Rate Coefficient Measurements for the Reaction OH + ClO → Products

Article abstract of DOI:10.1021/jp9904320

The rate coefficient for the reaction OH + ClO → products (1) was measured under pseudo-first-order conditions in OH. A discharge flow system was used to produce ClO, and its concentration was measured by UV/visible absorption. OH was produced by pulsed laser photolysis of O₃ (or ClO) at 248 nm in the presence of H₂O and was monitored by laser-induced fluorescence. The value of k₁ between 234 and 356 K is given by k₁(T) = (8.9 ± 2.7) × 10^-12 exp[(295 ± 95)/T] cm³ molecule^-1 s^-1, where uncertainties are 95% confidence limits and include estimated systematic uncertainties. Our value is compared with those from previous investigations.

Full text of DOI:10.1021/jp9904320

5040
J. Phys. Chem. A 1999, 103, 5040-5048
Rate Coefficient Measurements for the Reaction OH + ClO f Products
†
,†
Carla S. Kegley-Owen, Mary K. Gilles,* James B. Burkholder, and A. R. Ravishankara*
Aeronomy Laboratory, National Oceanic and Atmospheric Administration, 325 Broadway,
Boulder, Colorado 80303, and CooperatiVe Institute for Research in EnVironmental Sciences,
UniVersity of Colorado, Boulder, Colorado 80309
ReceiVed: February 4, 1999; In Final Form: May 4, 1999
The rate coefficient for the reaction OH + ClO f products (1) was measured under pseudo-first-order
conditions in OH. A discharge flow system was used to produce ClO, and its concentration was measured by
UV/visible absorption. OH was produced by pulsed laser photolysis of O
of H O and was monitored by laser-induced fluorescence. The value of k
by k
limits and include estimated systematic uncertainties. Our value is compared with those from previous
investigations.
3
(or ClO) at 248 nm in the presence
2
1
between 234 and 356 K is given
-12
3
-1 -1
1
(T) ) (8.9 ( 2.7) × 10 exp[(295 ( 95)/T] cm molecule s , where uncertainties are 95% confidence
5
Introduction
The reaction of OH with ClO
Poulet et al. (at 298 K). Previous studies, all of which were
carried out under pseudo-first-order conditions in OH, used
discharge flow systems for the radical source (or sources)
followed by detection of OH via resonance fluorescence, laser-
induced fluorescence, or laser magnetic resonance. Several
experiments have used reaction 2 with an excess of Cl atoms
as the ClO source and assumed that the initial ozone concentra-
tion was equal to the initial ClO produced, [O3]0 ) [ClO]0.4
We will examine this assumption in this work. Other experi-
ments done in excess Cl atoms applied corrections4 to k1 to
account for the regeneration of OH through the reaction
OH + ClO f HO + Cl
(1a)
(1b)
2
f HCl + O2
,6,7
may play a significant role in the partitioning of chlorine in the
upper stratosphere. The conversion of ClO to Cl in reaction 1a
propagates ozone destruction via
,5,7
Cl + O f ClO +O2
(2)
3
Cl + HO f ClO + OH
(3a)
(3b)
2
while reactive chlorine (ClO and Cl) is converted to the reservoir
species HCl in reaction 1b. The branching ratios in reaction 1
are of particular importance, since they directly affect partition-
ing between active and reservoir, HCl and ClONO2, chlorine
f HCl + O2
A few studies have used excess O3, [O3] > [Cl]0, when
employing reaction 2 as the ClO radical source.
In this paper, we present results from the study of the
temperature dependence of the rate coefficient for reaction 1.
Reaction 1 was studied under pseudo-first-order conditions in
OH. ClO was generated via reaction 2 under conditions where
1
species in the upper stratosphere. Recently, Dubey et al.
3
,5,8
performed a box model sensitivity analysis using the currently
2
recommended value of k1 and found that a 7% branching ratio
for channel 1b reduces the modeled [ClO]/[HCl] ratio to that
observed in some field studies. Revisions in the value of k1 and
its temperature dependence would change the value of the
branching ratio required to bring field observations and model
calculations into agreement. Revisions in the overall value of
k1 would also have other consequences to chlorine chemistry.
Previous measurements of k1 contain discrepancies in both
its magnitude and temperature dependence. Determinations of
(
a) excess O3 was present or (b) where O3 was completely
depleted during ClO production. In these experiments ClO and
O3 were monitored simultaneously via UV/visible absorption
in situ.
Experiments
-
11
-11
k1 at room temperature range from 0.91 × 10 to 1.99 × 10
3
-1 -1 2
cm molecule s . The value of k1 currently recommended
Owing to the relatively slow self-reaction of ClO and the
high sensitivity with which OH can be detected by laser-induced
fluorescence, we studied this reaction under conditions of [ClO]
> 10[OH]0 (pseudo-first-order in OH). The apparatus used was
a combination of a discharge flow system for producing ClO,
a UV/visible absorption spectrometer to quantify ClO concen-
tration, and a pulsed laser photolysis and pulsed laser-induced
fluorescence system for producing and detecting OH. The
apparatus and methods of operation are described in detail in a
-
11
for atmospheric models is k1(T) ) 1.1 × 10 exp[(120 ( 150)/
3
-1 -1
-11
3
T] cm molecule cm
s
, with k1(298 K) ) 1.7 × 10
-
1
-1 2
molecule s . This recommendation is based on the data of
Hills and Howard3 (who observed a negative temperature
4
dependence (E/R ) 235 ( 46 K)), of Burrows et al. (who
reported no dependence of k1 on temperature), and the value of
*
To whom correspondence should be addressed. Address: NOAA/ERL,
R/E/AL2, 325 Broadway, Boulder, CO 80303. E-mail: Ravi@al.noaa.gov
or mgilles@al.noaa.gov.
9
recent publication.
†
ClO Production. Cl atoms were generated in a side arm of
a flow tube (2.54 cm i.d.) by passing a dilute mixture of Cl2 in
Also affiliated with the Department of Chemistry and Biochemistry,
University of Colorado, Boulder, CO 80309.
1
0.1021/jp9904320 CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/15/1999

OH + ClO f Products
J. Phys. Chem. A, Vol. 103, No. 26, 1999 5041
kinetic experiments. The concentration of ClO was calculated
using the Beer-Lambert law:
A ) σl[ClO]
(I)
where A was the absorbance, σ was the absorption cross section
-
18
2
-1 2
(
σ253.7nm ) 4.25 × 10
cm molecule ), and l was the path
length (28.0-38.8 cm). After each OH temporal profile was
recorded, the photolysis laser was blocked and the beam from
the deuterium lamp passed through the cell to measure the
absorbance due to ClO and O3. The absorbance due to the initial
O3 could be measured by turning off the microwave discharge.
Finally, the flows of O3 and Cl2 were shut off and I0, the
intensity in the absence of absorbers, was measured. These
spectra were used to determine the initial ozone concentration
(
[O3]0) prior to ClO formation, the concentration of ClO, and
the concentration of ozone remaining after ClO production,
[O3]excess”. The absorbance due to ClO was extracted from the
“
measured spectrum by subtracting a scaled ClO reference
spectrum until all of the ClO structured absorption was
completely removed. Since the ClO spectra is structured, this
is easily accomplished. The spectral subtraction was normally
done by eye immediately after taking the spectra to check for
any variation in the deuterium lamp intensity. An example of
the spectral subtraction is shown in the lower part of Figure 1.
The residual shown in the middle and lower part (×1000) of
Figure 1 was obtained by subtracting the remaining O3. The
abundance of ClO (as inferred from its known absorption cross
section at 253.7 nm) subtracted from the reference spectra to
obtain an unstructured residual was used to determine [ClO].
3 2
Figure 1. (a) Reference spectra of O , ClO, and Cl . (b) Example of
14
-3
the spectral subtraction used to obtain [ClO]. (c) Residual (×1000)
obtained after subtracting the scaled ClO and O reference spectra from
the measured spectrum that contained both ClO and O
The concentrations of Cl2 used, (0.2-3) × 10 molecule cm ,
were too small to be measured by UV absorption and were
calculated from calibrated flows and the total pressure.
3
3
.
Some experiments were carried out under conditions in which
ozone was consumed during the ClO radical production. These
measurements were performed to test the dependence of k1 on
the presence of excess ozone and the assumption of stoichio-
metric conversion of O3 to ClO (reaction 2). Under these
conditions, there was no O3 remaining in the reaction cell. The
concentration of remaining Cl atoms was not measured. We
could ensure that all O3 was depleted by comparing the ratio of
the measured ClO absorbance under these conditions at 260.0
and 253.7 nm. The absorbance at these two wavelengths for
ClO was determined from ClO spectra measured in a large
excess of Cl by adding Cl2O or O3 to be 1.20 ( 0.01. Because
He through a microwave discharge. Within the flow tube, Cl
atoms reacted with O3 added through the movable injector via
-
11
3
-1
reaction 2, where k2(298 K) ) 1.2 × 10
cm molecule
-1 2
s . Typical flow rates of He in the flow tube were 16-38
3
-1
STP cm s at pressures of 4.2-15.4 Torr. This resulted in
linear gas flow velocities in the flow tube and reaction cell (also
-
1
2
.54 cm i.d.) of 330-800 cm s . ClO concentrations ranged
13
13
-3
from 0.3 × 10 to 12 × 10 molecule cm .
UV Absorption Spectroscopy. The absorbance due to ClO
was measured in situ by UV/visible absorption spectroscopy.
The output of a 30 W D2 lamp was collimated and passed
through the cell along the axis of the flow (counter to the
direction of the photolysis laser beam and perpendicular to the
probe laser beam). The D2 beam was then focused onto the
entrance slit of a 0.35 m spectrometer. The spectrometer
employed a 600 grooves/mm grating blazed at 300 nm and a
cooled 1024 element diode array detector. For monitoring ClO
and O3, the spectrometer was set to observe from 250 to 365
nm at a spectral resolution of 0.8 nm (fwhm). The wavelength
was calibrated using the emission lines from a low-pressure Hg
lamp. The ClO concentration was quantified using the unstruc-
tured portion of its spectrum, since the cross section in this
region is well-known and is independent of temperature and
2
the O3 cross section at 253.7 nm is ∼2.4 times larger than that
of ClO, the ratio would deviate from 1.20 if O3 were present.
Because the measured ratio of absorbance was essentially 1.20,
11
-3
twh ee pc el al lc .e an upper limit of 7 × 10 molecule cm of O3 in
OH Production. In most experiments, OH was produced via
photolysis of O3 in the presence of H2O that was added
downstream of the flow tube:
1
O + hν f O( D) + O
(4)
(5)
3
2
1
2
O( D) + H O f 2OH
resolution. However, the spectral subtraction, discussed later,
2
utilized the structured region of the spectrum.
Reference spectra of O3, ClO, and Cl2 are shown in Figure 1
and are in good agreement with literature values. A ClO
reference spectrum was measured prior to each experiment by
titrating O3 (or in some cases Cl2O) with a large excess of Cl
atoms. These spectra were recorded under flow, pressure, and
temperature conditions that were identical to the subsequent
2
1
The rate coefficient for the reaction of O( D) with H2O is k5
) 2.2 × 10
2
-10
3
-1 -1
cm molecule s , and this reaction went to
>95% completion in <3 µs. Vibrationally excited OH produced
in reaction 5 was quenched rapidly by H2O. For example, the
1
0
quenching rate coefficient for OH (v′′ ) 1) by H2O is 1 ×
-
11
3
-1 -1
10 cm molecule and higher levels are quenched even
s

5042 J. Phys. Chem. A, Vol. 103, No. 26, 1999
Kegley-Owen et al.
1
1
-3
faster. With insufficient [H2O], the vibrational quenching of OH
could be observed as a rise in the OH fluorescence signal.
Although the exact concentration of H2O was not needed for
the kinetics studies, it was calculated from flows, vapor
pressures, and total pressures. H2O concentrations in the reaction
[OH]0 (<3 × 10 molecule cm ) was produced by pulsed laser
photolysis of H2O2 at 248 nm. The concentration of CH4 {(1-
1
6
-3
13) × 10 molecule cm } was calculated from flow rates and
pressure. The value obtained for k9(298 K) ) (6.77 ( 0.26) ×
-
15
3
-1 -1
10
cm molecule
s
(uncertainty is 2σ precision) agrees
15
-3
2
cell were (3-20) × 10 molecule cm . Typical [OH]0 were
with previous determinations.
1
1
-3
(
1-3) × 10 molecule cm from this OH source. However,
Materials. The Cl2 (1% Cl2 (>99.99%) in He (>99.997%)),
N2 (>99.9995%), and He (>99.997%) were all obtained
commercially. Helium was passed through a liquid nitrogen trap,
and all other gases were used as supplied. Ozone was produced
in a commercial ozonizer and stored on a silica gel trap kept in
a dry ice/ethanol bath. Cl2O was synthesized following tech-
1
1
in ∼10% of the temporal profiles [OH]0 was ∼ 5 × 10
-
3
molecule cm .
An alternative OH source was the 193 nm photolysis of
HNO3:
HNO + hν f OH + NO2
(6)
niques in the literature, and its purity (>92%) was determined
14
3
via UV absorption. Mass flow controllers were used to measure
the gas flows into the cell, and pressure was measured with
capacitance manometers. HNO3 was introduced into the reaction
cell by bubbling He through a mixture of concentrated sulfuric
and nitric acids.
The quantum yield for OH is 0.3 at this wavelength, and the
1
3
other products are oxygen atoms [O( D) + O( P) ) 0.8] and
11
HONO along with very small yields of hydrogen atoms. Initial
1
1
OH concentrations were calculated to be roughly 6 × 10
-
3
molecule cm from the measured photolysis laser fluence, the
absorption cross section, and the estimated HNO3 concentration.
For this calculation, we assumed that O( D) reacted predomi-
Results
1
OH temporal profiles in the presence of excess ClO and O3
were influenced by the reactions
nantly with HNO3 to produce OH, since this would provide an
3
upper limit for [OH]0. O( P) would have reacted with ClO,
producing Cl atoms that would recycle ClO via reaction with
OH + ClO f products
OH f loss
(1)
-
12
O3, and HONO would be lost slowly (k(298 K) ) 4.5 × 10
3
-1 -1 2
cm molecule s ) through reaction with OH radicals. Hence,
none of the secondary chemistry should have produced OH on
a time scale comparable to that for reaction 1. The HNO3
(10)
where reaction 10 represents the first-order rate coefficient for
loss of OH due to diffusion and flow out of the detection region.
Other reactions that could influence the OH temporal profile
were
15
-3
concentration (∼1 × 10 molecule cm ) was estimated from
the first-order rate constant for OH loss in the absence of ClO
due to the reaction
OH + HNO f products
(7)
3
OH + Cl f HOCl + Cl
(11)
(12)
2
which has a rate coefficient of ∼1 × 10 cm molecule s-1
-
13
3
-1
OH + O f HO + O
2
3
2
2
at 298 K and ∼8 Torr.
In experiments where all of the ozone was depleted during
ClO production, OH was produced via the 248 nm photolysis
of ClO,
Contributions to the OH loss rate coefficients from reactions
1 and 12 were estimated from the calculated [Cl2] and
measured [O3]excess along with the respective rate coefficients.
1
2,15
1
These losses were estimated to be <1% and <2%, respectively,
ClO + hν f O( D) + Cl
(8)
of the loss due to reaction 1 and were neglected in the data
1
1
-3
1
12
analysis. Normally, [OH] was <5 × 10 molecule cm and
followed by O( D) reaction with H2O. The quantum yield for
0
1
the loss due to OH self-reaction
O( D) production at this wavelength is ∼1. Photolysis fluences
-
1
2
were adjusted [(1.5-4.8) mJ pulse cm ] when [ClO] was
varied to maintain initial OH concentrations of (2.9-5.3) ×
OH + OH f products
(13)
11
-3
13
14
1
0
molecule cm . ClO concentrations were 10 -10
-
3
was small. All OH temporal profiles were single exponential
decays (see Figure 2) and were fit to
molecule cm . We calculated that <2% of ClO was destroyed
by photolysis. Hence, photolysis did not significantly affect
[
ClO].
ln[OH] - ln[OH] ) -k ′t
(III)
t
0
1
OH temporal profiles were monitored by laser-induced
fluorescence. OH was excited by the frequency-doubled output
from a pulsed Nd:YAG pumped dye laser. Fluorescence from
where [OH]t and [OH]0 represent the OH fluorescence signals
at times t and zero, respectively, and k1′ ) k1[ClO] + kd. kd
represents the first-order rate coefficient for reaction 10,
reactions with impurities, and reactions 11, 12, and 13. Plots
of ln[OH]t vs time at 298 K are shown in Figure 2. Values of
k1′ were obtained from the slope of such plots determined at
various [ClO]. The second-order rate coefficient, k1, was
obtained from the slopes of plots of k1′ vs [ClO] (Figure 3).
The intercepts, kd, from these plots were always within the
uncertainty of the OH loss rate coefficients measured in the
absence of ClO. Details of the experiments carried out at 298
K in the presence of excess O3 are given in Table 1.
2
+
2
2 +
the OH (A Σ , V′ ) 1) f OH (X Π, v′′ ) 1) and OH (A Σ ,
2
V′ ) 0) f OH (X Π, V′′ ) 0) transitions passed through a
band-pass filter (308 ( 10 nm) and was detected by a
photomultiplier tube.13 Temporal profiles were obtained by
varying the delay time between the photolysis and probe lasers
from 50 µs to 12 ms.
Because this was the first measurement of an OH reaction
rate coefficient on this particular apparatus, we measured the
well-known rate coefficient for the reaction
OH + CH f H O + CH
(9)
4
2
3
In some experiments at 298 K done in excess O3, OH was
produced by HNO3 photolysis at 193 nm (reaction 6). For these
k9 was measured under pseudo-first-order conditions in OH.

OH + ClO f Products
J. Phys. Chem. A, Vol. 103, No. 26, 1999 5043
a
TABLE 1: ClO + OH Experiments at 298 K in the Presence of Excess O
3
OH source
[ClO]
[O
3
]
excess
[ClO]/[OH]
0
pressure
V
k
1
1
O( D) + H
2
O
1.5-8.0
0.24-1.1
0.25-1.4
0.25-1.4
0.8-7.2
0.8-6.5
0.4-0.8
0.8-20
0.08-0.5
0.09-0.5
1.2-2.3
0.1-0.5
0.1-0.7
60-240
60-180
50-240
70-150
70-180
40-280
70-170
130-760
90-600
50-80
4.2
7.9
7.9
8.4
8.8
8.8
15.4
4.6
8.2
7.7
4.8
4.7
715
760
760
330
680
680
360
790
440
450
770
760
2.56 ( 0.18
2.38 ( 0.14
2.43 ( 0.12
2.49 ( 0.08
2.49 ( 0.16
2.52 ( 0.18
2.40 ( 0.31
2.40 ( 0.10
2.41 ( 0.04
2.47 ( 0.14
2.31 ( 0.18
2.30 (0.10
2.43 ( 0.15
2.52 ( 0.38
2.42 ( 0.12
2.42 ( 0.24
2.45 ( 0.09
1.4-7.0
1.4-7.0
0.6-8.0
0.6-8.0
1.9-11
2.5-7.2
3.0-9.0
2.0-9.0
2.0-8.0
2.0-8.3
1.5-8.8
50-150
40-150
average
average
HNO
3
+ hν
1.1-8.7
0.5-1.1
0.6-1.2
0.6-2.4
15-100
10-95
20-120
7.7
6.5
8.1
350
350
350
1
1
.5-7.8
.9-9.4
a
ClO] is in units of 10 molecule cm-3, [O
13
excess is in units of 1014 molecule cm-3, pressure is in Torr, V is in cm s-1, and k
is in units of 10-11
1
[
3
]
3
-1 -1
cm molecule
s
with the uncertainty of (2σ precision from the fit.
ClO. As seen in Table 1, the rate coefficients determined by
producing OH via HNO3 photolysis in the presence of excess
O3 are in agreement with those obtained using reactions 4 and
5
for OH production.
k1 was measured at 10 temperatures between 234 and 356
K. At temperatures lower than 234 K, H2O condensed on the
reaction cell surfaces. Experimental conditions and results are
summarized in Table 2 and shown in Figure 4. A linear least-
squares fit of ln k1 vs 1/T yielded the Arrhenius expression
-
12
3
-1
k1(T) ) (8.9 ( 2.2) × 10 exp[(295 ( 65)/T] cm molecule
-1
s . The uncertainty in the preexponential factor, A, is 2σA where
σA ) AσlnA and σlnA is the precision in ln A from the fit of ln k1
vs 1/T. For comparison, the results from previous studies
performed in excess O3 are also shown in Figure 4. The greatest
source of possible systematic error in this study involves the
determination of the ClO radical concentration. Sources of
systematic uncertainties include the absorption cross section of
ClO ((5%) and determination of the absorption path length
Figure 2. Plots of ln[OH]
conditions of excess O . Values of k
such plots at various ClO concentrations.
t
vs time at 298 K for different ClO under
3
1
′ were obtained from the slope of
(
(3%). Other uncertainties (2σ) arise due to the measurement
of temperature ((2 K, which results in an uncertainty of (1%
in [ClO]), loss of OH via reaction with Cl2 and O3 ((2%), and
the ClO concentration gradient ((5%). The concentration
gradient is discussed in detail in prior publications).9 The
uncertainty in the spectral subtraction was calculated from 2σ
of the mean of the residual absorption (Figure 1c) over the entire
wavelength range. This uncertainty was attributed to ClO
absorption and used to calculate the percent difference ((6%)
in the resulting [ClO]. This is larger than reported in earlier
studies because the spectrum of O3 contains some structure while
,16
9
that of Cl2O does not. Adding these to the systematic
uncertainties gives an overall uncertainty of 22% in [ClO].
Combining this in quadrature with the 2σ precision yields k1 )
-
12
3
-1 -1
(
8.9 ( 2.7) × 10
exp[(295 ( 95/T] cm molecule s .
-1
-2
Varying photolysis laser fluence (0.02-0.5 mJ pulse cm )
and probe laser fluence (by a factor of 2), increasing [H2O]
1
5
-3
{
(3-20) × 10 molecule cm }, and increasing [O3]excess (a
Figure 3. Plot of k
1
′ vs [ClO], which yields a value for k
1
(298 K).
factor of 5) did not affect the measured value for k1(298 K). In
addition, doubling the cell pressure while maintaining a constant
flow velocity, and varying the pressure (4.2-15.4 Torr) and
These experiments were done under conditions of O
3
(excess). [O ]
3 0
1
3
-3
varied from (3-40) × 10 molecule cm . Measurements were made
with different linear gas flow velocities, pressure, and laser fluence as
noted in the figure.
-
1
flow velocity (330-790 cm s ) did not affect k1. Changing
the OH source to HNO3 photolysis and doubling [HNO3] also
had no influence on k1. Earlier tests on the path length used to
determine [ClO] from the absorption measurement are described
experiments the measured kd also included a loss due to reaction
6
, which was used to estimate [HNO3]. The intercepts in the
9
plots of k1′ vs [ClO] were always within the uncertainty of the
elsewhere, and additional tests done in the current study are
measured OH loss rate coefficient obtained in the absence of
presented in Discussion.

5044 J. Phys. Chem. A, Vol. 103, No. 26, 1999
Kegley-Owen et al.
TABLE 2: Determination of k in the Presence of Excess O
1
3
as a Function of Temperaturea
photolysis
fluence
T (K)
[ClO]
[O
3
]
excess
[ClO]/[OH]
0
pressure
V
k
1
234
236
236
245
252
252
263
272
298
317
336
356
1.3-12
2.0-9.0
1.8-5.7
1.6-8.5
0.3-8.0
1.1-12
2.7-8.1
1.2-6.7
0.2-1.4
2.0-2.7
2.0-3.5
0.9-4.9
0.3-1.0
0.5-1.4
1.8-3.0
0.8-2.5
60-150
20-70
7.3
7.7
7.7
7.3
7.3
7.8
7.5
6.5
460
330
500
480
530
330
440
410
0.09
0.16
0.06-0.13
0.20
0.08
0.3
0.21
0.08-0.23
3.07 ( 0.16
3.24 ( 0.14
2.97 ( 0.08
3.24 ( 0.12
2.84 ( 0.16
2.74 ( 0.14
2.72 ( 0.30
2.77 ( 0.20
2.44 ( 0.07
2.22 ( 0.12
1.98 ( 0.41
1.99 ( 0.12
60-150
30-120
80-480
10-100
20-60
b
15-420
1.5-8.0
0.3-7.3
1.0-8.0
0.6-2.0
0.8-2.4
0.5-1.7
30-200
20-800
30-100
7.5
7.3
7.7
440
450
430
0.10-0.17
0.19
0.25
a
ClO] is in units of 10 molecule cm-3, [O
13
excess is in units of 10 molecule cm , pressure is in Torr, linear velocity V is in cm s-1, photolysis
14
-3
[
3
]
-
1
2
-11
3
-1 -1
s
. b The OH source was nitric acid photolysis.
fluence is in mJ pulse cm , and k
1
( 2σ is in units of 10 cm molecule
TABLE 3: Determinations of k
1
in Experiments Where O
3
Was Depleteda
photolysis
T (K) [ClO] [ClO]/[OH]
0
pressure
V
fluence
1
k (T)
2
2
2
2
2
2
2
3
50 3.0-10
98 1.6-8.5
98 2.0-8.0
98 1.5-6.0
98 0.6-8.0
98 3.0-8.0
98 4.0-7.4
90-105
80-95
50-60
40
50
60
7.8
7.4
7.6
4.8
8.8
4.4
4.4
7.7
430 2.0
460 2.2
460 1.6-3.5 2.22 ( 0.10
800 4.8
420 3.5
780 3.0
780 3.0
460 1.5
2.85 ( 0.08
2.22 ( 0.08
2.33 ( 0.08
2.34 ( 0.16
2.31 ( 0.16
2.24 ( 0.08
1.58 ( 0.16
55
47 1.5-7.7 125
a
ClO] is in units of 10 molecule cm-3, [O
13
[
3
]
excess is in units of
-3
-1
1
014 molecule cm , pressure is in Torr, linear velocity V is in cm s
,
-
1
2
photolysis fluence is in mJ pulse cm , and k
1
(T) ( 2σ is in units of
-
11
3
-1 -1
1
0
cm molecule
s .
Figure 4. Plots of k
conditions of excess O
1
(T) versus 1000/T for experiments done under
: this experiment [data (solid squares), fit (solid
3
3
line)]; Hills and Howard [data (open circles), fit (dashed line)]; Poulet
5
8
et al. [data (solid circles)]; Lipson et al. [data (solid diamonds), fit
solid line)]. Uncertainties are 2σ precision from the respective papers.
(
Discussion
There are two categories of potential sources of error in
determining k1 under pseudo-first-order conditions in OH. Side
reactions can either increase or decrease the loss rate of OH,
resulting in an erroneous value for k1. In addition, any
uncertainty in [ClO] affects the measured rate coefficient. These
issues, both in the present and previous experiments, are
discussed in this section.
When ClO is made in an excess of Cl, as done in numerous
4-7
previous experiments, OH is regenerated via reaction 3a. The
branching ratio for OH production in reaction 3 ranges from
Figure 5. Plot showing the fit to k
1
(T) versus 1000/T obtained under
conditions of excess O (solid line). The filled triangles are data points
3
1
7-20
2
0
.01 to 0.28.
The current recommendation yields a
taken when all ozone was depleted during the production of ClO.
17
branching ratio of 0.22 for channel 3a. Several previous studies
carried out with excess Cl atoms used this value to correct their
measured k1 for OH regeneration.4 These corrections to k1
,5,7
using UV/visible absorption. The average of six different k1′
-
11
3
are significant and ranged from 13 to 30% between 248 and
vs [ClO] plots yielded k1(298 K) ) (2.28 ( 0.11) × 10 cm
7
-1 -1
3
1
35 K in the experiments of Ravishankara et al. and from 6 to
molecule
s
(2σ precision only). These experiments were
4
3% between 243 and 298 K in those of Burrows et al. and
repeated at 250 and 347 K. Details of the experimental
conditions are given in Table 3, and results are shown in Figure
5 along with the Arrhenius expression obtained in excess O3.
These rate coefficients were all slightly lower than those
measured in the presence of excess O3. Since the rate coefficient
for reaction 3a, which regenerates OH, increases with increasing
temperature, we expected the largest difference at the highest
temperatures. At 347 K, k3a is about 25% larger than at 298 K.
5
were ∼28% at 298 K in the study of Poulet et al. However,
6
during the experiments of Leu and Lin it was not known that
reaction 3 could regenerate OH. Hence, they did not correct
their data for reaction 3a and their value for k1(298 K), 0.91×
-
11
3
-1 -1
1
0
cm molecule s , is lower than other determinations.
In a number of experiments at 298 K we intentionally depleted
O3 during ClO production and verified that O3 was depleted

OH + ClO f Products
J. Phys. Chem. A, Vol. 103, No. 26, 1999 5045
ClO + HO f HOCl + O (16)
TABLE 4: Facsimile Model Used to Test for Contribution
of Reaction 15 to OH Lossa
2
2
reaction
Ab
E/R
However, the measurements at 324 and 345 K had the largest
[OH]0 (and larger uncertainties in [OH]0). At the highest
calculated [OH]0, the k1 measured could be up to 10% larger
than the true value because of enhanced OH loss arising from
reaction 15. Deleting the data from these two temperatures
(along with any other individual measurements of k1′, which
could have had >3% contribution due to reaction 15) from the
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
12
12
12
11
13
12
12
12
12
14
11
11
13
11
11
12
11
11
13
OH + ClO f Cl + HO
2
8.9 × 10
3.8 × 10
1.6 × 10
4.8 × 10
4.8 × 10
4.2 × 10
2.9 × 10
2.6 × 10
3.0 × 10
1.1 × 10
4.1 × 10
1.8 × 10
2.3 × 10
2.9 × 10
1.1 × 10
2.5 × 10
3.0 × 10
1.0 × 10
1.7 × 10
-295
1228
940
-250
-700
240
160
350
500
500
OH + Cl
OH + O
OH + HO
HO
OH + OH f H
OH + H f H
OH + HCl f H
OH + HOCl f H
2
f HOCl + Cl
3
f HO
f H
2
+ O
O + O
2
2
2
2
2
+ ClO f HOCl + O
O + O
O + HO
O + Cl
O + ClO
f OH + 2O
+ Cl f ClO + OH
+ Cl f HCl + O
+ HO f H + O
f ClO + O
f HCl + HO
2
2
2
O
2
2
2
-12
2
Arrhenius plot yielded k1(T) ) (8.0 ( 2.6) × 10
exp [(327
2
3
-1 -1
(
84)/T] cm molecule (2σ precision). This value for k1
s
HO
HO
HO
HO
2
2
2
2
+ O
3
2
is not significantly different from the value for k1(T) obtained
by including these data. Hence, we believe that interference from
reaction 15 does not significantly influence our measured value
for k1.
450
2
-170
-600
260
980
130
-70
3300
0
2
2
O
2
2
Cl + O
Cl + H
3
2
2
O
2
2
1
4
-3
ClO concentrations were e1.2 × 10 molecule cm . Yet,
Cl + HOCl f products
some ClO was lost due to self-reaction:
O + ClO f Cl + O
O + HCl f OH + Cl
O + HOCl f OH + ClO
2
ClO + ClO f ClOO + Cl
f Cl + O
(17a)
(17b)
a
The first five reactions were used to simulate every OH temporal
2
2
profile. The remaining reactions were used to test for their influence
on the OH loss rate. b These are from the Arrhenius expression k(T) )
f OClO + Cl
(17c)
(17d)
3
-1 -1
A exp[(-E/R)(1/T)]. A is in units cm molecule
s , and E/R is in K.
M
9
8 Cl O
2
2
In Figure 5 the measured value in the absence of O3 at the
highest temperature is significantly lower than the Arrhenius
expression obtained with excess O3. These experiments were
carried out to test if we could observe a different value of k1
measured in an absence of excess O3. Since we did not measure
where k (298 K) ) 2.2 × 10 cm molecule s-1 at 8 Torr.
-
14
3
-1
1
7
Under our experimental conditions (8 Torr, flow velocity of
20 cm s , residence time of 80 ms in the absorption/reaction
-1
4
cell) the total decrease in [ClO] across the cell due to
self-reaction was calculated to be <12% under conditions of
excess O3 and <30% when O3 was completely consumed during
the formation of ClO. The difference in the ClO loss arises
because channels 17a and 17c produce Cl atoms that recycle
rapidly to ClO in the presence of O3. However, the concentration
gradient across the length of the absorption cell has little effect
on the measured [ClO], since absorption measures a column
abundance that, when divided by the path length, yields an
[Cl] in these experiments, we cannot correct the measured value
of k1. However, it does show that OH can be regenerated to
yield a lower value of k1 in the presence of Cl atoms.
When k1 is determined under conditions of excess ozone, OH
could be regenerated via
HO + O f OH + 2O
2
(14)
2
3
-
15
3
-1 -1 2
where k14(298 K) ) 2.0 × 10
experiments, [O3]excess ranged from 8.0 × 10 to 21 × 10
molecule cm so that reaction 14 should not significantly
cm molecule s . In our
21
average concentration. FACSIMILE simulations and calcula-
1
2
14
9,16
tions described previously
show that [ClO] in the reactive
-
3
volume (intersection of probe and photolysis lasers) located at
the midpoint of the absorption cell is the same (within 2% under
conditions of excess O3 and within 11% in the absence of excess
O3) as the integrated column average measured by absorption.
Previous tests determining a rate coefficient while measuring
3
,5,8
influence our measured k1. In previous measurements
of k1
carried out in excess ozone, [O3] was sufficiently low that
reaction 14 should be negligible.
A secondary reaction that could result in an erroneously large
value for k1 is
[
ClO] both prior and after the reaction cell showed that the
difference in the determined rate coefficients was consistent with
ClO loss via self-reaction alone.
OH + HO f H O + O
(15)
2
2
2
9
The rate coefficient for reaction 17d, the association reaction
to form Cl2O2, is greater at lower temperatures. FACSIMILE
simulations using reactions 2 and 17 under our experimental
conditions showed that at the lower temperatures (234-272 K)
-
10
3
-1 -1 2
where k15(298 K) ) 1.1 × 10 cm molecule s . Our OH
21
temporal profiles were simulated using FACSIMILE with the
reactions shown in Table 4. Only the first five reactions in Table
significantly influenced OH temporal profiles under our
experimental conditions. Every OH temporal profile was
simulated using these five reactions, and [ClO]0, [O3]excess, [Cl2]0,
and [OH]0 were set to match the experimental conditions. To
provide a conservative (i.e., largest) estimate of the error and
to account for uncertainties in these rate coefficients and in
4
1
4
-3
and highest ClO concentrations (∼1 × 10 molecule cm ),
[Cl2O2]/[ClO] ranged from 0.03 to 0.05. Higher [Cl2O2] would
have been observable in the absorption spectra; we did not
observe any absorption attributable to Cl2O2. These [Cl2O2] are
high enough to possibly affect the OH decay. There are no
reported rate coefficients for the reaction
[
OH]0, any increase in the OH loss rate due to reaction 15 was
doubled. In the majority of the OH profiles loss due to reaction
OH + Cl O f products
(18)
1
1
5 was insignificant (<2% of the OH loss rate due to reaction
). Although k1 is about 5 times less than k15, the maximum
2
2
-
11
cm molecule s-1 would result
3
-1
possible HO2 concentration was determined by [OH]0, and
ClO]/[OH]0 was always >10, (usually >50). Therefore, HO2
reacted predominantly with ClO
A value of k18 of 5 × 10
[
in an overestimation of our measured k1 by 10%. However, since
all of the k1′ vs [ClO] plots were linear and the [Cl2O2] to [ClO]

5
046 J. Phys. Chem. A, Vol. 103, No. 26, 1999
Kegley-Owen et al.
assume
TABLE 5: Summary of Measurements of k a
1
ClO made
w/excess Cl?
corrected for OH
regeneration?
k
1
(298 K)
A
E/R
[ClO]
0
3 0
) [O ] ?
Leu and Lin6
Ravishankara et al.
Burrows et al.
0.91 ( 0.26
1.17 ( 0.33
1.19 ( 0.09
1.75 ( 0.31
1.77 ( 0.33
yes
yes
yes
no
yes
yes
no
no
yes
yes
not needed
yes
yes
not needed
not needed
yes
yes
yes
calibration
no
7
-(66 ( 200)b
-(235 ( 46)
(9.2 ( 6.5)
c
(8.0 ( 1.4)
4
3
Hills and Howard
5
Poulet et al.
1
.99 ( 0.25
.89 ( 0.21
no
1
calibration
Lipson et al.8
this work
1.46 ( 0.23
2.28 ( 0.55
(5.5 ( 1.6)
(8.9 ( 2.7)
-(292 ( 72)
-(295 ( 95)
no
no
no
no
d
depleted O
no
3
no
2.44 ( 0.63
not needed
a
is in units of 10-11 cm molecule s-1, A is in 10-12 cm molecule s-1, and E/R is in K. Uncertainties given are 2σ precision as quoted
3
-1
3
-1
k
1
b
from the authors, except those of Hills and Howard and ours, which include 2σ precision and estimated systematic uncertainties. Ravishankara et
7
c
al. measured a negligible temperature dependence and preferred to quote a temperature-independent value. Burrows et al. measured k
1
over the
as temperature-independent. d We were unable to correct for OH regeneration, since we did not
1
temperature range 243-298 K but reported k
measure [Cl].
ratio changes dramatically with [ClO], we believe that the rate
coefficient for reaction of Cl2O2 with OH is not very large and
that our measured rate coefficient is not affected by reaction
determining [ClO] are reflected in k1. Some earlier measure-
ments of k1 generated ClO in an excess of Cl atoms via
Cl + O f ClO + O
2
(2)
1
8.
3
Additional tests on the absorption cell path length were also
or
performed. Since the gases flowing through the flow tube and
into the absorption/reaction cell may not continuously replenish
the extreme ends of the cell (i.e., a stagnant region), we
measured an effective absorption cell length l. Briefly, a second
cell of a fixed path length (50 cm) with no room for stagnation
was added in series with the reaction cell. The absorbance of
O3 across both of these cells was measured simultaneously. The
Cl + Cl
2
O f ClO + Cl2
(19)
[
O3] or [Cl2O] were measured prior to reaction with Cl, and it
4,6,7
was assumed that [ClO]0 ) [O3]0 or [ClO]0 ) [Cl2O]0.
All
of these experiments reported significantly smaller values for
k1 than reported here or by Hills and Howard or Poulet et al.;
3
5
[
O3] calculated from the 50 cm long absorption cell was used
we believe an overestimation of [ClO] could be the source of
this discrepancy. Reaction 2 is very exothermic and can produce
vibrationally excited (ClO ). Burkholder et al. observed that
as much as 40% of ClO could be lost in excess Cl and suggested
that the reaction
to obtain l of the reaction cell. These cells were attached to one
another and had identical diameters; therefore, pressure gradients
were <2%. We should note that the difference between the
effective path length and simply measuring the entire physical
length of the reaction cell was not large, <5%.
Other tests of the path length included adding flush gases at
the ends of the reaction cell to ensure that there was no stagnant
region and recalibrating the effective path length. In another
test, windows were inserted into the absorption/reaction cell to
the positions where the flow entered and exited the reaction
cell and the physical path length was used. In yet another
experiment, the physical path length was increased by 13 cm.
For each of these tests k1(298 K) was remeasured; all of these
measurements were within 10% of one another.
†
22
†
Cl + ClO f O + Cl
(20)
2
5
was responsible for the loss of ClO. Poulet et al. noted ClO
losses of up to 50% using reactions 2 and 19 under conditions
of excess Cl atoms. Although reaction 19 is not exothermic
†
enough to produce ClO , the production of Cl2 was consistent
with losses in the injector, most likely because of recombination
of ClO (reaction 7).
In our experiments, we determined the concentration of ClO
(after its production had gone to completion), the initial
Although somewhat redundant, since changes in temperature
only change the number density in the reaction cell, we
remeasured this path length under each set of temperatures and
flow conditions. For each of these temperatures the path length
was also calculated using the physical lengths of the sections
held at different temperatures (these lengths were obtained by
measuring the temperature gradient along the cell). These
calculated lengths were within 3% of the calibrated effective
path lengths.
concentration of ozone ([O3]0), and the concentration of any
remaining ozone ([O3]excess) by UV/visible absorption. Figure
6a displays a plot of the measured [ClO] vs [O3]0 obtained under
conditions where O3 was depleted. If the amount of ClO
produced were equal to [O3]0, the data points would fall on the
1
-to-1 line. This figure shows that even at low [ClO] the 1-to-1
conversion was not obeyed. The low conversion is consistent
5
22
with the observations of Poulet et al. and Burkholder et al.
1
3
over the [ClO] ranges of (0.25-2.0) × 10 and (0.4-1.35) ×
On the basis of all these tests, we are confident that [ClO]
was determined within 10% if the absorption cross section of
ClO at 253.7 nm is exact. The ClO cross section has been
measured by many different techniques to obtain the same value
13
-3
1
0
molecule cm , respectively. Figure 6b is a plot of [ClO]
vs ∆[O3] for a few experiments done with O3 in excess. In both
of these types of experiment the conversion efficiency varied
depending upon experimental conditions and was somewhat
irreproducible. The conversion efficiency may very well change
with temperature. The temperature dependence of the conversion
efficiency could be one explanation for the discrepancy between
those experiments that report a small or no temperature
(
within (5%). Therefore, we expect our measured [ClO] and
k1 to be accurate.
In kinetic studies of reaction 1 carried out under pseudo-
first-order conditions in OH, aside from influences on the OH
temporal profiles, any uncertainties in [ClO] or errors in
4
,7
dependence and those that report a negative temperature

OH + ClO f Products
J. Phys. Chem. A, Vol. 103, No. 26, 1999 5047
tests showed that their ClO calibration should have been accurate
within a few percent. As discussed earlier, we observed a wide
range of conversion efficiencies for the production of ClO from
3
O3. Hills and Howard did not observe this affect with ClO
1
3
-3
concentrations of (0.14-1.25) × 10 molecule cm .
5
Poulet et al. measured k1(298 K) using a discharge-flow
system and laser-induced fluorescence to detect OH and mass
spectrometry to detect ClO and HCl. They measured k1 in three
different ways. The first measurement was relative to the
reaction
OH + OClO f products
(23)
-
12
where k23(298 K) was first determined to be (6.9 ( 0.5) × 10
3
-1 -1
cm molecule s . Determination of k1 required correction
(
∼28%) for the regeneration of OH via reaction 3a and yielded
-11 3 -1 -1
a value of k1 ) (1.77 ( 0.33) × 10
cm molecule s . In
the second method, ClO was generated via reactions 2 or 9 in
an excess of Cl atoms. ClO was titrated with NO and the
resulting NO2 signal compared to a calibrated NO2 signal to
determine [ClO], as in the experiment by Hills and Howard.3
Again, this required correction (∼28%) for the regeneration of
-
11
3
OH and yielded a value of (1.99 ( 0.25) × 10
cm
-1 -1
molecule s . In the third method, ClO was made via reaction
with a slight excess of O3. In these experiments ClO was not
2
titrated with NO because the Cl atoms produced (reaction 22)
can react to regenerate ClO (reaction 2) and because NO reacts
with O3 to produce NO2. Instead, a mass spectrometric calibra-
tion of ClO was used to determine [ClO]. However, for this
calibration, Cl atoms were added until the O3 signal disappeared,
and then it was assumed that [ClO] ) [O3]0. From this
3 0
Figure 6. (a) Plot of the measure [ClO] vs [O ] obtained under
conditions where O
ClO] vs ∆[O ] for a few experiments done with O
the symbols represents data taken during different days or under varying
experimental conditions. If the production of ClO from O were
3
was depleted during ClO production. (b) Plot of
[
3
3
in excess. Each of
3
-
11
stoichiometric, the data points would fall on the 1-to-1 lines shown.
experiment they obtained k1(298 K) ) (1.89 ( 0.21) × 10
3
-1 -1
cm molecule s .
5
3
Both Poulet et al. and Hills and Howard report values for
k1 larger than those from any previous experiments. Each of
these groups carefully chose their experimental conditions to
avoid complications from secondary chemistry and tested for
the accuracy of their ClO concentration. Although these
measurements do overlap within the 2σ uncertainties of one
another, our measured k1(298 K) in excess O3 is larger than
their reported values by ∼20% and ∼30%, respectively. A
possible source for a systematic discrepancy could be the
necessity of both groups to calibrate [ClO] under conditions of
a slight excess of Cl atoms while measurements of k1 were
performed under conditions of excess O3. If there were a
systematic error in the ClO calibration, this could also affect
dependence.3,8 In all of our experiments, the source region for
ClO production was held at 298 K. Our experiments, along with
those of Burkholder et al.22 and Poulet et al., show that one
cannot assume unit conversion of ozone to ClO, and those of
5
5
Poulet et al. showed that this was also true when reaction 19
was used to produce ClO.
Hills and Howard3 were the first to report a negative
-12
temperature dependence (k1(T) ) (8.0 ( 1.4) × 10 exp[(235
-1
-1
(
46)] cm3 molecule s ) for reaction 1. They used a
discharge-flow system with laser magnetic resonance capable
of detecting OH, ClO, and HO2, although not simultaneously.
The rate coefficient was determined by measuring OH decays
in an excess of ClO. The experiments were performed in a slight
excess of O3 to prevent the regeneration of OH from reaction
of HO2 with Cl. In general, the ClO concentration was calibrated
from a plot of ClO signal vs [O3]0 obtained under conditions of
excess Cl atoms and assuming [ClO] ) [O3]0. As a test, Hills
23
other rate coefficients that relied on a similar calibration.
8
More recently, Lipson et al. used the turbulent flow technique
with chemical ionization mass spectrometry to measure k1 and
-
12
report a value k1(T) ) (5.5 ( 1.6) × 10
exp[(292 ( 72)]
3
-1 -1
cm molecule s . In their experiments, ClO was generated
from reaction 2. The ClO signal was calibrated with NO via
reaction 22, and ethane was added to the scavenge Cl atoms,
3
and Howard verified the [O3]0 by converting it to NO2 via
reaction with NO,
O + NO f NO + O
2
(21)
3
2
Cl + C H f HCl + C H
5
(24)
2
6
2
and compared the NO2 signal to an absolute calibration for NO2.
Similarly, to test the accuracy of the ClO concentration obtained
from the calibration plot, they converted ClO to NO2 via reaction
with NO,
preventing regeneration of ClO from reaction 2. As they noted,
one complication with this titration is that NO2 can react with
C2H5 formed in reaction 24:
NO + C H f products
(25)
2
2
5
ClO + NO f NO + Cl
(22)
2
They modeled this titration system to correct for the underes-
timation of the ClO concentration and report that the correction
factor was almost always <15%. Simulation of this titration
scheme requires accurate knowledge of the concentrations of
and compared the NO2 signal to an absolute calibration for NO2.
Finally, they titrated ClO with small amounts of NO and
observed the falloff in ClO signal as ClO was consumed. These

5048 J. Phys. Chem. A, Vol. 103, No. 26, 1999
Kegley-Owen et al.
NO, ClO, OH, O3, Cl2, Cl, and C2H6 for each temporal profile.
Therefore, we were unable to evaluate the model used in their
titration scheme.
(2) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for Use in Stratospheric
Modeling; Jet Propulsion Laboratory: Pasadena, CA, 1997.
2
From the enthalpies of formation, ∆fH°298, and k3a of (9.1
(3) Hills, A. J.; Howard, C. J. J. Chem. Phys. 1984, 81, 4458.
-
12
3
-1 -1
(
(
1.3) × 10
cm molecule s , we calculated ∆fH°298-
(4) Burrows, J. H.; Wallington, T. J.; Wayne, R. P. J. Chem. Soc.,
-1
HO2) to be 3.0, 3.2, 3.1, and 3.3 kcal mol , respectively, using
Faraday Trans. 2 1984, 80, 957.
3
5
k1 measured by us, Hills and Howard, Poulet et al., and Lipson
(5) Poulet, G.; Laverdet, G.; LeBras, B. J. Phys. Chem. 1986, 90, 159.
(6) Leu, M. T.; Lin, C. L. Geophys. Res. Lett. 1979, 6, 425.
8
et al., respectively. For this calculation the channel producing
(
7) Ravishankara, A. R.; Eisele, F. L.; Wine, P. H. J. Chem. Phys.
983, 78, 1140.
8) Lipson, J. B.; Elrod, M. J.; Beiderhase, T. W.; Molina, L. T.;
Molina, M. J. J. Chem. Soc., Faraday Trans. 1997, 93, 2665.
9) Turnipseed, A. A.; Gilles, M. K.; Burkholder, J. B.; Ravishankara,
HCl was neglected, since it is, at most, a minor channel. These
1
values all lie within the currently accepted value of (2.8 ( 0.5)
(
-
1
8
kcal mol , although the value of Lipson et al. is at the limit
of the uncertainty. Although our value of k1 is higher than
previously reported values, it is consistent with the enthalpy of
formation for HO2.
(
A. R. J. Phys. Chem. A 1997, 101, 5517.
(10) Baulch, D. L.; Cox, R. A.; Crutzen, P. J.; Hampson, R. F., Jr.;
Kerr, J. A.; Troe, J.; Watson, R. T. J. Phys. Chem. Ref. Data 1982, 11,
This is the first measurement of k1 where [ClO] was
determined by its absorption spectrum. The value determined
here is larger than those reported previously. Our in situ
measurement of [ClO] by UV/visible absorption eliminated the
need to titrate ClO to determine its concentration. Therefore, it
3
73.
11) Turnipseed, A. A.; Vaghjiani, G. L.; Thompson, J. E.; Ravishankara,
A. R. J. Chem. Phys. 1992, 96, 5887.
(
(12) Davis, H. F.; Lee, Y. T. J. Phys. Chem. 1996, 100, 30.
(13) Vaghjiani, G. L.; Ravishankara, A. R. J. Phys. Chem. 1989, 93,
1
appears that our value of k1 is accurate. Dubey et al. recently
1948.
showed (using a smaller rate coefficient than was measured in
this work) that a branching ratio for channel 1b as small as 7%
could be significant in the chlorine partitioning in the middle
and upper stratosphere. By use of our rate coefficient, a similar
influence could be achieved with a much smaller branching ratio
for channel 1b. Other implications of the larger value of k1 await
analysis by modeling studies.
(14) Cady, G. H. Inorg. Synth 1957, 5, 156.
(15) Gilles, M. K.; Burkholder, J. B.; Ravishankara, A. R. Int. J. Chem.
Kinet., in press.
(16) Gilles, M. K.; Turnipseed, A. A.; Burkholder, J. B.; Ravishankara,
A. R.; Solomon, S. J. Phys. Chem. A 1997, 101 5526.
(17) Lee, Y. P.; Howard, C. J. J. Chem. Phys. 1982, 77, 746.
(18) Burrows, J. P.; Cliff, D. I.; Harris, G. W.; Thrush, B. A.; Wilkinson,
J. P. T. Proc. R. Soc. London 1979, A368, 463.
19) Cattell, F. C.; Cox, R. A. J. Chem. Soc., Faraday Trans. 2 1986,
82 1413.
(20) Dobis, O.; Benson, S. W. J. Am. Chem. Soc. 1993, 115, 8798.
(
Acknowledgment. This work was funded in part by NASA
Upper Atmospheric Research Program. This work constituted
part of the Ph.D. thesis submitted by C. S. K.-O. to the
University of Colorado, Boulder.
(21) Malleson, A. M.; Kellett, H. M.; Myhill, R. G.; Sweetenham, W.
P. FACSIMILE.; A. E. R. E. Harwell Publications Office: Oxfordshire,
990.
1
(
22) Burkholder, J. B.; Hammer, P. D.; Howard, C. J.; Goldman, A. J.
References and Notes
Geophys. Res. 1989, 94, 2225.
(
1) Dubey, M. K.; McGrath, M. P.; Smith, G. P.; Rowland, R. S. J.
(23) Stimpfle, R. M.; Perry, R. A.; Howard, C. J. J. Chem. Phys. 1979,
71, 5183.
Phys. Chem. A 1998, 102 3127.