Kinetics of the Cl(2PJ) + CH4 Reaction: Effects of Secondary Chemistry below 300 K

Article abstract of DOI:10.1021/jp9913259

Absolute rate data for the Cl(²P_J) + CH₄ → HCl + CH₃ reaction have been obtained from 218 to 298 K by using the discharge flow resonance fluorescence technique in helium at 1 Torr total pressure. The result at 298 K is (10.1 ± 0.6) × 10^-14 cm³ molecule^-1 s^-1. The temperature dependence in Arrhenius form is (6.5 ± 0.9) × 10^-12 exp[(-1235 ± 34)/T]. The errors given are one standard deviation; overall experimental error is estimated at ± 15%. Because of the relatively large disagreement among earlier measurements at low temperatures, the results were examined for possible effects of non-Boltzmann spin distribution and vibrational excitation of CH₄, secondary chemistry of CH₃ radicals, and impurities in the Cl atom and CH₄ sources. There was no significant change in the observed rate constant when an efficient spin quencher, CF₄, was added, and estimates indicate that vibrational partitioning in CH₄ should be at the ambient reactor temperature before the start of the reaction. The results were also independent of the source of Cl atoms (microwave discharge or thermal decomposition of Cl₂) and whether CH₄ was purified in situ. However, the observed rate constant did depend on initial Cl atom concentrations and to a lesser extent on CH₄ concentrations. Numerical simulations were used to assess the importance of secondary chemistry over a range of reactant concentrations.

Full text of DOI:10.1021/jp9913259

7460
J. Phys. Chem. A 1999, 103, 7460-7469
2
Kinetics of the Cl( PJ) + CH4 Reaction: Effects of Secondary Chemistry below 300 K
J. J. Wang and Leon F. Keyser*
Atmospheric Chemistry Element, Earth and Space Sciences DiVision, Jet Propulsion Laboratory,
California Institute of Technology, Pasadena, California 91109
ReceiVed: April 22, 1999; In Final Form: July 22, 1999
Absolute rate data for the Cl(2P
using the discharge flow resonance fluorescence technique in helium at 1 Torr total pressure. The result at
) + CH
J
4
3
f HCl + CH reaction have been obtained from 218 to 298 K by
-
14
3
-1 -1
2
(
98 K is (10.1 ( 0.6) × 10 cm molecule s . The temperature dependence in Arrhenius form is (6.5
-
12
0.9) × 10
exp[(-1235 ( 34)/T]. The errors given are one standard deviation; overall experimental
error is estimated at (15%. Because of the relatively large disagreement among earlier measurements at low
temperatures, the results were examined for possible effects of non-Boltzmann spin distribution and vibrational
excitation of CH
There was no significant change in the observed rate constant when an efficient spin quencher, CF
added, and estimates indicate that vibrational partitioning in CH should be at the ambient reactor temperature
before the start of the reaction. The results were also independent of the source of Cl atoms (microwave
discharge or thermal decomposition of Cl ) and whether CH was purified in situ. However, the observed
rate constant did depend on initial Cl atom concentrations and to a lesser extent on CH concentrations.
4
, secondary chemistry of CH
3
radicals, and impurities in the Cl atom and CH
4
sources.
4
, was
4
2
4
4
Numerical simulations were used to assess the importance of secondary chemistry over a range of reactant
concentrations.
Introduction
yield linear Arrhenius plots above and below room temperature
while the other techniques observed nonlinear plots. At 298 K
In the atmosphere from 15 to 50 km the reaction of atomic
chlorine with methane is the major pathway by which reactive
these various methods agree within 6-13%; this is within the
experimental uncertainties of (15 to ( 20%. However, at 220
K the averages of the discharge flow (including the very-low-
pressure reactor data) and the flash photolysis results are higher
than the competitive chlorination results by about 15% and 35%,
respectively. This disagreement is at the margin of the combined
experimental uncertainties and might not be significant. But the
agreement among the studies within each technique combined
with the rather large discrepancies at low temperatures suggests
that systematic errors may be present in the low-temperature
studies of Cl + CH4 by one or more of these techniques.
Several possibilities exist that could account for the disagree-
ment at low temperatures. Nonequilibration between the atomic
2
Cl( P ) + CH f HCl + CH
(1)
J
4
3
chlorine is converted to the relatively inert reservoir species,
1
HCl. During winter in the polar regions, abnormally high levels
of reactive chlorine can be formed by reactions initiated on the
surfaces of stratospheric clouds. In many cases, reaction 1 also
controls the recovery rate from this perturbed condition to a
normal partitioning of reactive and reservoir chlorine species.
The Cl + CH4 reaction also has a major influence on the isotopic
composition of CH4 and CO in the stratosphere.
Because of its importance, reaction 1 has been the subject of
many studies using various experimental techniques over a wide
temperature range.12,13 For the present study of Cl + CH4
applied to stratospheric chemistry, we will consider only
temperature-dependent studies below 300 K. In this region there
have been several determinations of the absolute rate constant:
three studies using the discharge flow resonance fluorescence
technique under laminar14,15 or turbulent flow conditions; one
2-8
9-11
2
2
21
chlorine spin states ( P1/2 and P3/2) has been suggested. Since
vibrational excitation of CH4 can greatly increase the rate of
24,25
reaction,
incomplete vibrational relaxation at low temper-
atures could lead to observed rate constants that are too high.
Observed decays of atomic chlorine could also be affected by
secondary reactions (eqs 2-3) of molecular and atomic chlorine
15,16
with the methyl radicals produced by reaction 1.
Reaction
16
1
7
study using discharge flow mass spectrometry; one study using
Cl + CH + M f CH Cl + M
(2)
3
3
18
the very-low-pressure reactor technique; and four studies using
flash photolysis resonance fluorescence.1
9-22
In this temperature
Cl + CH f Cl + CH Cl
(3)
2
3
3
range, there are also two measurements that used competitive
chlorination1
7,23
to obtain the Cl + CH4 rate constant relative
of hydrogen atom impurities (eq 4) in the chlorine atom source
-11
to Cl + C2H6; to convert to absolute values we use 7.7 × 10
15
could interfere by regenerating chlorine atoms. Competition
3
-1 -1
12
exp(-90/T) cm molecule
s
for the C2H6 rate constant.
Within each technique the agreement is good over the entire
temperature range; but the competitive chlorination experiments
H + Cl f Cl + HCl
(4)
2
for chlorine atoms by residual ethane (eq 5) in the methane
source could interfere especially at low temperatures.
*
To whom correspondence should be addressed. Fax: (818) 393-5019.
E-mail: Leon.F.Keyser@jpl.nasa.gov.
1
0.1021/jp9913259 CCC: $18.00 © 1999 American Chemical Society
Published on Web 08/31/1999

2
Kinetics of the Cl( PJ) + CH4 Reaction
J. Phys. Chem. A, Vol. 103, No. 37, 1999 7461
Figure 1. Schematic diagram of discharge flow resonance fluorescence apparatus: BsBaratron pressure gauge (1∼10 Torr); MIsmovable injector
for addition of CH kinetics runs, F
4
+ He during Cl + CH
4
2
+ He during Cl atom calibration runs, and Cl
2
+ He during Cl atom wall loss runs;
Fsfurnace; Cscapillary tube; Psphotomultiplier with BaF
A4sinterface; A5scomputer.
2
filter; A1sfast-preamplifier; A2samplifier/discriminator; A3sdual counter/timer;
Cl + C H f HCl + C H
5
(5)
of Cl atoms, a fixed inlet for the He carrier, and a movable
inlet for CH4 + He . The inner surface of the reactor and outer
surface of the movable inlet are coated with halocarbon wax
2
6
2
The purpose of the present study is to test the discharge flow
resonance fluorescence technique for possible effects of second-
ary chemistry at temperatures below 300 K where the disagree-
ment among the previous studies is worst. To check for spin
equilibration, an efficient spin quencher, CF4, is added to the
Cl + CH4 reaction system. Vibrational deactivation is discussed
in terms of known quenching rate constants. To test for possible
secondary chemistry involving methyl radicals, initial concen-
trations of atomic chlorine and methane are varied over wide
ranges. Since in the present study atomic chlorine is the limiting
reactant ([Cl]o , [CH4]), methyl radical concentrations and,
thus, the importance of reactions 2 and 3 can be controlled by
varying the initial chlorine atom concentrations. To check for
impurities in the chlorine atom source and possible interference
from reaction 4, chlorine atoms are generated by using a
microwave discharge or thermal dissociation. Ethane impurities
in methane are minimized by using research grade methane and
by further purifying it in situ to double check for impurities
possibly added in the connecting vacuum lines. Finally a
numerical model is used to assess the importance of secondary
chemistry at varying concentrations of chlorine atoms and
methane.
(Series 15-00, Halocarbon Corp.) to minimize wall loss of Cl
atoms. The main helium carrier flow bypasses the discharge to
minimize production of impurity atoms such as H or O. Total
3
-1
helium flow rates are around 1900 to 2100 cm min at STP
-1
to establish flow velocities of 1000 to 1500 cm s . The flow
-1
system is pumped by a trapped 38 L s rotary pump; a throttling
valve is used to maintain a total pressure around 1 Torr at the
above flow velocities. Methane is added through a 1.5 cm o.d.
3
-1
movable injector at flow rates from 30 to 290 cm min at
STP. Other ports upstream of the reactor allow various reagents,
such as CF4, C2H6, Cl2, H2, and F2, to be added as needed to
the reactor tube. Temperatures in the reaction zone were
maintained within (2 K by using refrigerated bath circulators
(Neslab, ULT-80DD or RTE-110) to pass heat exchange fluids
(water or methanol) through the cooling jacket. Temperatures
were monitored by two thermocouples (Type E, chromel-
constantan) located inside each end of the cooling jacket.
Atomic Chlorine Sources. Chlorine atoms are generated in
dilute mixtures of Cl2 in He by using a microwave (2.45 GHz)
discharge or thermal decomposition. An uncoated 1 cm i.d.
Suprasil quartz tube is used in the microwave source, which is
3
-1
operated at 60 W. Total flow rates are about 300 cm min at
STP and the pressure is about 1 Torr. Typically 43 to 55%
dissociation efficiencies are obtained at Cl atom concentrations
Experimental Section
The present study was carried out by using a fast flow system
with resonance fluorescence detection. A schematic diagram of
the apparatus is presented in Figure 1. The experimental
approach is similar to that used in a previous study, and only
important modifications will be discussed in detail.
Reactor. The temperature-controlled Pyrex reactor has an
internal diameter of 5.04 cm and is 60 cm in length. At the
downstream end, it is connected by means of an epoxy seal to
a stainless steel resonance fluorescence cell. Pressure is
measured by using a 10 Torr capacitance manometer connected
to a port between the reactor and the fluorescence cell. At the
upstream end are connections to microwave and thermal sources
9
11
-3
6
.8 × 10 to 4.1 × 10 atoms cm .
For the thermal dissociation source, we use a 1 cm i.d. quartz
15
tube with a capillary region, 0.62 mm i.d. by 4 cm in length
located near the downstream end of a tubular shaped furnace,
which is operated at 1325 K and monitored by a thermocouple
(Type K, chromel-alumel). The heated section of the tube is
uncoated, but immediately downstream the walls are coated with
3
-1
phosphoric acid. Total flow rates are about 450 cm min at
STP and the total pressure is approximately 400 Torr. Dissocia-
tion efficiencies are 25 to 34% for Cl atom concentrations
1
0
11
-3
9.6 × 10 to 2.8 × 10 atoms cm .

7462 J. Phys. Chem. A, Vol. 103, No. 37, 1999
Wang and Keyser
Atomic Chlorine Detection. Cl atoms are detected by
by reference to an oil manometer. All thermocouples used in
the experiments were calibrated at 273 and near 195 K by using
ice plus water and CO2 plus methanol baths, respectively.
Barometric corrections were used to obtain the CO2 equilibrium
temperature. The temperature in the reaction zone was measured
by using a thermocouple probe in place of the movable inlet.
At low temperatures the probe reading was 1 to 2 K lower than
the two thermocouples in the cooling jacket, while at 298 K
the probe temperature was within 0.2 K. All temperatures
reported are based on the probe readings.
resonance fluorescence immediately downstream of the reactor.
The fluorescence is excited by radiation from a 50 W microwave
discharge in a chlorine resonance lamp. A mixture of about 0.5%
Cl2 in helium is passed through the lamp at pressures near 1.5
Torr. The resonantly emitted photons are detected at a 90° angle
to the lamp by means of a photomultiplier, PMT, (EMR 541G-
0
8-18) sensitive between 105 and 220 nm. A barium fluoride
filter window is placed before the PMT to cut off radiation
below 136 nm and eliminate interference from oxygen and
hydrogen atom impurities. In this spectral region a strong
chlorine transition occurs near 139 nm. As shown in Figure 1,
the output of the PMT is fed to a photon counting system and
then to a computer by a standard RS-232 interface. The scattered
signal is significantly reduced by using light baffles in front of
the chlorine atom lamp and the PMT and Wood’s horns at
positions opposite to the lamp and the PMT. During the Cl +
CH4 runs, background fluorescence signals were determined
with the Cl2 flow turned off, discharge or furnace on, and C2H6
added to scavenge any residual Cl atoms.
Corrections. The data were corrected for the viscous pressure
drop between the reaction zone and the pressure measurement
2
9
port. Pressure corrections were less than 0.5%. Observed
pseudo-first-order rate constants were also corrected for axial
and radial diffusion by using the method described previ-
3
0,31
1.75
ously.
For these corrections a value of 0.0237 × T Torr
2
-1
cm s was used for the diffusion coefficient of Cl atoms in
3
2
He. Diffusion corrections ranged from 2 to 14% at all
temperatures studied.
Reagents. Gases used were chromatographic grade He
(99.9999%), research grade Cl2 (99.99%), Matheson research
Atomic Chlorine Calibration. The Cl atom detection
sensitivity is calibrated by generating a known amount of Cl
atoms either from the reaction F + Cl2 f Cl + FCl (k )
grade CH4 (99.99%), CF4 (99.9%), H2 (99.9995%), a 10.33%
mixture of C2H6 in He and a 1% mixture of F2 in He. He was
further purified by passage through a molecular sieve (Linde
3A) trap at 77 K just prior to use. CH4 was purified in situ as
described below.
-10
3
-1 -1
26
1
.6 × 10 cm molecule
H + Cl2 f Cl + HCl (k ) 2.6 × 10
); in both cases, the F atoms or H atoms are produced
s
at room temperature ) or from
-
11
3
-1
cm molecule
-
1 27,28
s
in a microwave discharge and used in large excess over known
concentrations of Cl2. The observed signals contain contributions
from Cl atom fluorescence produced by the calibrating reactions
and fluorescence from other background sources of Cl atoms
and from scattered light. The sum of the background and
scattered signals is determined by turning off the Cl2 flow.
Semilog plots of the calibration signal vs the reaction length
are linear, and loss of Cl atoms is accounted for by extrapolating
these plots to the detector position with a linear least-squares
analysis. The two methods agree within 5 to 25%. Since lower
background losses are observed in the F atom reaction, it is
used in most of the calibrations. Typical detection sensitivities
Results
The present experiments were carried out with methane in
large excess at temperatures between 218 and 298 K. Helium
was used as the carrier gas at total pressures of (1.02 ( 0.02)
Torr. Molecular chlorine, Cl2, was added at concentrations from
1
0
11
-3
1.1 × 10 to 4.1 × 10 molecules cm ; after dissociation by
microwave discharge or thermal decomposition, this resulted
9
in initial chlorine atom concentrations, [Cl]0, from 6.8 × 10 to
1
1
-3
4.1 × 10 atoms cm . Methane concentrations were between
14
15
-3
5.7 × 10 and 4.5 × 10 molecules cm . Initial stoichiometric
3
5
ratios, [CH4]/[Cl]0, ranged from 4.8 × 10 to 1.6 × 10 . Under
these conditions, the loss of atomic chlorine is pseudo-first-
order and may be written
-
8
-1
-3
are about 8.0 × 10 counts s /atoms cm with background
-
1
signals at about 380 counts s . For a typical 50 s counting
time, this is equivalent to a minimum detectable [Cl] of 5 ×
7
-3
1
0 atoms cm at a signal-to-noise ratio of unity. Plots of
-d[Cl]/dt ) k [CH ] [Cl] + k [Cl]
(6)
1
4
L
fluorescence signal vs [Cl] are linear over the entire range
9
11
-3
studied from about 2 × 10 to 3 × 10 atoms cm .
Since the Cl atom resonance fluorescence intensity, I(Cl), was
found to vary linearly with [Cl], we can write
Chlorine Atom Wall Loss. Since Cl atoms are added and
detected at fixed points, wall losses are not directly observed
in our experiments, but they are needed for the diffusion
corrections and numerical simulations described below. These
measurements are performed under the same conditions of
temperature, pressure, and flow velocity as those used in the
Cl + CH4 experiments, except, of course, without added CH4.
Cl atoms are formed in a microwave discharge of Cl2 + He in
the movable inlet and added at various reaction lengths. The
resulting linear semilog plots of the fluorescence signals vs
length show that the wall loss is first order in [Cl]. A linear
least-squares analysis is then used to obtain the wall loss rate
constants from the slopes of these plots. After diffusion
corrections the values observed (( one standard deviation) are
k ≡ -d ln[I(Cl)]/dt ) k [CH ] + k
(7)
+
1
4
L
where k+ is the total first-order rate constant for loss of atomic
chlorine with CH4 added through the movable injector, k1 is
the bimolecular rate constant for reaction 1, and kL accounts
for changes in the Cl atom signal other than by reaction with
CH4. The latter effect is due to changes in the flow velocity, Cl
atom wall loss, and reaction with impurities in the He carrier
gas up and down stream of the movable injector. Values of k+
were determined from the slopes of ln[I(Cl)] vs l plots by linear
least-squares analysis when CH4 was present; here l is the
reaction length, which under plug flow conditions determines
the reaction time, t ) l/V, where V is the average flow velocity.
Reaction lengths were 4 to 26 cm and reaction times ranged
from 2.7 to 26 ms. The range of Cl atom decays depended on
[CH4] and the temperature; between these reaction times the
-
1
1
2
1.7 (( 2.9), 14.0 (( 4.1), and 4.2 (( 2.8) s at 218, 261, and
98 K, respectively.
Calibrations. All mass flow controllers and meters were
calibrated for N2, He, or CH4 by using the volume change at
constant pressure (bubble meter) method or by the pressure rise
at constant volume method. Pressure gauges were calibrated
-
0.4
-4.0
decays were from about e
to e . Most semilog plots of
the Cl decay were linear over the entire range; if nonlinearity

2
Kinetics of the Cl( PJ) + CH4 Reaction
J. Phys. Chem. A, Vol. 103, No. 37, 1999 7463
TABLE 1: Summary of Observed Rate Constants for
fitting. The Arrhenius plot shown in Figure 3 results in the
following expression:
Cl + CH
4
1
k )
(10- cm molecule s-1
14
3
-1
-
12
k1 ) (6.5 ( 0.9) × 10
T (K)
runs
averagea,b
slopea,c
exp[(-1235 ( 34)/T] cm molecule s-1 (9)
3
-1
298
279
261
239
218
24
12
38
13
29
10.1 ( 0.60
7.56 ( 0.70
5.98 ( 0.43
3.69 ( 0.20
2.20 ( 0.28
10.2 ( 0.20
6.28 ( 0.48
6.46 ( 0.18
3.49 ( 0.10
2.25 ( 0.08
for 218 e T e 298 K; the errors given are one standard
deviation obtained from an unweighted least-squares analysis.
Overall experimental uncertainty is estimated to be (15%.
a
Errors are one standard deviation. b Average of individual k
1
4
′/[CH ].
c
Discussion
From plots of k ′ ) k [CH ] vs [CH ].
1
1
4
4
Spin Equilibration. Chlorine atoms can be in either the
2
2
ground P3/2 or the excited P1/2 spin state, which is higher in
-
1
33,34
energy by 882 cm (2.52 kcal/mole).
In the stratosphere
the spin states are expected to be in thermal equilibrium;3
however, in the laboratory they may not be, and in order to
apply measured rate coefficients to the atmosphere, it is
3
2
important to check for spin equilibration. If Cl( P1/2) reacts with
CH4 at a rate sufficiently greater than Cl( P3/2), the spin
2
distribution could affect the measurements. If interconversion
2
Cl( P ) + CH f HCl + CH
(1a)
(1b)
3
/2
4
3
2
Cl( P ) + CH f HCl + CH
1
/2
4
3
between the Cl states is rapid compared to reaction with CH4,
then spin equilibration can be maintained and the observed rate
3
4
constant is at its maximum and is given by
k (spin equilibrium) ) (k + K × k )/(1 + K) (10)
1
1a
1b
where the equilibrium constant, K ) 0.5 exp(- 2520/RT).
In the case of very slow interconversion, the two states decay
independently with distinct rate constants, k1a and k1b. Even if
the states are near equilibrium at the start of the Cl + CH4
reaction (which might be the case in the present experiments
because of a 50 ms delay between Cl formation and CH4
addition), only a small fraction, 0.7% at 298 K, of Cl would be
2
in the P1/2 state and the observed signal will be due mostly to
2
the P3/2 state. Under slow spin equilibration conditions, the
Figure 2. Plot of pseudo-first-order rate constant vs methane concen-
tration; all of the data have been corrected for viscous pressure drop
upper state would not be repopulated fast enough and the
observed rate constant would be just k1a, not the equilibrium
rate constant (eq 10) needed for atmospheric models.
and diffusion (see text); open symbols are without added CF
symbols are with added CF ; circles indicate experiments in which CH
was purified in situ (see text for details); triangles represent experiments
in which CH was not purified in situ; lines are least-squares fits of
4
; filled
4
4
In a flash photolysis resonance fluorescence study of the
2
1
Cl + CH4 reaction, Ravishankara and Wine reported that at
low temperatures the observed Cl losses depended on the
composition of the reaction mixture. Under conditions where
spin equilibration could be maintainedsat low [CH4] (low Cl
loss rates) or when efficient spin quenchers were presentsthey
observed higher values for k1′. They suggested that the differ-
ences among the various experimental techniques at low
temperatures could be explained by the absence of quenchers
and, thus, a non-Boltzmann distribution of the spin states in
both the discharge flow and competitive chlorination studies.
Since measurements using the discharge flow technique are
carried out at low pressures where diffusion is rapid, reactive
species can undergo many wall collisions during their lifetime.
4
the data; temperatures are given next to each plot.
was observed, the initial slopes were used. Values of kL were
determined from the slopes of ln[I0(Cl)] vs l plots by linear
least-squares analysis when no CH4 was present in the system.
Generally kL was less than about 10% of k+. I0(Cl) is the Cl
atom fluorescence intensity without CH4; it was measured by
replacing the CH4 flow with an equivalent flow of helium
through the movable injector. A few minutes delay time was
used after the CH4 shut off in order to allow the flow system to
stabilize. Then from eq 7, the pseudo-first-order rate constant,
k1′, is given by
-
6
Collision efficiencies less than 5 × 10 have been reported
k ′ ) k [CH ] ) k - k
(8)
2
35
1
1
4
+
L
for deactivation of Cl( P1/2) on Teflon coated walls. No spin
deactivation rates have been reported for the wax coated walls
used here; however, if they are as low as the reported values
on Teflon, wall collisions would be too inefficient to equilibrate
the states in our system. On the other hand, recent measurements
of gas-phase spin quenching rates have shown that He and CH4
A summary of observed rate constants is listed in Table 1
and shown in Figure 2.
No significant differences were observed between values
obtained from the averages of individual k1′/[CH4] points and
from the slopes of k1′ vs [CH4] plots by linear least-squares
2
33
itself are efficient quenchers of the P1/2 state. To apply the

7464 J. Phys. Chem. A, Vol. 103, No. 37, 1999
Wang and Keyser
Figure 3. Arrhenius plot of present results compared to earlier studies using several experimental techniques. Absolute values for the competitive
-
11
3
-1 -1
12
chlorination studies were calculated using 7.7 × 10 exp(-90/T) cm molecule
s
for the rate constant of the reference reaction, Cl + C
2
H .
6
The line is the unweighted linear least-squares fit of the present results only.
discharge flow measurements of k1 to stratospheric chemistry
with confidence, it is crucial to confirm that Cl is indeed at
thermal equilibrium under the conditions of the present study.
Nonequilibrium could be revealed by nonlinear Cl decay curves
and by changes in the observed rate constant when spin
quenchers are added. At all temperatures studied, no significant
curvature was observed in the ln [Cl] vs time plots. In addition,
an efficient spin quencher, CF4, was added to the reaction
TABLE 2: Effect of Added Spin Quencher, CF
4
, on the
Rate Constant, k
1
1
k )
(10- cm molecule s-1
14
3
-1
without CF a
a
T (K)
runs
4
runs
with CF
4
2
2
2
98
79
61
18
7
17
9
10.2 ( 0.68
7.30 ( 0.80
6.04 ( 0.45
3.65 ( 0.17
2.16 ( 0.33
6
5
21
4
10.0 ( 0.43
7.92 ( 0.34
5.94 ( 0.42
3.77 ( 0.25
2.25 ( 0.21
239
18
2
15
14
2
2
Cl( P ) + CF T Cl( P ) + CF
(11)
a
1/2
4
3/2
4
1 4
Errors are one standard deviation; average of individual k ′/[CH ].
2
mixture to quench any excess Cl( P1/2) and maintain spin
equilibration. Measurements of the forward rate constant for
tion. Estimates and numerical simulations using the above
quenching rate also show that most of the spin equilibration
can be maintained by CF4 alone even if k1b is as much as 100
times greater than k1a. In our system CH4 and He also act to
-
11
3
-1 -1 33,36
reaction 11 average about 9 × 10
cm molecule s .
1
3
CF4 was added at concentrations averaging about 5.5 × 10
-
3
-1
33
molecules cm to yield a quenching rate of nearly 5000 s .
Since there is about a 50 ms delay between the addition of CF4
to Cl atoms and the start of the Cl + CH4 reaction, there is
more than sufficient time to bring the spin states into equilibra-
maintain the equilibrium. The results of adding CF4 are
summarized in Table 2. No significant change in the rate
constant was observed when quencher was added at tempera-
tures from 218 to 298 K. These observations and simulations

2
Kinetics of the Cl( PJ) + CH4 Reaction
J. Phys. Chem. A, Vol. 103, No. 37, 1999 7465
show that the spin states were in thermal equilibrium in the
present measurements.
Vibrational Deactivation. Recent theoretical and experi-
mental studies have shown that the rate of reaction 1 is very
sensitive to vibrational excitation of CH4. Ab initio calculations
using variational transition state theory predict that population
-1
of the umbrella bending mode (ν4) near 1306 cm can increase
the reaction rate over the ground state.37 Enhancement factors
as large as 16 and 52 are calculated at 300 and 200 K,
24
respectively. A molecular beam study at low collision energies
has shown that speed distributions and spatial anisotropies of
product CH3 are consistent with reaction of CH4 with vibrational
excitation in the ν4 mode and/or in the torsional mode (ν2) near
-
1
1
534 cm ; moreover, excitation to one of these modes enhances
25
the reaction rate by a factor of 200. At 298 K the thermal
population of the two modes is 0.66%; at 218 K it is 0.062%.
Because of the large enhancement, it is important that at the
start of the reaction, the CH4 vibrational population be the same
as the nominal temperature of the carrier gas. In our system,
CH4 is precooled in the movable injector before it is added to
Cl atoms. Residence times in the injector were 10 to 13 ms
16
16
with average [CH4] and [He] near 1.6 × 10 and 3.7 × 10
-3
molecules cm , respectively. Vibrational deactivation can occur
by wall collisions as well as by gas-phase collisions with He
and CH4 itself. Numerical simulations using quenching rate
-
14
-15
3
constants at 215 K of 1.1 × 10
and 6.2 × 10
cm
-
1
-1
38,39
molecule
s
for CH4 and He collision partners,
show that
the 10-13 ms time period is sufficient to deactivate the 298 K
population of the ν2 and ν4 modes and vibrational excitation
should not interfere with our results.
1
Figure 4. Plot of observed k vs initial chlorine atom concentration;
solid lines are linear least-squares fits to the data; dashed lines are
obtained from numerical simulations described in the text; symbols
have the same meaning as in Figure 2.
Methyl Radical Reactions. To check for secondary chemistry
involving methyl radicals, initial chlorine atom and methane
concentrations were varied over a wide range. Since [Cl]0 ,
TABLE 3: Effect of Two Methods Used to Remove
Interference from Reaction 2
[CH4], [Cl]0 determines the maximum [CH3] that can be
produced and, hence, the importance of reactions 2 and 3. We
can estimate the expected interference from reaction 2 at various
concentrations by requiring the Cl atom loss by this reaction to
be less than 10% of the loss by reaction 1. That is,
(10-14 cm molecule s-1)a
3
-1
k
1
exclude points
simulations predict
errors > 5%
extrapolate
observedb
c
b,d
T (K)
98
261
to zero [Cl]
0
2
10.1 ( 0.60
5.98 ( 0.43
2.20 ( 0.28
9.66 ( 0.26
5.70 ( 0.08
2.16 ( 0.08
9.80 ( 0.14
5.81 ( 0.32
2.20 ( 0.30
k [CH ] < 0.1 k ′
(12)
2
3
1
2
18
Since [CH3] ≈ [Cl]0, we have
Cl] e 0.1 k ′/k
2
a
Errors are one standard deviation. b Average of individual k
1 4
′/[CH ].
c
See Figure 4. d See Figure 6.
[
(13)
0
1
change in observed k1 as [Cl]0 is increased at three temperatures;
not enough data points were taken at 239 and 279 K to give
reliable plots. At 298 K average values of k1, represented by
the linear least-squares fit, increase by about 20% when [Cl]0
Values for k2 have been obtained at temperatures from 298 to
23 K by flash photolyzing mixtures of Cl2, CH4, and CO2 and
using UV absorption to monitor CH3; computer simulations of
CH3] vs time profiles and product analyses were then used to
4
1
0
11
-3
[
varies from 1 × 10 to 4 × 10 cm ; at 261 K the change is
also about 20%, while at 218 K observed k1 changes by only
6% over the same range of [Cl]0. Since reaction 2 represents
an added loss of Cl, the increase in observed k1 with increasing
[Cl]0 indicates that this reaction may be interfering at high [Cl]0.
By extrapolating the k1 vs [Cl]0 curves to zero [Cl]0, we can
estimate k1 free of interference from reaction 2. These values
are compared to average values in Table 3. The extrapolated
values are lower than the averages, but the effect is less than
5% at the three temperatures plotted. Computer simulations were
used to further investigate interference from this reaction, and
the results are discussed below.
40
-10
3
determine k2. At 298 K k2 was found to be 2 × 10
molecule s-1 and is independent of CO2 pressures between
0 and 300 Torr. For the purposes of the present discussion,
cm
-
1
5
we extrapolate these results to lower temperatures by using the
reported temperature dependence; k2 then is given by 3.7 ×
-
10
3
-1 -1
1
0
exp(-185/T) cm molecule s . Using this value for k2
at 298 K, eq 13 becomes
8
[
Cl] e 5 × 10 × k ′
(14)
0
1
To avoid interference, [Cl]0 e 1.2 × 10 cm-3 for k1′ ) 25
10
-
1
11
-3
-1
Similar estimates can be made for reaction 3; in this case we
have
s
and [Cl]0 e 1.5 × 10 cm for k1′ ) 300 s . Over the
concentration ranges used in the present experiments, eq 14
predicts that some interference from reaction 2 is to be expected.
This is confirmed by the plots in Figure 4, which show the
k [CH ][Cl ] e 0.1 k [Cl][CH ]
(15)
3
3
2
1
4

7
466 J. Phys. Chem. A, Vol. 103, No. 37, 1999
Wang and Keyser
TABLE 5: Reactions Used in Numerical Simulations
TABLE 4: Effect of Chlorine Atom Source and Purification
of Methane
reaction
rate constanta
refs
a
k
1
Cl + CH
4
f HCl + CH
3
measured in this study,
see text
k
1
a, purify CH
yes
4
in situ
no
microwave
discharge
source
-
10
thermal
source
CH
CH
CH
3
3
3
+ Cl f CH
+ Cl f CH
+ CH + He f
+ He
Cl + Cl + He f Cl
Cl + wall f products
3
Cl
3.7 × 10 exp(-185/T)
40
41
-
12
T (K) runs
runs
6
(17 runs)
(21 runs)
2
3
Cl + Cl
5.0 × 10 exp(-267/T)
-30
3
2.3 × 10 T exp(+185/T) 45, 46
2
2
2
79
61
39
6
7.94 ( 0.41
7.17 ( 0.75
C
2
H
6
at 1 Torr
5.79 ( 0.31 6.14 ( 0.45
-34
2
+ He
6.1 × 10 exp(+906/T)
measured in this study,
see text
27
7
3.71 ( 0.12
6
3.66 ( 0.27
a
0-14 cm molecule s-1; average of individual k
3
-1
1
1 4
′/[CH ]; errors
-
11
Cl + CH
Cl + C
+ Cl
Cl + C
+ C
+ C
CH + wall f products
3
Cl f HCl + CH
f HCl + C
f Cl + C
f HCl + C
f products
2 5
H f products
2
Cl 3.2 × 10 exp(-1250/T) 12
are one standard deviation.
-
11
2
H
6
2
H
5
7.7 ×10 exp(-90/T)
12
41
47
48
48
-
11
C H
2 5
2
2
H
5
Cl
4
1.3 × 10 exp(+152/T)
noting that [CH3] ≈ [Cl] and that [Cl2] ) (1/2fd) (1 - fd) [Cl]0,
-
-
-
10
11
11
2
H
5
2
H
2.4 × 10
2.0 × 10
5.0 × 10
we have
C H
2 5
2 5
H
CH
3
[
Cl] e (0.1/k ){2f /(1 - f )}k ′
(16)
0
3
d
d
1
varies, see text
3
a
Units are cm molecule s-1 and cm molecule s-1 for third-
6
-2
3
-1
where fd is the fractional dissociation of Cl2, which ranges
order and second-order reactions, respectively.
typically from about 0.25 to 0.55; for this estimate we take
4
1
-12
fd ) 0.33; using the literature value for k3: 5.0 × 10
concentrations. The results can then be used to assess under
which conditions the observed rate coefficients are free from
significant interference. Most of the simulations were carried
3
-1 -1
exp(-267/T) cm molecule s , at 298 K we have
10
[
Cl] e 5 × 10 k ′
(17)
0
1
42
out using the Chemical Kinetics Simulator (CKS). In contrast
-
1
to the customary numerical integration approach, CKS is based
on a stochastic algorithm. The output from CKS was compared
with results obtained by using two differential equation integra-
The minimum k1′ at 298 K was 27 s and this requires
1
2
-3
[
Cl]0 e 1.3 × 10 cm to prevent interference from reaction
11
-3
3
. The maximum [Cl]0 used was 3.1 × 10 cm , which is
43
tors based on standard Gear algorithms: ACUCHEM and
well below the estimate at 298 K. Similar results are found at
the lower temperatures and these indicate that reaction 3 should
not interfere with the present measurements.
4
4
CHEMRXN. The results from all three programs were
essentially the same.
The reactions and rate constants listed in Table 5 were used
to model the present study of Cl + CH4 at 218, 261, and 298
K. Initial [Cl] and [CH4] ranged over those used in the
experimental runs. Input rate constants for reaction 1, k1(in),
were the values obtained by extrapolating to zero [Cl]0 (see
Figure 4 and Table 3). For Cl wall loss rates, we used the
observed values given in the Experimental Section. The model
output consists of [Cl] vs reaction time profiles. These were
treated in the same way as experimental data to obtain the model
prediction of k1: plots of ln [Cl] vs reaction time were fit by
linear regression over time ranges similar to those used in the
experiments; the slope gives a value for -d ln [Cl]/dt and, hence,
from eq 7 a value for k1′(out), and, knowing [CH4], k1(out) can
be obtained.
Chlorine Atom Source. A microwave discharge was used
to produce atomic Cl in most of the experimental runs. To
minimize production of impurities, such as H or O atoms, most
of the He flow bypassed the discharge region. If produced in
sufficient concentrations, H atoms could interfere by forming
atomic Cl by reaction 4. To check this possibility, a cleaner,
thermal source of Cl atoms was used in some of the experiments.
The results are summarized in Table 4. At both temperatures
studied, no significant differences were observed using these
two sources of atomic Cl. This shows that reaction 4 does not
interfere with the present measurements.
Purification of Methane. Since C2H6 and higher alkanes
react much more rapidly with Cl atoms than CH4, it is important
to use very high purity CH4, especially at low temperatures.
For example, at 298 K, k5/k1 is about 600 and at 218 K the
ratio is about 2800. Thus, to keep interference from reaction 5
less than 1% requires that impurity levels of C2H6 in CH4 be
less than 17 and 4 ppm at 298 and 218 K, respectively. Batch
analyses of the CH4 supplied by the manufacturer found less
than 1 ppm C2H6 and less than 2 ppm C3H8; these levels should
prevent interference even at the lowest temperatures studied.
However, to remove possible impurities added from the vacuum
lines, CH4 was purified in situ just upstream of its addition point
during all (except for some runs at 261 K, see below) of the
experiments below 298 K. This was done by passing the CH4
through a molecular sieve (Linde 3A) trap at 195 K; earlier
work showed that this method reduced C2H6 to less than 1 ppm
with no trace of higher alkanes.1 At 261 K, in situ purification
was done during 17 runs and no purification was done during
As a first test of the model, we simulated the observed
increase in k1 with increasing [Cl]0 by varying k2; to make the
system as sensitive as possible to CH3 reactions we assumed
that CH3 wall loss is zero. As shown in Figure 5, the value of
-
11
3
k2 in the model must be greater than about 5 × 10
molecule to fit the observations at 298 K. The best fit
occurs in the range (1 to 2) × 10 . This is close to the value
cm
-
1
-1
s
-
10
-
10
3
-1 -1
40
of 2 × 10
at 298 K, and, as a worst-case test for interference from reaction
, we adopt this value along with their reported temperature
cm molecule
s
reported by Timonen et al.
2
dependence (see above and Table 5) in the rest of this discussion.
Except at very low [CH4], the value of k1(out) depends less
strongly on [CH4] than on [Cl]0. Using this observation we can
simplify the tests for interference from reaction 2 by setting
[CH4] at some intermediate value and fitting the model to the
k1 vs [Cl]0 plots shown in Figure 4. The upper dashed lines
near the plots for each temperature show the model results when
CH3 wall loss is set to zero; in this case the model results are
more sensitive to [Cl]0 than observed. The lower dashed lines
4,15
2
1 runs. The results are compared in the last two columns of
Table 4; the difference is about 6% and does not appear to be
significant even at the level of one standard deviation.
Computer Simulations. Numerical models were used to
evaluate further the sensitivity of the Cl + CH4 reaction system
to secondary chemistry over a range of temperatures and reactant
-
1
are obtained by using 75, 50 and 150 s for the CH3 wall loss
at 298, 261 and 218 K, respectively. The best fit occurs at CH3

2
Kinetics of the Cl( PJ) + CH4 Reaction
J. Phys. Chem. A, Vol. 103, No. 37, 1999 7467
Figure 5. Plot of k
1
vs initial chlorine atom and methane concentrations at 298 K; filled and open symbols represent experiments with and without
-
11
added CF , respectively; lower and upper mesh plots are obtained from numerical simulations described in the text with k
4
2
set at 5 × 10 and
-
10
3
-1 -1
2
× 10 cm molecule
s , respectively.
-
1
-12
3
-1 -1
wall loss rate constants of 50, 40, and 150 s ; these values
correspond to wall loss collision efficiencies (γ) of 0.005, 0.004,
and 0.1 at the three respective temperatures. What we call CH3
wall loss could conceivably also include CH3 reactions with
impurities in the reaction mixture.
8.3 × 10
cm molecule s , calculated from the thermo-
18
chemistry of a collinear C-H-Cl transition state. A linear
transition state is also predicted by ab initio calculations.4
Cold rotational state distributions observed in the product HCl
also indicate a linear transition state.
The present results are in excellent agreement with earlier
results from this laboratory using the same experimental
technique but different flow tubes, flow meters, pressure gauges
and temperature probes. In the earlier work the surface-to-
9-51
5
2
The models can now be used to estimate the degree of
interference from secondary chemistry over various ranges of
Cl]0 and [CH4] at each of the temperatures studied. In Figure
1
5
[
6
the best-fit model predictions for the percent changes in k1
-
1
induced by secondary chemistry are plotted vs [Cl]0 and [CH4].
The percent change in k1 is calculated using the relation,
del% ) {[k1(out)/k1(in)] - 1} × 100%, where k1(out) and k1-
volume ratio of the flow reactor was 1.6 cm compared to the
-
1
present value of 0.8 cm ; also, the previous study used a
phosphoric acid wall coating, not the halocarbon wax of the
present study. The good agreement shows that wall effects such
as a second-order surface reaction of Cl and CH4 and calibration
errors did not affect these measurements. The present measure-
ments are also in very good agreement with the combined results
of previous studies using discharge flow resonance fluores-
(
in) are the input and output of the model discussed above. The
results show that in order to avoid interference at the 10% level,
11
-3
[
Cl]0 should be less than about 1 × 10 atoms cm at 298
11 -3
and 261 K and less than about 2 × 10 atoms cm at 218 K.
Since the majority of the data points lie within the 5% contours
at each of the three temperatures, the measured k1 values should
not be seriously affected. This can be tested by removing all
data points beyond the 5% contours and recalculating the
averages for k1. The results, shown in Table 3, are lower than
the averages of all the data by less than about 3% and indicate
that the present measurements should not be influenced sig-
nificantly by secondary reactions.
1
4-16
17
cence,
discharge flow mass spectrometry, and very-low-
1
8
pressure reactor techniques. At 298 and 220 K the present
results are within 4 and 9%, respectively, of the combined data.
These differences are within the expected experimental errors.
The average value of k1 obtained near 220 K by the flash
1
9-22
photolysis studies,
after corrections for the small temper-
ature differences and for 70 ppm C2H6 impurity in ref 20, is
only about 6% higher than the present result. Estimates of
vibrational quenching using the rate constants of Siddles et al.38
show that, for bath gas pressures greater than 20 Torr, the
Comparison with Theory and Earlier Experimental Re-
sults. The preexponential factor in the Arrhenius expression (eq
) is in reasonably good agreement with the approximate value,
9

7468 J. Phys. Chem. A, Vol. 103, No. 37, 1999
Wang and Keyser
lifetime of vibrationally excited CH4 is less than about 1 ms.
22
Manning and Kurylo used 5 Torr of argon as a bath gas, which
gives vibrational lifetimes around 8 ms. These estimates
represent upper limits to the lifetimes since CH4 itself is an
3
8
efficient vibrational quencher; the good agreement among all
of the flash photolysis results suggests that sufficient CH4 was
used to preclude interference from vibrationally excited CH4.
Numerical simulations of the ground-state chemistry in the flash
photysis studies were done by using the CKS algorithm and
the reactions in Table 5 but with rate constants and initial
concentrations set at those used in the actual experiments. The
results show no apparent interference from secondary chemistry,
and the remaining differences may be due to combined
experimental errors.
The competitive chlorination experiments17 should not be
subject to interference from vibrationally excited CH4 because
of the high pressures used (high quenching rates), the length of
time the reactants are held at low temperature with no addition
of room-temperature CH4, and the low level of photolysis
radiation injected. Numerical simulations of these experiments
were carried out by using the CHEMRXN simulator. The
reactions in Table 5 were used with the addition of a Cl2
photolysis step, the reaction of Cl with C2H5Cl, and the back
reactions of HCl with CH3 and C2H5 radicals; the wall reactions
were not included. Rate constants and conditions were set similar
to the actual experiments. The model output values for the rate
constant ratio, R ≡ k(Cl + CH4)/k(Cl + C2H6), differed from
the input ratios by less than 1% at 296 and 220 K. This indicates
that there was no interference from the ground-state chemistry
included in the model. At 220 K our results are about 25%
higher than the competitive chlorination results. If the errors in
the ratio, R, are combined with errors in the reference rate
constant,
k(Cl + C2H6), the competitive chlorination results overlap the
present determination at the one sigma level. The reason for
the 25% difference remains unclear; but, as the present study
shows, it cannot be due to lack of spin or vibrational equilibra-
tion nor to known secondary chemistry.
Summary
The present study clearly demonstrates that the discharge flow
measurements of the Cl + CH4 reaction are not subject to
systematic errors arising from nonequilibration of Cl atom spin
states, vibrational excitation of CH4, or secondary chemistry.
Addition of CF4 at concentrations sufficient to bring about and
maintain spin equilibrium resulted in no observable change in
the Cl + CH4 rate constant. CH4 was vibrationally deactivated
by precooling it before adding it to Cl atoms; the extent of the
deactivation was checked by using numerical simulations with
known vibrational quenching rate constants. Secondary chem-
istry due to impurities in the Cl atoms was checked by using
both a microwave and a thermal source; impurities in the CH4
source were minimized by using an in situ purification method.
No significant difference was found in the results using either
Cl atom source or purification of CH4. The secondary chemistry
of CH3 radicals produced in the Cl + CH4 reaction was carefully
examined by variation of the reactant concentrations in com-
bination with numerical models. At the concentrations used in
the present study, the results show that Cl + CH3 interference
is less than 3%; and in order to prevent interference at the 10%
1
Figure 6. Increase in observed k due to secondary chemistry as
1
1
-3
level, [Cl]o should be less than about 1 × 10 atoms cm .
predicted by numerical simulations; contours are percent changes
calculated from the relation: del% ) (kout/kin - 1) × 100%, where kin
and kout are the input and output of the model; symbols are experimental
data and have the same meaning as in Figure 2. Panel (a), 298 K; panel
Acknowledgment. The research described in this article was
performed at the Jet Propulsion Laboratory, California Institute
of Technology, under a contract with the National Aeronautics
(b), 261 K; panel (c), 218 K.

2
Kinetics of the Cl( PJ) + CH4 Reaction
J. Phys. Chem. A, Vol. 103, No. 37, 1999 7469
and Space Administration. We thank W. B. DeMore for helpful
discussions on this work.
(20) Watson, R.; Machado, G.; Fischer, S.; Davis, D. D. J. Chem. Phys.
976, 65, 2126.
1
(
(
(
(
(
(
21) Ravishankara, A. R.; Wine, P. H. J. Chem. Phys. 1980, 72, 25.
22) Manning, R. G.; Kurylo, M. J. J. Phys. Chem. 1977, 81, 291.
23) Knox, J. H. Chem. Industry 1955, Dec. 10, 1631.
24) Duncan, W. T.; Truong, T. N. J. Chem. Phys. 1995, 103, 9642.
25) Kandel, S. A.; Zare, R. N. J. Chem. Phys. 1998, 109, 9719.
26) Appelman, E. H.; Clyne, M. A. A. J. Chem. Soc., Faraday Trans.
References and Notes
(
1) Brasseur, G.; Solomon, S. Aeronomy of the Middle Atmosphere;
Reidel: Dordrecht, 1984; Chapter 5.
2) Mickley, L. J.; Abbatt, J. P. D.; Frederick, J. E.; Russell, J. M., III
J. Geophys. Res. 1997, 102, 21479.
3) Rinsland, C. P.; Gunson, M. R.; Salawitch, R. J.; Michelsen, H.
(
1
975, 1 (71), 2072.
27) Baulch, D. L.; Duxbury, J.; Grant, S. J.; Montague, D. C. J. Phys.
Chem. Ref. Data 1981, 10.
28) Seeley, J. V.; Jayne, J. T.; Molina, M. J. Int. J. Chem. Kinet. 1993,
5, 571.
(
(
A.; Zander, R.; Newchurch, M. J.; Abbas, M. M.; Abrams, M. C.; Manney,
G. L.; Chang, A. Y.; Irion, F. W.; Goldman, A.; Mahieu, E. Geophys. Res.
Lett. 1996, 23, 2365.
(
2
(
(
(
(
29) Kaufman, F. Prog. React. Kinet. 1961, 1, 1.
30) Keyser, L. F. J. Phys. Chem. 1984, 88, 4750.
31) Brown, R. L. J. Res. Natl. Bur. Stand. (U.S.) 1978, 83, 1.
32) Perry, R. H., Chilton, C. H., Eds.; Chemical Engineer’s Handbook,
(
4) Douglass, A. R.; Schoeberl, M. R.; Stolarski, R. S.; Waters, J. W.;
Russell, J. M., III; Roche, A. E.; Massie, S. T. J. Geophys. Res. 1995, 100,
3967.
5) WMO Scientific Assessment of Ozone Depletion: 1994; World
1
(
5
th ed.; McGraw-Hill: New York, 1973; p 3-230 f.
33) Tyndall, G. S.; Orlando, J. J.; Kegley-Owen, C. S. J. Chem. Soc.,
Faraday Trans. 1995, 91, 3055.
34) Clark, R. H.; Husain, D. J. Chem. Soc., Faraday Trans. 2 1984,
0, 97.
Meteorological Organization Global Ozone Research and Monitoring
Project, Report No. 37, Chapter 3.
(
(
6) Prather, M.; Jaffe, A. H. J. Geophys. Res. 1990, 95, 3473.
(
(7) McElroy, M. B.; Salawitch, R. J.; Wofsy, S. C. Planet. Space Sci.
8
1
1
988, 36, 73.
(
(
(
35) M u¨ ller-Markgraf, W.; Rossi, M. J. J. Phys. Chem. 1991, 95, 825.
36) Clark, R. H.; Husain, D. J. Photochem. 1983, 21, 93.
37) Espinosa-Garcia, J.; Corchado, J. C. J. Chem. Phys. 1996, 105,
(8) Salawitch, R. J.; Wofsy, S. C.; McElroy, M. B. Geophys. Res. Lett.
988, 15, 871.
(9) Bergamaschi, P.; Br u¨ hl, C.; Brenninkmeijer, C. A. M.; Saueressig,
3
517.
G.; Crowley, J. N.; Grooss, J. U.; Fischer, H.; Crutzen, P. J. Geophys. Res.
Lett. 1996, 23, 2227.
(38) Siddles, R. M.; Wilson, G. J.; Simpson, C. J. S. M. Chem. Phys.
1
994, 188, 99.
(10) Saueressig, G.; Bergamaschi, P.; Crowley, J. N.; Fischer, H.; Harris,
(
(
39) Perrin, M. Y.; Jolicard, G. Chem. Phys. Lett. 1986, 127, 118.
40) Timonen, R.; Kalliorinne, K.; Koskikallio, J. Acta Chem. Scand.
G. W. Geophys. Res. Lett. 1996, 23, 3619. Saueressig, G.; Bergamaschi,
P.; Crowley, J. N.; Fischer, H.; Harris, G. W. Geophys. Res. Lett. 1995,
1
986, A 40, 459.
2
2, 1225.
11) M u¨ ller, R.; Brenninkmeijer, C. A. M.; Crutzen, P. J. Geophys. Res.
Lett. 1996, 23, 2129.
12) DeMore, W. B.; Golden, D. M.; Hampson, R. F.; Howard, C. J.;
(41) Timonen, R. S.; Gutman, D. J. Phys. Chem. 1986, 90, 2987.
42) IBM, CKS Chemical Kinetics Simulator, IBM Almaden Research
(
(
Center; IBM Corporation 1995 (Internet address is http://eagle.almaden.
ibm.com/st/).
(
Kolb, C. E.; Kurylo, M. J.; Molina, M. J.; Ravishankara, A. R.; Sander, S.
P. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling, EValulation No. 12; Jet Propulsion Laboratory, California Institute
of Technology: Pasadena, CA, 1997; JPL Publication 97-4.
(
43) Braun, W.; Herron, J. T.; Kahaner, D. K. Int. J. Chem. Kinet. 1988,
0, 51.
(44) Keyser, L. F. CHEMRXN, unpublished work.
(45) Slagle, I. R.; Gutman, D.; Davies, J. W.; Pilling, M. J. J. Phys.
2
(13) Atkinson, R.; Baulch, D. L.; Cox, R. A.; Hampson, R. F., Jr.; Kerr,
J. A.; Rossi, M. J.; Troe, J., J. Phys. Chem. Ref. Data 1997, 26, 521 and
earlier references cited therein.
Chem. 1988, 92, 2455.
(46) MacPherson, M. T.; Pilling, M. J.; Smith, M. J. C. Chem. Phys.
Lett. 1983, 94, 430.
(47) Kaiser, E. W.; Rimai, L.; Wallington, T. J. J. Phys. Chem. 1989,
93, 4094.
(
14) Zahniser, M. S.; Berquist, B. M.; Kaufman, F. Int. J. Chem. Kinet.
1
978, 10, 15.
(15) Keyser, L. F. J. Chem. Phys. 1978, 69, 214.
(
16) Seeley, J. V.; Jayne, J. T.; Molina, M. J. J. Phys. Chem. 1996,
(48) Mallard, W. G.; Westley, F.; Herron, J. T.; Hampson, R. F.; Frizzell,
D. H. NIST Chemical Kinetics Data Base 1998, Version 2Q98.
(49) Nyman, G.; Yu, H.-G.; Walker, R. B. J. Chem. Phys. 1998, 109,
5896.
(50) Dobbs, K. D.; Dixon, D. A. J. Phys. Chem. 1994, 98, 12584.
(51) Truong, T. N.; Truhlar, D. G. J. Chem. Phys. 1989, 90, 7137.
(52) Simpson, W. R.; Rakitzis, T. P.; Kandel, S. A.; Lev-On, T.; Zare,
R. N. J. Phys. Chem. 1996, 100, 7938.
1
1
1
00, 4019.
(17) Lin, C. L.; Leu, M. T.; DeMore, W. B. J. Phys. Chem. 1978, 82,
772.
(18) Heneghan, S. P.; Knoot, P. A.; Benson, S. W. Int. J. Chem. Kinet.
981, 13, 677.
(19) Whytock, D. A.; Lee, J. H.; Michael, J. V.; Payne, W. A.; Stief, L.
J. J. Chem. Phys. 1977, 66, 2690.