian Foundation for
Basic Research (project nos. 99-03-32250 and 00-03-
6
ditions. Dissociation only occurs at T > 1500 K because
the S–F bond energy in this molecule is ~600 kcal/mol
3
2979).
[
42]. Abid et al. [24] also found that sulfur hexafluoride
additives inhibited the propagation of the spherical
H + O flame. Chemiluminescence in the visible
2
2
APPENDIX
region with the characteristic time, which is of the same
order as the time of ignition, is due to the processes
Let us show that variations in the coefficient of
.
mutual diffusion in the DCS + O mixtures in the pres-
2
involving the O, H, and OH radicals, whereas chemilu-
ence of the SF additives (up to 10%) cannot cause an
6
minescence in the near IR region with much longer char-
increase in the lower self-ignition limit [16].
acteristic times may be caused by the H O, H O and
2
2
2
.
To determine the coefficients of mutual diffusion in
a mixture, one should know the self-diffusion coeffi-
cients of the mixture components. DO2 = 0.18 cm /s for
HO species, which emit at the finite combustion tem-
2
perature. Water formation is accompanied by emission
2
with λ = 0.823 µm, whereas hydrogen peroxide forma-
.
.
O under ordinary conditions [35]. The self-diffusion
2
tion during the recombination HO + HO
H O* +
2
2
2
2
coefficients for SF , DCS, and O can be estimated
6
2
O is accompanied by emission with λ = 0.852 µm [24,
2
from the equation [45]:
2
5] (cf. Fig. 2). These data suggest that the SF addi-
6
tives have a similar effect on hydrogen and dichlorosi-
D = V/(3 2nσ),
(3)
KINETICS AND CATALYSIS Vol. 42 No. 5 2001

KINETIC STUDY OF THE CHEMILUMINESCENCE SPECTRUM
599
where V is the average rate of molecules; n is the num- 12. Rubtsov, N.M., Tsvetkov, G.I., and Chernysh, V.I., Kinet.
3
19
–3
Katal., 1997, vol. 38, no. 4, p. 498.
ber of molecules in cm (2.69 × 10 cm under ordi-
2
nary conditions); and σ = πd /4, where d is the collision 13. Rubtsov, N.M., Tsvetkov, G.I., and Chernysh, V.I., Kinet.
diameter. The d value was determined from the equa-
Katal., 1998, vol. 39, no. 3, p. 330.
–
9
1/3
tion d = (9.26 × 10 )b [35], where b is the correction
1
4. Rubtsov, N.M., Ryzhkov, O.T., and Chernysh, V.I.,
to the volume in van der Waals equation equal to bSF6
Kinet. Katal., 1995, vol. 36, no. 5, p. 645.
3
3
15. Aivazyan, R.G., Azatyan, V.V., Kalachev, V.I., and
Sinel’nikova, T.A., Kinet. Katal., 1995, vol. 36, no. 2,
p. 186.
=
88.1 cm /mol, b
= 107.8 cm /mol, and b
=
O2
DCS
3
3
1.6 cm /mol. The D values obtained with due regard to
the experimental DO2 value under ordinary conditions
16. Karpov, V.P., Rubtsov, N.M., Ryzhkov, O.T., et al.,
Khim. Fiz., 1998, vol. 17, no. 4, p. 73.
2
2
were DDCS = 0.045 cm /s and D = 0.043 cm /s. Tak-
SF6
1
1
1
7. Aivazyan, R.G., Kinet. Katal., 1997, vol. 38, no. 1,
ing into account that DDCS ~ DSF6 , the mutual diffusion
p. 195.
coefficients can be estimated by the Blanck equation
8. Kustanovich, I.M., Spektral’nyi analiz (Spectral Analy-
[35]:
sis), Moscow: Vysshaya Shkola, 1967.
–
ÄB
1
9. Bauwell, C.N., Fundamentals of Molecular Spectros-
copy, London: McGraw Hill, 1983.
D
= f /D + f /D ,
(4)
A
A
B
B
where f and f are the molar fractions of A and B, and
A
B
20. Kuipers, G.A. and Smith, D.F., J. Chem. Phys., 1955,
D and D are the self-diffusion coefficients of A and B
vol. 25, no. 2, p. 275.
A
B
under ordinary conditions. For the 10% DCS + O and
2
2
2
1. Becker, K.H., Fink, H.H., Langen, P., and Schurath, U.,
1
0% DCS + 4% SF + O mixtures, the D coefficients
6 2 AB
J. Chem. Phys., 1974, vol. 60, no. 11, p. 4623.
2
are 0.14 and 0.13 cm /s, respectively. Hence, the rela-
tive change in the diffusion coefficient is ~8.5%, which
is insufficient for a 1.5-fold increase in the lower self-
2. Garvin, D., Broida, H.P., and Kostkowski, H.J., J. Chem.
Phys., 1960, vol. 38, no. 10, p. 1742.
ignition limit even in the diffusion-controlled termina- 23. Charters, P.E. and Polanyi, J.C., Can. J. Chem., 1960,
tion of the reaction chains.
vol. 38, no. 10, p. 1742.
2
2
2
4. Abid, M., Wu, J.S., Lin, J.B., et al., Combust. Flame,
1
999, vol. 116, no. 3, p. 348.
REFERENCES
5. Gardiner, M.C., J. Phys. Chem., 1982, vol. 86, no. 6,
1
. Cze, S., VLSI Technology, New York: Murray Hill,
981, vol. 1.
p. 1429.
1
6. Radtsig, A.A. and Smirnov, B.M., Spravochnik po Atom-
noi i Molekulyarnoi Fizike (Handbook of Atomic and
Molecular Physics), Moscow: Atomizdat, 1980.
2
. Till, W.C. and Luxon, J.T., Integrated Circuits. Materi-
als, Devices and Fabrication, Prentice Hall: Englewood
Cliffs, 1982.
2
2
2
3
3
3
3
7. Britton, L.F., Plant Operations Progress, 1990, vol. 9,
3
. Conner, C.P., Stewart, G.W., Lindsay, D.M., and Gole, S.H.,
J. Am. Chem. Soc., 1977, vol. 99, no. 8, p. 2540.
no. 1, p. 16.
8. Cashion, J.K. and Polanyi, J.C., J. Chem. Phys., 1961,
4
5
6
7
8
. Chesnokov, E.N., Khim. Fiz., 1992, vol. 11, no. 9,
vol. 35, no. 2, p. 600.
p. 1202.
9. Herbelin, J.M. and Emanuel, G., J. Chem. Phys., 1974,
. Chesnokov, E.N., Doctoral (Phys.–Math.) Dissertation,
Novosibirsk: Inst. of Chem. Phys., 1999.
vol. 60, no. 2, p. 689.
0. Cashion, J.K. and Polanyi, J.C., Proc. R. Soc. A (Lon-
. Velikanov, S.D., Zapol’skii, A.F., and Frolov, Yu.N.,
Kvant. Electron., 1997, vol. 11, no. 1, p. 11.
don), 1960, vol. 258, p. 529.
1. Burch, D.E., Gryvnak, D.A., Patty, R.R., and Bartly, C.E.,
. Khairutdinov, R.F., Usp. Khim., 1998, vol. 67, no. 2,
J. Opt. Soc. Am., 1969, vol. 59, no. 3, p. 267.
p. 125.
2. Aivazyan, R.G., Azatyan, V.V., Kalachev, V.I., et al., Fiz.
Gor. Vzryva, 1994, vol. 35, no. 1, p. 77.
. Aivazyan, R.G., Azatyan, V.V., Kalachev, V.I., and
Rubtsov, V.I., Fiz. Gor. Vzryva, 1994, vol. 30, no. 4,
p. 85.
3. Zel’dovich, Ya.B., Barenblatt, G.I., Librovich, V.B., and
Makhviladze, G.M., Matematicheskaya teoriya
goreniya i vzryva (Mathematical Theory of Combustion
and Explosion), Moscow: Nauka, 1980.
9
. Karmazin, M.V., Kochubei, V.F., and Senyuta, T.B.,
Mater. IX Vsesoyuz. simp. po goreniyu (Proc. IX All-
Union Symp. on Combustion), Suzdal, 1989, p. 20.
3
4. Karpov, V.P., Chernysh, V.I., Rubtsov, N.M., and
Temchin, S.M., Proc. XIX Int. Symp. on Combustion
Processes, Czestochova, 1995, p. 22.
1
0. Vartanyan, A.A., Cand. Sc. (Chem.) Dissertation, Cher-
nogolovka: Inst. of Structural Macrokinetics, 1991.
1
1. Azatyan, V.V., Vartanyan, A.A., Kapkanov, V.A., and 35. Fizicheskie velichiny: Spravochnik (Physical Values: A
Temchin, S.M., Kinet. Katal., 1991, vol. 32, no. 2,
Handbook), Grigor’ev, I.S. and Meilikhov, E.Z., Eds.,
Moscow: Energoatomizdat, 1991.
p. 511.
KINETICS AND CATALYSIS Vol. 42 No. 5 2001

6
3
00
CHERNYSH et al.
6. Frank-Kamenetskii, D.A., Diffuziya i teploperedacha v 41. Miller, D.R., Evers, R.L., and Skinner, G.B., Combust.
khimicheskoi kinetike (Diffusion and Heat Transfer in
Flame, 1963, vol. 7, no. 2, p. 137.
Chemical Kinetics), Moscow: Nauka, 1967.
4
4
4
4
2. Handbook of Chemical Lasers, Fross, R.W.F. and Bott, J.F.,
3
7. Eremin, E.I., Osnovy khimicheskoi kinetiki (Foundations
Eds., New York: Wiley, 1976.
of Chemical Kinetics), Moscow: Vysshaya Shkola,
1
976.
3. Combustion Chemistry, Gardiner, W.C., Ed., New York:
3
3
8. Karpov, V.P., Rubtsov, N.M., Tsvetkov, G.I., and
Springer, 1984.
Chernysh, V.I., Khim. Fiz., 2000, vol. 19, no. 8, p. 74.
4. Aivazyan, R.G., Dzhabiev, T.S., and Kir’yakov, N.V.,
9. Khimicheskaya kinetika i tsepnye reaktsii (Chemical
Kinetics and Chain Reactions), Kondrat’ev, V.N., Ed.,
Moscow: Nauka, 1966.
Kinet. Katal., 1995, vol. 36, no. 2, p. 179.
5. Levshina, E.S. and Novitskii, P.V., Elektricheskie izme-
reniya fizicheskikh velichin (Electrical Measurements of
Physical Values), Leningrad: Energoatomizdat, 1983.
4
0. Fenimore, C.P. and Jones, G.W., Combust. Flame, 1964,
vol. 13, no. 3, p. 231.
KINETICS AND CATALYSIS Vol. 42 No. 5 2001