Tunable diode laser study of the reaction OH + ClO → HCl + O2

Article abstract of DOI:10.1021/jp013410y

The main pathways for the formation of HCl in the stratosphere are through the reaction of chlorine atoms with methane and formaldehyde: Cl + CH₄ → HCl + CH₃. The production of HCl in the reaction of OH and ClO radicals was measured using time-resolved tunable diode laser spectroscopy. OH and ClO radicals were formed from the 308 nm laser photolysis of mixtures containing O₃, Cl₂, and H₂O. The rate coefficient of 1.25 × 10^-12 cc/molecule-sec for the reaction at 298 K was derived. This reaction rate corresponded to a branching fraction of 6.5% relative to the currently recommended rate coefficient for the overall reaction, indicating that the reaction has a significant impact on the partitioning of chlorine compounds in the stratosphere.

Full text of DOI:10.1021/jp013410y

J. Phys. Chem. A 2002, 106, 1567-1575
1567
Tunable Diode Laser Study of the Reaction OH + ClO f HCl + O2
†
G. S. Tyndall,* C. S. Kegley-Owen, J. J. Orlando, and A. Fried
Atmospheric Chemistry DiVision, National Center for Atmospheric Research, P.O. Box 3000,
Boulder, Colorado 80307
ReceiVed: September 5, 2001; In Final Form: December 6, 2001
Time-resolved tunable diode laser spectroscopy has been used to measure the production of HCl in the reaction
of OH and ClO radicals. The OH and ClO radicals were formed from the 308 nm laser photolysis of mixtures
3 2 2
containing O , Cl , and H O. Computer simulations of the chemistry were used to derive a rate coefficient of
-
12
3
-1 -1
(
1.25 ( 0.45) × 10 cm molecule
s
for the title reaction at 298 K. This reaction rate, which corresponds
to a branching fraction of (6.5 ( 3.0)% relative to the currently recommended rate coefficient for the overall
reaction, confirms that the reaction will have a substantial impact on the partitioning of chlorine compounds
in the stratosphere.
Introduction
occurs with a yield of only a few percent, it is capable of
exerting an influence on the ozone levels. Inclusion of the
The concentration of ozone in the stratosphere is governed
by a delicate balance between its production through oxygen
photolysis and its loss via catalytic free radical cycles. The major
cycles involve nitrogen oxides (NO and NO2), odd hydrogen
radicals (OH and HO2), and halogen-containing radicals (Cl,
ClO, Br, and BrO). The halogen species in the stratosphere are
predominantly man-made and consequently have been changing
on a much faster time scale than the other species. In the mid-
stratosphere, the major loss of ozone is currently due to chlorine
reaction in atmospheric models was found to partially account
for the underprediction of ozone at 40 km altitude and also to
improve model estimates of the magnitude of the observed
5
-8
decadal ozone trends.
Stronger evidence for the occurrence of channel 3b came in
the way of satellite measurements of the partitioning between
members of the chlorine family, ClO, ClONO2 and HCl. Such
measurements indicated a substantially improved fit for models
8
-11
atom chemistry.1-3 Loss of ozone is tempered by the formation
of reservoir species which temporarily convert the active species
into an inactive form. In the case of chlorine, the reservoirs are
if reaction 3b was included with a branching ratio of ∼7%.
It has also been pointed out that reaction 3b can play a role in
the rate of recovery of the ozone hole by accelerating the
conversion of active chlorine to HCl under denitrified conditions,
when chlorine nitrate concentrations are low and ClO high.12
(in order of decreasing lifetime) HCl, ClONO2 and HOCl. The
first of these, HCl, does not undergo photolysis to any extent
in the lower- to mid-stratosphere and reacts only slowly with
OH, and so is a particularly effective sink. The main pathways
for formation of HCl in the stratosphere are through reaction
of chlorine atoms with methane and formaldehyde:
Laboratory measurements of the rate coefficient of reaction
1
3-21
3
have been available since the late 1970s.
The most recent
1
9-21
determinations
seem to indicate a higher rate coefficient
than thought previously. A number of the early studies
13,15,16
quantified k3a via HO2 production, inferring branching fractions
k3b/k3 that were less than 0.15. The exception was the study of
Cl + CH f HCl + CH3
(1)
(2)
4
1
7
Poulet et al., in which HCl yields of <2% were reported.
However, large corrections had to be made for secondary
production of HCl by the reaction Cl + HO2.
Cl + HCHO f HCl + HCO
Any reaction which converts active chlorine species to HCl will
thus be of potential importance in the mid-stratosphere. The
reaction of OH with ClO (3) offers a possible mechanism for
this to occur. The major reaction pathway forms HO2 + Cl,
but it is also thermodynamically possible to form HCl + O2 in
More recently, three groups have reported direct studies of
1
8,22
reaction 3b. Lipson et al.
used a high-pressure flow tube
with chemical ionization mass spectrometry and found yields
of 7% relative to their measurement of k3. Wang and Keyser23
used a low-pressure discharge-flow system with tunable diode
laser detection and reported an absolute HCl yield (independent
-
1
reaction 3b, which is exothermic by 56 kcal mol to produce
3
-1
1
O2( Σ), or by 33 kcal mol to produce O2( ∆):
2
0
of k3) of 9%. Most recently, Bedjanian et al. reported a yield
of 3.5% at room temperature. All these studies used flow tubes,
which can be subject to loss or production of “sticky”
compounds such as HCl. The present study uses a different
approach, by producing HCl in a flash system and observing
its formation on a time scale faster than diffusion to or from
the walls of the vessel. In this sense, it can be considered a
OH + ClO f Cl + HO2
(3a)
(3b)
OH + ClO f HCl + O2
The latter pathway removes members of both the odd hydrogen
and active chlorine families. Brasseur and co-workers were
among the first to suggest in the mid-1980s that if channel 3b
4
“
wall-less” reactor. Both OH and ClO radicals were produced
by the same excimer laser pulse, and HCl was detected by time-
resolved tunable diode laser spectroscopy in direct absorption.
Diode laser absorption is a sensitive and specific technique that
can be used in time-resolved mode for either free radicals or
*
Corresponding author. Tel: 303-497-1472. Fax: 303-497-1411. E-
mail: tyndall@acd.ucar.edu.
†
Present address: Department of Chemistry, University of Nebraska,
Kearney, NE 68849.
1
0.1021/jp013410y CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/02/2002

1568 J. Phys. Chem. A, Vol. 106, No. 8, 2002
Tyndall et al.
Figure 1. Schematic of experimental setup. The laser diode is housed in the liquid nitrogen dewar pointing downward. A gold-coated off-axis
ellipsoidal mirror located under the dewar collects and focuses the beam, and the second one (shown) collimates it through the cell.
stable products.2
4-26
The final value for the rate coefficient of
liquid nitrogen. The exact temperature and current of the diode
(typically 81 K and 0.35 A) were adjusted externally to control
the output wavelength of the laser, which was swept by applying
a sawtooth ramp to the current at a frequency of 1 kHz. The
ramp provided both the spectroscopic frequency axis within a
given sweep, and the kinetic time base for the experiments via
the generation of consecutive sweeps through the HCl line 1
ms apart. The total DC laser power reaching the detector was
measured several times a day by inserting a mechanical chopper
at the primary focus of the beam (see Figure 1). The mode purity
reaction 3b was obtained by computer modeling of the total
amount of HCl produced, taking into account primary and
secondary HCl sources.
Experimental Section
The experiments involved using tunable diode laser (TDL)
spectroscopy to measure the HCl produced by pulsed excimer
24
laser photolysis. The experiments were carried out in a 47.7
cm jacketed cell, 5 cm in internal diameter, fitted with wedged
BaF2 windows (Figure 1). External dielectric-coated mirrors
(
which was ∼96%) was checked from time to time by flowing
HCl through the photolysis cell at a high enough concentration
(
Creative Optics, Texas) reflected the TDL beam twice through
1
4
-3
(>10 molecule cm ) to absorb >98% of the laser light.
the cell to give an optical path length of 95.4 cm. The dielectric
mirrors transmitted UV radiation, allowing photolysis of the
gas mixture down the length of the cell. After passing through
the cell, the TDL beam was detected on a liquid nitrogen cooled
InSb detector/preamplifier (Cincinnati Electronics SDD-7854).
A band-pass filter positioned after the TDL focusing optics
protected the diode from scattered UV light. The filter, set at
Secondary checks of mode purity were made at the start and
finish of each day using a sealed cell containing a known amount
of OCS.
The output of the laser was swept through the R(1) absorption
line of HCl, allowing a differential (peak to baseline) signal to
be measured in direct absorption. Sweeping the diode output
reduced uncertainties potentially caused by transient (<2 ms)
fluctuations in diode laser wavelength caused by electrical pick-
up from the excimer laser pulse, or offset on the detector due
to scattered excimer radiation. The sweep of the diode laser
was synchronized to the excimer laser trigger so that the
absorption was always measured at the same delay times relative
to the excimer laser. Hence, the signal could be averaged over
consecutive excimer laser flashes, and by averaging 50 traces,
4
5° to the path of the laser beam, also served to reflect a small
fraction of the beam onto a second InSb detector which was
used to remove baseline fluctuations. Both detectors were fitted
with long-pass filters to remove scattered excimer light.
Transient signals from the detectors were averaged in a LeCroy
450 digital oscilloscope and sent to a PC for analysis.
The radiation used to detect the reaction product HCl was
9
generated using a quaternary lead-salt diode laser, cooled by

OH + ClO f HCl + O2
J. Phys. Chem. A, Vol. 106, No. 8, 2002 1569
peak-to-peak noise levels of 1 × 10-5 absorbance units could
the temperature of which was constantly monitored. At such
flow rates, it can safely be assumed that the helium was fully
9
be obtained, corresponding to about 3 × 10 HCl molecule
-
3
27
cm . The ultimate detection limit was governed by drifts in
the laser output frequency, which caused the noncancellation
of interference fringes.
saturated with H2O vapor. The flow of water vapor was
2
8
calculated using tabulated vapor pressures and the measured
total pressure in the aspirator. The water vapor concentration
was also measured in the reaction cell using absorption of the
The sweep current imposed on the diode corresponded to an
-1
27
amplitude of ∼0.11 cm . The change in laser power associated
with the sweep resulted in a sloping background which was
often larger than the HCl absorption feature, limiting the number
of bits available for recording the absorption. Two approaches
were used to remove this background modulation. For HCl
calibration experiments involving the photolysis of Cl2-C2H6-
O2, concentrations of HCl in excess of 1012 molecule cm were
produced, and a single detector was used. Averages were
alternately taken with Cl2 present (signal) and without Cl2
mercury emission line at 184.9 nm as described previously;
the measured water vapor agreed with that calculated from its
vapor pressure to within 2%. Chlorine from a lecture bottle was
made into a 10% mix in helium; the mole fraction of Cl was
2
occasionally checked off-line using a UV-visible diode array
spectrometer. A flow of 40-80 sccm ultrahigh purity O was
2
-3
passed through a dry ice trap and then through an ozonizer which
was known to produce 3.8% O . The ozone mixing ratio in the
3
absorption cell was also checked daily using absorption of a
(
background), and the background trace was digitally subtracted
254-nm Penray Hg lamp. UV absorption cross sections and
quantum yields for Cl and O were taken from the NASA JPL
from the signal trace to remove the slope. For the measurement
of the very small amounts of HCl in the OH + ClO experiments,
two detectors were used. Part of the beam was picked off before
entering the cell and sent to the reference detector. The output
from both detectors was fed to a differential amplifier which
subtracted the signals to remove the slope in real time. The
resultant differential signal was amplified by a factor of 67, and
the amplified signal was fed to the digital oscilloscope for
averaging. Again, averages with and without Cl2 present were
subtracted digitally. Extensive tests showed that the amount of
HCl measured in both direct and differentially amplified modes
agreed in both peak height and integrated area and that the
amplifier did not cause any clipping of the signal.
Even after averaging and digital subtraction, small fluctuations
remained in the baseline associated with residual 120 Hz
vibrations. These oscillations could be removed by co-adding
the individual sweeps to provide one pre-photolysis and one
post-photolysis measurement of HCl. Thus, the primary mea-
sured quantity was the change in HCl produced by a given [OH]
and [ClO].
2
3
2
9
compilation. Pressures in the reaction cell and the water
bubbler were measured using capacitance manometers. Calcu-
lated residence times in the cell were 3-5 s. Flows were
maintained using mass flow controllers that had been calibrated
using a commercial bubble meter for the exact gases and range
2
of flow rates used. Gases used, with purities were: Cl (UHP
99.5%, Matheson), He (UHP, US Welding), O (UHP, US
2
Welding), and H O (deionized in house, 18 Mohm).
2
Since the branching ratio to produce HCl was expected to
be small, the utmost care had to be taken to minimize
background HCl in the gas flow and also any HCl produced by
secondary reactions. The Cl2 was found to contain about 100
ppm HCl, which resulted in background levels roughly 10 times
higher than the amount produced from the reaction. Initial
attempts to purify the Cl2 by distillation were not very
successful. A better solution was found to be the inclusion in
the flow path of a short glass trap containing a small amount
(about 100 mg) of Hopcalite catalyst (Cu-Mn-O) held between
glass wool plugs. By trial and error, it was found that an
appropriate amount could be used that removed essentially 100%
of the HCl while still passing the Cl2 (as verified by UV
spectroscopy). Since Hopcalite is used in atmospheric chemistry
field experiments as an ozone “killer”, it was obviously
necessary to avoid passing the ozonized O2 through the trap
The reaction was initiated by a pulse from an excimer laser
Questek model 2440) operating at the 308 nm XeCl line. The
(
excimer laser was positioned several meters from the reaction
cell, so the UV beam was wide and filled the cell fairly
uniformly. The laser power was measured using a pyroelectric
power meter that was cross-calibrated against two meters of
the disk-calorimeter type used by the NOAA Aeronomy lab
(
map out and measure the fluence directly. Fluctuations over
the beam profile were less than 5%. Excimer laser powers of
1
1
(see Figure 1).
HCl was detected using the R(1) line of the fundamental
2
R. Talukdar, Pers. comm.). A 4-cm square mask was used to
3
5
-1
stretching mode of H Cl at 2925.897 cm . The integrated line
3
0
strength SJ was taken from the HITRAN listing, which is in
turn based on the work of Pine et al. For the R(1) line of the
3
1
00-140 mJ/pulse were used, which resulted in fluences of
-
19
-1
-
2
HCl fundamental, SJ is 4.19 × 10
cm molecule , with an
.0-1.5 mJ cm measured behind the cell.
uncertainty of 2-3%. The HCl was quantified using both
integrated line strength and peak-to-baseline cross sections (see
next section). The absolute HCl could be compared to the
amount of HCl produced in situ by the reaction of Cl atoms
with C2H6. The frequency of the diode laser output was
calibrated relative to well-characterized OCS absorption lines.
The reactants ClO and OH radicals were produced concur-
rently by the flash photolysis of mixtures containing Cl2, O3,
and H2O
Cl + hν f Cl + Cl
(4)
2
1
O + hν f O + O( D)
(5a)
(5b)
(6)
For the majority of the experiments, a time base of 20 ms
was used with a 1 kHz sweep rate of the diode laser, and 800
digitized pixels were recorded (25 µs per pixel). Six diode laser
sweeps were acquired prior to the excimer pulse and 12 after.
Due to the rapidity of the chemistry, this only allowed the net
change of HCl to be measured. For some experiments, however,
the sweep rate was increased to 4 kHz to allow time-resolved
HCl profiles to be acquired. The repetition rate of the excimer
laser was 0.2 Hz, allowing the contents of the cell to be swept
out between pulses. This was necessary to avoid the build-up
of both HCl and ClO (which reacts only slowly with itself).
3
2
3
O + hν f O + O( P)
3
2
Cl + O f ClO + O2
3
1
O( D) + H O f OH + OH
(7)
2
The main carrier gas was a flow of helium (250-400 sccm),
which was passed through an aspirator near room temperature
to saturate it with water. The aspirator stood in a water bath,

1570 J. Phys. Chem. A, Vol. 106, No. 8, 2002
Tyndall et al.
The Cl2-C2H6 experiments could be used to show that HCl
from the previous pulse had been removed before data acquisi-
tion began and that any residual HCl observed was residual in
the Cl2 flow breaking through the Hopcalite filter.
could be calculated for any given reaction mixture, and the
corresponding cross section obtained:
γ P
γ P
γH2OPH2O
He He
O2 O2
Pequ
)
+
+
(B)
γN2
γN2
γN2
Results
Calibration of HCl. The most critical parameters in the
determination of k3b are the quantification of HCl and the
measurement of the excimer laser fluence (i.e., the determination
of OH and ClO). Particular care was taken to characterize the
cross section of HCl under the conditions used, and to relate
the excimer laser fluence to the amount of HCl produced in a
known reaction. In this way, the system was internally cali-
brated, and systematic errors were reduced. The HCl concentra-
tion was determined using either peak cross section or integrated
line strengths.
-1
Broadening coefficients γ (half-width at half-maximum in cm
-1
atm ) for nitrogen and air are known with high accuracy as a
function of rotational quantum number, J, from the work of
Pine and Looney, from which a half-width for oxygen can be
obtained. However, since most of the broadening in air is due
to N2, values of γO2 obtained in this manner may not be very
accurate. Babrov et al. measured broadening half-widths for
32
33
pure O2 and N2, as a function of J at low resolution, and found
a smooth trend of γ with rotational quantum number. Hence,
the values for O2 deduced from Pine and Looney were smoothed
according to the trend given by the Babrov data. Broadening
coefficients for HCl lines by helium, γHe, have been measured
by Rank et al. and Babrov et al. These values were used as
the initial guesses for this work but optimized as described
below.
Concentrations of HCl in different gas mixtures were derived
from the integrated absorbance, as described in the previous
section, and a peak absorption cross section was derived. A cross
section was then calculated from the nonlinear fit of cross
section versus N2 pressure using an equivalent N2 pressure as
defined in eq B. The broadening coefficients were then varied
to simultaneously mimimize the differences between all the
measured and calculated cross sections. Such line width
measurements were made using both flowing mixtures of
authentic HCl and HCl produced in situ from the photolysis of
Cl2-C2H6-O2-He mixtures. In this way, broadening coef-
ficients for H2O could be determined and those for He and O2
refined. The half-widths for broadening by He, O2 and H2O
Measurement of HCl Using Integrated Line Strengths. The
concentration of HCl was most readily derived experimentally
from the integrated area under the resolved rotational line used.
Raw voltages were converted to absorbance, A, using the
chopped single-mode laser power and the Beer-Lambert law.
The tuning rate of the diode laser with current was measured
using a germanium etalon with a free spectral range of 0.015 98
34
33
-1
cm , calculated from the refractive index at the exact frequency
and laboratory temperature used. The HCl R(1) line was
recorded using an expanded oscilloscope time base with 400
pixels per sweep, enabling the peak to be fully resolved. The
absorbance was integrated over a defined number of pixels, dp,
and converted to concentration using equation A
1
∂ω
A0
[HCl] )
∫
A dp )
(A)
SJl ∂p
σ l
0
where SJ is the integrated line strength, l is the path length,
ω/∂p is the diode tuning rate per pixel, A0 is the absorbance at
-
1
-1
were found to be 0.026, 0.052, and 0.22 cm atm , respec-
tively, which can be compared to the literature values for He
of 0.022 and 0.020 and for O2 of 0.041 and 0.051. The
∂
line center, and σ0 the peak absorption cross section.
3
4
33
32
33
For the kinetics measurements, a time base of 20 ms per trace
was used, so multiple sweeps through the HCl line were
recorded, separated by 1 ms. Using this slower time base, about
coefficient γH2O is similar in magnitude to that for self-
-
1
-1 30
broadening of HCl, 0.24 cm atm . The broadening coef-
ficients are expected to be accurate to about 20%. It should be
noted, though, that values for γ were only determined under
conditions closely resembling those where experiments were
conducted and that they were only used to interpolate measured
cross sections by a few percent to match experimental condi-
tions. Hence, the results are not sensitive to the exact values of
the broadening coefficients. The calculated Doppler-limited peak
1
5 pixels were recorded across the HCl absorption line. The
area under the absorption feature was integrated using the central
1 pixels, defining a baseline starting 7 pixels on either side of
1
the peak. By comparing traces recorded under the same
conditions but with slow and fast time bases, it was shown that
the area defined in the kinetics scans captured >95% of the
true area for all except the experiments at 40 Torr with added
H2O, where the line width is largest. For such cases a correction
was made for the “missing” area. The sweep routine was also
tested using known flows of CH4, which is not lost in flow
controllers. The CH4 concentrations derived using the HITRAN
-
17
2
cross section of HCl R(1) at 296 K is 6.51 × 10
cm
-
1
molecule . The measured HCl cross sections under the
-
17
2
experimental conditions ranged from 4.85 to 3.55 × 10 cm
-
1
molecule , depending on the total pressure and the amount of
water vapor present.
-
1
line strengths (four lines near 2926.8 cm ) agreed with those
from flows to within a few percent. The measured areas were
also found to be independent of whether the differential amplifier
was used or not.
In Situ Calibration of HCl. The reaction of Cl atoms with
C2H6 was used to measure the amount of HCl produced by a
given concentration of Cl2 and excimer power. This reaction is
known to be rapid and provided quantitative conversion of Cl
atoms to HCl under the exact conditions of the OH + ClO
experiment
Calculation of HCl Cross Sections. The concentration of
HCl was also derived using peak absorption at line center. First,
the peak HCl absorption cross section was calculated at 10
different pressures of N2 between 0 and 40 Torr using a Voigt
profile and the broading coefficients given by Pine and
Looney.32 The variation of this cross section with pressure was
then fitted with a second-order polynomial to allow the cross
section to be interpolated at an arbitrary pressure. Then, using
known or measured broadening parameters γ (as described
below) for He, O2, and H2O, an equivalent pressure of N2, Pequ,
Cl + C H f HCl + C H
5
(8)
(9)
2
6
2
C H + O + M f C H O + M
2
5
2
2
5
2
The HCl concentration derived from the TDL absorption
measurement was corrected for a small amount of regeneration

OH + ClO f HCl + O2
J. Phys. Chem. A, Vol. 106, No. 8, 2002 1571
Figure 2. Comparison of measured (using infrared absorption) and
calculated (from the excimer laser fluence) HCl concentration produced
in the flash photolysis of a Cl
pressure. The measurements were made both in the absence (filled
symbols) or presence (open symbols) of H
2
-C
2
H
6
-O -He mixture at 40 Torr total
2
16
-3
2
O (9.9 × 10 molec cm ).
The peak HCl absorption cross sections derived from the measurements
-17
2
-1
-17
2
-1
were 4.3 × 10 cm molecule (dry) and 3.6 × 10 cm molecule
with water).
(
of Cl atoms by the reaction of C2H5 + Cl2 as described
previously.24
The amount of HCl measured using the TDL was always
12-15% higher than that calculated from the excimer laser
fluence and the Cl2 concentration. Since the excimer fluence
may not be perfectly spatially homogeneous and reflections
occur inside the cell, the measured HCl was used to quantify
the fluence, rather than the HCl being obtained from the
measured laser power. The Cl2-C2H6 experiments were not
performed every day, to reduce the risk of having residual C2H6
in the system, which could react with Cl to produce HCl. Day-
to-day variations in the total excimer power exiting the cell were
normalized to the previous HCl calibration. Figure 2 shows a
plot of measured HCl versus calculated HCl for conditions of
Figure 3. Measurement of HCl formed in the flash photolysis of Cl
2
15
15
16
-3
(5.3 × 10 ), O
3
(6.5 × 10 ), and H
2
O (9.6 × 10 ) molecule cm at
a total pressure of 42.9 Torr (5.1 Torr O
2
, balance He). (a) Time series
10
-3
showing production of 3.0 × 10 molecule cm of HCl. The excimer
laser fired at 6.8 ms. Average from 60 excimer pulses. (b) Composite
showing spectrally resolved HCl peak, obtained by subtracting co-added
sweeps before excimer pulse from co-added sweeps after excimer pulse.
40 Torr total pressure in the presence and absence of about 3
Torr water vapor. The HCl concentration measured by integrated
area agrees in both cases, but is about 13% higher than
calculated from the excimer fluence measured behind the cell.
be avoided. The reaction of Cl atoms with impurities was shown
not to be a serious problem from experiments involving the
12
photolysis of Cl -O -H O. In the presence of about 5 × 10
2
2
2
12
-3
-3
10
-3
The precision in measuring ∼10 molecules cm of HCl was
Cl atoms cm , less than 5 × 10 molecule cm of HCl were
produced. Assuming that the chlorine atoms were in excess,
the latter value gives an estimate of the impurity (presumably
organic) present
1
0
-3
on the order of 1 × 10 molecules cm .
Measurement of OH + ClO at 298 K. Photolysis of Cl2-
O3-O2-H2O mixtures clearly led to the formation of HCl on
10
-3
the order of 1-4 × 10 molecule cm . Figure 3a illustrates
the time-resolved production of HCl in an experiment at 43 Torr.
The HCl measurements are 1 ms apart, as a result of ramping
the diode current at a frequency of 1 kHz. After the excimer
laser pulse occurs at a delay of 6.8 ms, a clear increase in the
HCl concentration is observed. Figure 3b gives a composite of
the spectrally resolved HCl peak, obtained by co-adding the
diode laser sweeps before and after the excimer pulse and then
subtracting the former from the latter to remove residual baseline
oscillations. The HCl concentration was measured relative to
the gently sloping baseline.
Under the conditions used, HCl can be formed both by the
primary reaction, OH + ClO, and by secondary reactions. Under
the experimental conditions used, secondary production was
similar in magnitude to that from reaction 3b, but the sensitivity
was sufficiently high to determine 3b reasonably accurately.
Secondary production of HCl can occur from two sources:
reaction of Cl with impurities in the system, or from reactions
that are intrinsic to the source chemistry and thus cannot easily
Cl + RH f HCl + R
(10)
In the presence of 1 × 10 molecule cm-3 O3, though, the Cl
atoms would react exclusively with O3, so the HCl would be
well below the detection limit for this study and would be
formed with a time constant on the order of 1 µs, controlled by
Cl + O3. Control experiments were performed using different
combinations of the gases Cl2, H2O, and O3, but no evidence
was found for extraneous HCl sources that would impact the
chemistry of reaction 3. Other secondary sources of HCl, such
as the reaction of Cl with HO2, were shown by computer
modeling to be much smaller than that discussed above, i.e.,
16
8
-3
<
5 × 10 molecule cm
Cl + HO f HCl + O
(11a)
(11b)
2
2
Cl + HO f OH + ClO
2
The production of HCl by intrinsic chemistry was a much more

1572 J. Phys. Chem. A, Vol. 106, No. 8, 2002
Tyndall et al.
TABLE 1: Major Reactions Describing the Production of
HCl
value, since the OH decay is largely controlled by O3 and Cl2.
The excimer fluence used in the model was corrected for
absorption down the length of the cell by Cl2 and O3, the
calculated fluence at the center of the cell being used in the
model. Corrections were always less than 6%. For the Cl2-
C2H6 experiments, the sample was always optically thin at 308
nm, so this gave a measure of the incident fluence.
reaction
rate coefficient at 298 Ka,b
1
O
O
Cl
3
+ hν f O( D) + O
2
Φ(O) ) 0.78
Φ(O) ) 0.22
Φ(Cl) ) 2.0
3
3
+ hν f O( P) + O
2
2
+ hν f Cl + Cl
1
-10
O( D) + H
2
O f OH + OH
2.2 × 10
1
3
-11
O( D) + O
2
f O( P) + O
2
4.0 × 10
Under the conditions used, the decay rate of OH radicals was
largely controlled by the reactions OH + O3 and OH + Cl2,
-
-
11
11
Cl + O
3
f ClO + O
2
1.2 × 10
1.8 × 10
OH + ClO f HO
2
+ Cl
-
1
-
12
and the loss rate for OH decay was between 600 and 1000 s
OH + ClO f HCl + O
2
1.25 × 10 (c)
3
-11
O( P) + OH f H + O
2
3.3 × 10
3
-14
O( P) + Cl
2
f ClO + Cl
O( P) + ClO f Cl + O
4.2 × 10
OH + O f HO + O
2
(16)
(17)
3
2
3
-11
2
3.8 × 10
-
14
OH + Cl
2
f HOCl + Cl
f HO + O
f HCl + O
f OH + ClO
f OH + O
f HCl + Cl
6.3 × 10 (d)
OH + Cl f HOCl + Cl
2
-
-
-
-
-
14
11
12
11
11
OH + O
Cl + HO
Cl + HO
3
2
2
7.8 × 10
3.2 × 10
9.0 × 10
2.9 × 10
2
2
2
In fact, only about 20-40% of the OH actually reacted with
ClO for any experiment. However, this essentially means that
the rate of production of HCl is being measured against the
rate of reaction of OH with Cl2 and O3, two reactions for which
the rate coefficient is reasonably well-known. Thus, the values
of k3b reported here are to a large extent independent of the
value of k3, which is still uncertain to (20%.
Experiments on a given day were conducted with fixed flows
of O3 and H2O, varying the Cl2 over a range of flows (usually
a factor of 3 or 4). The experimental conditions used and results
obtained are summarized in Table 2. In addition to the various
concentrations, Table 2 gives the total change in measured HCl
H + O
H + Cl
3
2
2
1.9 × 10 (e)
a
3
-1 -1 b
Rate coefficient units are cm molecule
s . Taken from ref 29
c
d
e
unless noted. This work. Weighted mean of literature values. From
3
8
39
40
Michael and Lee, Bemand and Clyne, and Wagner et al.
difficult problem to overcome, and the amount of HCl from
these reactions was roughly equal to that produced from reaction
3
3
b. The major source was linked to the presence of O( P) atoms,
which were produced both in the photolysis of O3 (reaction 5b,
1
yield 20%) and by quenching of O( D) by O2
(
estimated precision (20%), the modeled contributions from
1
3
O( D) + O f O( P) + O
(12)
2
2
reactions 3 and 14, and the rate coefficient k3b required to
reproduce the measured [HCl]. The initial [OH], derived from
3
12
-3
O( P) + OH f O + H
(13)
(14)
(15)
the complete model, was around 1.8-2.1 × 10 molecule cm ,
2
12
while the ClO was varied in the range 2.4-9.7 × 10 molecule
-
3
H + Cl f HCl + Cl
cm . The range of k3b reported is fairly small, despite a wide
change in initial [ClO], leading to confidence in the measure-
ments. The unweighted mean of the 27 determinations shown
2
H + O f OH + O2
3
-
12
3
-1 -1
in Table 2 is (1.23 ( 0.25) × 10 cm molecule s , which
1
Quenching of O( D) was minimized by keeping the ratio [H2O]/
O2] as high as possible. Unfortunately, since the rate determin-
ing step in the production of this secondary HCl was the reaction
is a branching ratio of 6.5% expressed relative to the recom-
2
9
[
mended overall rate coefficient k3. Figure 4 shows the
measured HCl as a function of [ClO] at 30 and 40 Torr with
lines indicating the HCl level expected for 0, 5% and 10%
yields. The lines are derived from computer simulations
3
of O( P) with OH, production from H + Cl2 and from OH +
ClO occurred on the same time scale, so computer modeling
was required to derive the rate coefficient. The reactions listed
in Table 1 summarize the major competitive loss reactions of
1
beginning with typical values of O( D) and H2O at each
pressure. The exact HCl produced will depend on the particular
combination of conditions, but the lines give an indication of
the magnitude and trend to be expected with the given variation
in [ClO]. As can be seen, the total HCl produced for a given
O3 and H2O is largely independent of the [ClO]. This can be
rationalized since the OH also reacts with O3 and Cl2, and the
ClO is proportional to the Cl2 concentration. Thus, the partition-
ing of OH reactivity between Cl2 and ClO does not change
markedly as the Cl2 concentration is increased. Likewise, the
secondary HCl from H + Cl2 (given by the lines marked 0% in
Figures) is largely independent of [ClO] because the competition
between O + ClO and O + OH and between H + Cl2 and H
+ O3 tend to offset one another. The fact that the observed
yields of HCl are relatively independent of [ClO] (and hence
[Cl]o) also suggests that the HCl is not being formed in reactions
of Cl atoms with impurities and that secondary chemistry has
been adequately accounted for.
1
O( D), Cl, OH, and H. This subset of reactions essentially
describes the formation of HCl for the conditions used. The
complete reaction scheme used in the numerical model
Acuchem ) contained about 50 reactions, including ground and
electronically excited states of O and O2. Rate coefficients were
taken from the JPL or IUPAC evaluations whenever possible,
and pressure-dependent rate coefficients were estimated for the
appropriate pressure of helium where exact values were not
available. However, the pressure-dependent reactions (H + O2
and O + O2) did not play a great role at 30-40 Torr, so accurate
values of the rate coefficients were not necessary.
3
5
(
29
36
For the reaction of OH with Cl2, a weighted average of the
values in the literature was used, which gave a rate coefficient
-
14
3
-1 -1
of 6.3 × 10
cm molecule
s
(6% greater than the most
3
7
recent measurement by Gilles et al. ). For H + Cl2, the direct
studies of Michael and Lee, Bemand and Clyne, and Wagner
38-40
-11
3
et al. are in good agreement.
A value of 1.9 × 10
was chosen as the average of the room-
temperature rate coefficients from these studies. The overall rate
cm
Reactions of Excited Species. The potential exists for several
species (OH, ClO, HCl) to be produced in vibrationally excited
states, which may have an effect on the chemistry or detection
of these species. The primary source of OH is the reaction of
-
1
-1
molecule
s
-
11
3
-1
coefficient for reaction 3 was set to 1.9 × 10 cm molecule
-
1
29
1
s , the currently recommended value. However, as will be
shown later, the HCl yield is not strongly dependent on this
O( D) with H2O, which produces OH radicals in vibrational
4
1
levels up to V ) 3. The present experiments were conducted

OH + ClO f HCl + O2
J. Phys. Chem. A, Vol. 106, No. 8, 2002 1573
TABLE 2: Summary of Experimental Conditions and Results at 296 K
a
OHb
ClOc
HCld
(OH+ClO)
HCle
(H+Cl )
2
pressure
Torr)
O
3
Cl
2
H
2
O
HClmeas
k
3b
(1015 cm-3
)
(1015 cm-3
)
(1016 cm-3)
(1012 cm-3
)
(1012 cm-3
)
(10 cm-3
10
)
(10-12)
(
3
1
4.69
4.07
5.39
2.73
6.72
8.67
4.06
6.05
4.73
7.38
5.39
4.06
3.77
5.23
7.80
10.3
2.65
5.25
8.73
11.3
7.90
5.33
10.7
5.15
8.51
7.07
3.55
7.90
6.45
6.41
6.49
6.40
6.34
6.43
6.40
6.43
6.38
6.41
6.43
6.50
9.55
9.48
9.44
9.68
9.57
9.55
9.51
9.61
9.73
8.95
8.60
8.51
9.51
9.56
9.45
1.93
1.85
1.95
1.80
1.72
1.92
1.82
1.88
1.77
1.84
1.89
1.99
2.07
1.99
1.92
2.19
2.08
1.99
1.92
2.02
2.11
1.75
1.86
1.75
1.91
2.00
1.88
4.5
5.8
3.1
7.0
8.8
4.5
6.5
5.1
7.6
5.8
4.5
4.3
4.5
6.5
8.5
2.4
4.5
7.1
9.0
6.5
4.5
9.7
5.0
7.9
6.7
3.6
7.4
2.80
3.44
2.82
3.06
3.06
2.65
3.58
3.40
3.62
3.03
2.91
3.05
3.07
2.35
2.85
2.16
3.00
3.09
2.88
2.94
2.84
2.35
2.31
3.20
3.24
2.55
2.55
1.12
1.85
1.26
1.54
1.69
0.99
2.03
1.78
2.14
1.44
1.29
1.31
1.66
0.98
1.58
0.93
1.58
1.74
1.63
1.55
1.41
1.19
0.97
1.92
1.93
1.20
1.28
1.68
1.59
1.56
1.52
1.37
1.66
1.55
1.62
1.48
1.59
1.62
1.74
1.41
1.37
1.27
1.23
1.42
1.35
1.25
1.39
1.43
1.16
1.34
1.18
1.31
1.35
1.27
0.92
1.43
1.27
1.12
1.21
0.82
1.53
1.43
1.51
1.11
1.10
1.06
1.60
0.84
1.28
1.23
1.52
1.44
1.30
1.31
1.33
0.86
0.83
1.55
1.48
1.16
0.95
4
4
4
4
4
4
4
4
4
4
4
.66
.72
.65
.61
.68
.65
.68
.64
.66
.68
.66
4
3
3
8
6.47
6
6
6
6
6
6
6
6
.42
.40
.56
.48
.47
.44
.51
.59
4.95
4
4
4
4
4
.75
.71
.89
.92
.86
a
2
.
OH from model extrapolated back to zero time. ClO from model at roughly half of OH decay. d Modeled HCl produced
b
c
O
3
) 0.038 × O
from OH + ClO (10 cm ). Modeled HCl produced from H + Cl
2
(1010 cm ).
1
0
-3
e
-3
-
1 41-43
s ),
and the OH is expected to be fully relaxed within 2
µs. Some secondary production of OH by H + O3 occurs,
leading to OH excited up to V ) 9. However, higher vibrational
levels of OH should be quenched very rapidly by both O2 and
4
1,43,44
H2O,
and the computer model shows that the amount of
H produced is in any case very small compared to the primary
OH source.
The reaction of Cl atoms with O3 produces vibrationally
45
excited ClO. Quenching of ClO(v) has not been studied widely.
Rate coefficients for vibrational quenching of ClO by Cl2 and
-
12
3
-1 -1
O3 are thought to be around 2 × 10
cm molecule s ,
and complete vibrational quenching of ClO to the V ) 0 state
should be complete for our conditions within 200 µs, based on
4
5
the data in Figure 5 of Matsumi et al. Several groups have
commented on the possibility of the reaction of Cl with
vibrationally excited ClO, which can lead to incomplete
conversion of Cl atoms to ClO when reaction 6 is used.2
Such studies tended to be in low pressure flow tubes. The effect
will probably not be important here as a result of the high overall
pressures and high [O3] used, which leads to a chemical lifetime
for Cl of a few microseconds. Thus, conversion of Cl into ClO
should be quantitative. It is not known what effect vibrationally
excited ClO would have on the OH + ClO reaction, although
vibrational energy would be expected to inhibit the reaction,
since the overall k has a negative activation energy.29
3,46
Finally, the product HCl could be formed in excited vibra-
tional states (V < 4 if singlet O2 is formed). The extended
geometry of the transition state also suggests vibrational
4
7,48
excitation of the HCl product.
Quenching of HCl(v) by H2O
-
11
3
-1 -1 49
is rapid, though, k ) 1.5 × 10 cm molecule s , and no
excited HCl should remain on the time scale of the diode laser
measurements.
Figure 4. Measured [HCl] plotted as a function of [ClO] at (a) ∼30
and (b) ∼40 Torr total pressure. The lines indicate the level of HCl
expected to be formed for 0% (secondary production only), 5%, and
Sensitivity Tests. The most critical experimental parameter
in the determination is probably the measurement of the excimer
laser fluence, since it essentially enters into the rate coefficient
k3b quadratically (since it is linked to both the OH and ClO
concentrations). Thus, even though the fluence is calibrated in
1
1
0% yields of HCl in reaction 3b. The initial [OH] was around 2 ×
12 -3
0
molecule cm .
in the presence of 2 Torr of water vapor. Quenching of OH(v)
by H2O is known to be fast (>1 × 10-11 cm molecule
3
-1

1
574 J. Phys. Chem. A, Vol. 106, No. 8, 2002
Tyndall et al.
situ relative to HCl production, relative errors do not cancel
significant amounts of HCl is in accord with three other recent
20,22,23
(as they would in a straightforward quantum yield experiment)
studies.
The methodology in the current study differs from
since the calibration reaction is first order and the reaction under
study is second order. However, the fact that the HCl line
strength is very well-known constrains the fluence more tightly
than a direct measurement of laser power would.
the previous ones in that a slowly flowing mixture is used with
a pulsed radical production. While wall reactions were mini-
mized in the present study, the presence of large radical
concentrations led to an increased contribution from secondary
chemistry. However, the results agree very well with those
obtained using discharge-flow techniques.
The rate coefficient k3b derived here is in essence measured
relative to OH + Cl2 and OH + O3 and is not particularly
sensitive to the overall value of k3. To test this, computer
simulations were run with k3 10% higher or lower; the calculated
HCl concentration changed by less than 1%. Further tests were
conducted with the rate coefficients for H + Cl2, OH + O3,
and OH + Cl2 adjusted separately and together. As expected,
an increase in H + Cl2 led to an increase in the amount of
secondary HCl produced, requiring a lower value of k3b to match
the experimental data. The effect was most pronounced at low
Four other groups have attempted to measure the HCl yield
17
directly. Poulet et al. used a traditional discharge-flow system,
with electron-impact mass spectrometric detection. They mea-
sured HCl in pairs of experiments with either Cl atoms or O3
in excess. All of the HCl observed could be accounted for by
reactions such as Cl + HO2, and a branching fraction of (2 (
1
2)% was quoted. The determination suffered from variable
backgrounds of HCl from the Cl2 discharge, and uncertainties
in the amount of OH and Cl probably also played a role.
Lipson et al.18,22 made two determinations of the HCl
[
Cl2] where H atoms react mostly with O3. The simulations
-
11
3
indicated that if the rate coefficient is as high as 2.2 × 10
3
-1 -1
-12
cm molecule s , k3b could be reduced to 1.15 × 10
cm
production rate, the former using deuterated hydroxyl radicals,
OD (to discriminate against background levels of HCl) and the
latter using OH. The studies were conducted in a high-pressure
turbulent flow tube at pressures between 100 and 200 Torr,
which should minimize wall contact. A background of HCl was
present and the change in HCl was measured above this level.
The reaction was not driven to completion, and the rate of HCl
production was equated to the rate of OH loss. Absolute
calibrations were required for ClO, OH, and HCl using chemical
ionization mass spectrometry. It should be noted that the overall
rate coefficient measured by Lipson et al. is 30% lower than
molecule s-1. Increases in both OH + O3 and OH + Cl2
resulted in an increased loss rate for OH, and consequently less
occurrence of O + OH. The decreased amount of secondary
OH required an increase in k3b. The recommended rate coef-
ficient for OH + O3 has recently been revised upward from
-1
-14
-14
3
-1 -1 29
6
.8 × 10 to 7.8 × 10 cm molecule s . A 10% change
in either OH + O3 or OH + Cl2 led to 4-5% changes in k3b in
the same direction. The combined uncertainty in the HCl yield
related to the kinetics parameters is thus estimated as being
around 20%, similar to the random error in the determinations
and the uncertainty due to the excimer fluence. Thus, we
estimate an overall uncertainty of 35%, leading to a value of
the currently recommended value and up to 40% lower than
19-21
the most recent determinations.
It is not clear whether there
-
12
3
-1 -1
k3b ) (1.25 ( 0.45) × 10
cm molecule s . Expressed as
is a systematic calibration error that propagates into both the
overall rate coefficient and the branching fraction. Lipson et
a branching ratio, this becomes (6.5 ( 3.0)% relative to the
2
9
-11
3
recommended overall rate coefficient of 1.9 × 10
cm
-13
3
al. quote a rate coefficient k3b of (9.5 ( 1.6) × 10
cm
-
1
-1
molecule s , where an uncertainty of (25% in the overall
rate coefficient has been included.
-1 -1
molecule
s
at room temperature, which is similar to that
measured here. Expressed relative to their overall rate coef-
ficient, this corresponds to a branching fraction of 7%, or 5%
Other Experimental Tests. A number of other experiments
were conducted to confirm that the HCl was being produced
from the reactions 3 and 14. Some experiments were attempted
using a faster diode laser sweep rate (4 kHz) to improve the
time resolution. Production of HCl occurred on a time scale
consistent with that predicted by the model, but the data were
not of sufficient quality to obtain any kinetic information.
However, the HCl was clearly produced on a time scale of
29
relative to the NASA-JPL recommended value of k3; this
determination is also the basis for the current recommendation
for k3b.
Wang and Keyser23 used a low-pressure discharge flow tube
and measured HCl using a tunable diode laser. They also
measured the initial OH concentration and the ClO, using
resonance fluorescence and UV absorption, respectively. The
method did not require a knowledge of the overall rate
coefficient, since it involved the complete reaction of OH with
ClO, and a determination of the ratio of HCl produced to the
initial OH. However, secondary production of HCl occurred,
and the quoted branching fraction (9.0 ( 4.8)% was obtained
by extrapolating the HCl back to zero OH. Expressed relative
1
00s of µs, too long for it to be due to reaction of Cl atoms
with impurities, since Cl atoms are scavenged extremely rapidly
by O3.
Experiments were also conducted at 265 and 333 K (unfor-
tunately hampered by cracks in the cell window). The amount
of HCl produced at lower temperature, for otherwise identical
conditions, was more than at room temperature. This can be
rationalized in terms of the temperature dependences of the
major reactions. Reactions 16 and 17, which consume OH, both
have normal temperature dependences,29,36 while reaction 3 has
a negative temperature dependence, so the fraction of OH
reacting with ClO should be higher at 265 K. Again, the
dependence of the HCl on experimental conditions points toward
its being formed from reaction 3, but uncertainties in the cell
pressure caused by the leak preclude further analysis.
-
12
3
to the overall reaction, a rate coefficient k3b ) 1.7 × 10 cm
-
1 -1
molecule
s
is obtained, which is somewhat larger than both
22
the present measurement and that of Lipson et al. The very
recent study of Bedjanian et al., which used a low-pressure
discharge flow tube, gave the lowest yield of HCl from the
2
0
recent studies, 3.5% at 298 K.
Overall, the four most recent determinations clearly show the
22
occurrence of channel 3b. The studies of Lipson et al., Wang
2
3
20
and Keyser, and Bedjanian et al., which covered wider
temperature ranges than the present study, indicate that the HCl
yield is largely independent of both temperature and pressure
down to 220 K. There is also good evidence from theoretical
calculations of a low energy pathway to give HCl + O2 on a
Discussion
The results presented here clearly show that HCl is formed
in reaction 3 with a yield of 6-7% (based on the currently
recommended rate coefficient). The observation of small, but
2
2,47
singlet surface,
although a third theoretical study found the

OH + ClO f HCl + O2
J. Phys. Chem. A, Vol. 106, No. 8, 2002 1575
transition state to be too energetic to allow any reaction.48 The
agreement between the measurements of k3b is only about a
factor of 2 if each determination is taken as reported, although
the actual agreement is probably better. Each determination has
potential problems with calibration and background levels of
HCl, which were all addressed differently. The present technique
uses a different approach from the others, incorporating rapid
time response on a time scale shorter than that for diffusion to
the walls. The determination used an absolute calibration of HCl
based on known spectroscopic constants and also tied the
determination of excimer laser fluence to the HCl measurement.
However, the complexity of the chemistry, coupled with the
need to numerically simulate the results, leads to an additional
uncertainty. A further weakness of this study is that neither OH
or ClO was measured, but the calculation of their concentration
is tied very closely to the HCl calibration. An ideal measurement
might utilize a flowing discharge source of ClO with in situ
(12) Lary, D. J.; Chipperfield, M. P.; Toumi, R. J. Atmos. Chem. 1995,
2
1, 61.
13) Leu, M. T.; Lin, C. L. Geophys. Res. Lett. 1979, 6, 425.
14) Ravishankara, A. R.; Eisele, F. L.; Wine, P. H. J. Chem. Phys.
(
(
1983, 78, 1140.
(15) Hills, A. J.; Howard, C. J. J. Chem. Phys. 1984, 81, 4458.
(
16) Burrows, J. P.; Wallington, T. J.; Wayne, R. P. J. Chem. Soc.,
Faraday Trans. 2 1984, 80, 957.
17) Poulet, G.; Laverdet, G.; Le Bras, G. J. Phys. Chem. 1986, 90,
159.
(
(18) Lipson, J. B.; Elrod, M. J.; Beiderhase, T. W.; Molina, L. T.;
Molina, M. J. J. Chem. Soc., Faraday Trans. 1997, 93, 2665.
(19) Kegley-Owen, C. S.; Gilles, M. K.; Burkholder, J. B.; Ravishankara,
A. R. J. Phys. Chem. A 1999, 103, 5040.
(20) Bedjanian, Y.; Riffault, V.; Le Bras, G. Int. J. Chem. Kinet. 2001,
3, 587.
3
(
(
21) Wang, J. J.; Keyser, L. F. J. Phys. Chem. A 2001, 105, 10544.
22) Lipson, J. B.; Beiderhase, T. W.; Molina, L. T.; Molina, M. J.;
Olzmann, M. J. Phys. Chem. A 1999, 103, 6540.
(23) Wang, J. J.; Keyser, L. F. J. Phys. Chem. A 2001, 105, 6479.
(
24) Kegley-Owen, C. S.; Tyndall, G. S.; Orlando, J. J.; Fried, A. Int.
J. Chem. Kinet. 1999, 31, 766.
25) Stickel, R. E.; Nicovich, J. M.; Wang, S.; Zhao, Z.; Wine, P. H. J.
19
UV detection, and flash production of OH coupled with time-
resolved diode laser detection of HCl as used here.
(
Phys. Chem. 1992, 96, 9875.
In the atmosphere, the concentrations of OH and ClO are
controlled by reactions other than reaction 3. Thus, for
atmospheric models the absolute rate coefficient for reaction
(26) Rim, K. T.; Hershberger, J. F. J. Phys. Chem. A 1999, 103, 3721.
(
27) Cantrell, C. A.; Zimmer, A.; Tyndall, G. S. Geophys. Res. Lett.
997, 24, 2195.
28) CRC Handbook of Chemistry and Physics, 65th ed.; Weast, R. C.,
1
(
3b is required, rather than a branching ratio. The recent
Ed.; CRC Press: Boca Raton, FL, 1985.
determinations of k3b confirm that HCl production occurs at a
(29) Sander, S. P.; Friedl, R. R.; DeMore, W. B.; Golden, D. M.; Kurylo,
M. J.; Hampson, R. F.; Huie, R. E.; Moortgat, G. K.; Ravishankara, A. R.;
Kolb, C. E.; Molina, M. J. JPL Publication 00-003; NASA Jet Propulsion
Laboratory: Pasadena, CA, 2000.
(30) Rothman, L. S.; et al. J. Quant. Spectrosc. Radiat. Transfer 1998,
60, 665.
9-12
level of significance to stratospheric chemistry.
However,
there is still a factor of 2 discrepancy in the rate coefficients
for reaction 3b as reported by the individual groups. This may
represent real systematic differences between the studies, or may
simply reflect the difficulties inherent in measuring this
parameter.
(
0.
31) Pine, A. S.; Fried, A.; Elkins, J. W. J. Mol. Spectrosc. 1985, 109,
3
(
(
32) Pine, A. S.; Looney, J. P. J. Mol. Spectrosc. 1987, 122, 41.
33) Babrov, H.; Ameer, G.; Benesch, W. J. Chem. Phys. 1960, 33, 145.
Acknowledgment. This work was supported by NASA’s
Upper Atmosphere Research Program. C. S. Kegley-Owen held
a NSF Training Grant (EAR-9256339) during preliminary stages
of the work. The authors thank Michael Coffey and Dirk Richter
of NCAR for comments on the manuscript. NCAR is sponsored
by the National Science Foundation.
(34) Rank, D. H.; Eastman, D. P.; Rao, B. S.; Wiggins, T. A. J. Mol.
Spectrosc. 1963, 10, 34.
35) Braun, W.; Herron, J. T.; Kahaner, D. K. Int. J. Chem. Kinet. 1988,
0, 51.
36) 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.
37) Gilles, M. K.; Burkholder, J. B.; Ravishankara, A. R. Int. J. Chem.
Kinet. 1999, 31, 417.
(
2
(
(
(
(
38) Michael, J. V.; Lee, J. H. Chem. Phys. Lett. 1977, 51, 303.
39) Bemand, P. P.; Clyne, M. A. A. J. Chem. Soc., Faraday Trans. 2
References and Notes
1
977, 73, 394.
40) Wagner, H. G.; Welzbacher, U.; Zellner, R. Ber. Bunsen-Ges. Phys.
Chem. 1976, 80, 902.
41) Silvente, E.; Richter, R. C.; Hynes, A. J. J. Chem. Soc., Faraday
Trans. 1997, 93, 2821.
42) Smith, I. W. M.; Williams, M. D. J. Chem. Soc., Faraday Trans.
1985, 81, 1849.
43) Raiche, G. A.; Jeffries, J. B.; Rensberger, K. J.; Crosley, D. R. J.
Chem. Phys. 1990, 92, 7258.
(
1) Scientific Assessment of Ozone Depletion: 1998; World Meteoro-
logical Organization: Geneva, 1999.
2) Brasseur, G. P., Orlando, J. J., Tyndall, G. S., Eds. In Atmospheric
Chemistry and Global Change; Oxford University Press: New York, 1999.
3) Finlayson-Pitts, B. J.; Pitts, J. N., Jr. Chemistry of the Upper and
Lower Atmosphere; Academic Press: San Diego, 2000.
(
(
(
(
(
2
(4) Brasseur, G.; De Rudder, A.; Tricot, C. J. Atmos. Chem. 1985, 3,
(
2
61.
(
(
(
5) McElroy, M. B.; Salawitch, R. J. Science 1989, 243, 763.
6) Toumi, R.; Bekki, S. Geophys. Res. Lett. 1993, 20, 2447.
7) Chandra, S.; Jackman, C. H.; Douglass, A. R.; Fleming, E. L.;
(
(
44) Chalamala, B. R.; Copeland, R. A. J. Chem. Phys. 1993, 99, 5807.
45) Matsumi, Y.; Nomura, S.; Kawasaki, M.; Imamura, T. J. Phys.
Chem. A 1996, 100, 176.
(46) Burkholder, J. B.; Hammer, P. D.; Howard, C. J.; Goldman, A. J.
Geophys. Res. 1989, 94, 2225.
(47) Dubey, M. K.; McGrath, M. P.; Smith, G. P.; Rowland, F. S. J.
Phys. Chem. A 1998, 102, 3127.
(48) Sumathi, R.; Peyerimhoff, S. D. Phys. Chem. Chem. Phys. 1999,
1, 5429.
(49) Chen, H.-L.; Moore, C. B. J. Chem. Phys. 1971, 54, 4072.
Considine, D. B. Geophys. Res. Lett. 1993, 20, 351.
(
8) Natarajan, M.; Callis, L. B. J. Geophys. Res. 1991, 96, 9361.
9) Michelsen, H. A., et al. Geophys. Res. Lett. 1996, 23, 2361.
(
(
10) Chance, K.; Traub, W. A.; Johnson, D. G.; Jucks, K. W.; Ciarpallini,
P.; Stachnik, R. A.; Salawitch, R. J.; Michelsen, H. A. J. Geophys. Res.
996, 101, 9031.
11) Khosravi, R.; Brasseur, G. P.; Smith, A. K.; Rusch, D. W.; Waters,
J. W.; Russell, J. M., III. J. Geophys. Res. 1998, 103, 16203.
1
(