Kinetics of the IO radical. 1. Reaction of IO with ClO

Article abstract of DOI:10.1021/jp970914g

The rate coefficient for the IO + ClO → products (5) reaction was measured by coupling discharge flow and pulsed laser-induced fluorescence techniques. Rate coefficients were measured from 200 to 362 K by monitoring the temporal profile of IO in an excess of ClO. The rate coefficient is described by the expression: k₅(T) = (5.1 ± 1.7) × 10^-12 exp[(280 ± 80)/T] cm³ molecules^-1 s^-1 where the quoted uncertainties include estimated systematic errors. Atomic iodine was identified as a major product of reaction 5. A branching ratio of Φ = 0.14 ± 0.04 at 298 K was obtained for the sum of channels which do not produce I atoms.

Full text of DOI:10.1021/jp970914g

J. Phys. Chem. A 1997, 101, 5517-5525
5517
Kinetics of the IO Radical. 1. Reaction of IO with ClO
Andrew A. Turnipseed,† Mary K. Gilles, James B. Burkholder, and A. R. Ravishankara*
Aeronomy Laboratory, National Oceanic and Atmospheric Administration, Boulder, Colorado 80303, and
CooperatiVe Institute for Research in EnVironmental Sciences, UniVersity of Colorado,
Boulder, Colorado 80309
ReceiVed: March 12, 1997; In Final Form: May 9, 1997X
The rate coefficient for the IO + ClO f products (5) reaction was measured by coupling discharge flow and
pulsed laser-induced fluorescence techniques. Rate coefficients were measured from 200 to 362 K by
monitoring the temporal profile of IO in an excess of ClO. The rate coefficient is described by the
expression: k5(T) ) (5.1 ( 1.7) × 10-12 exp[(280 ( 80)/T] cm3 molecule-1 s-1 where the quoted uncertainties
include estimated systematic errors. Atomic iodine was identified as a major product of reaction 5. A branching
ratio of Φ ) 0.14 ( 0.04 at 298 K was obtained for the sum of channels which do not produce I atoms.
Introduction
of iodine can make a significant contribution to ozone loss at
low altitudes.
The reactivity of chlorine- and bromine-containing radicals
are of interest in atmospheric chemistry due to their ability to
catalytically destroy stratospheric ozone. However, iodine-con-
taining radicals have received much less attention. It is known
that the oceans are a significant source of iodine to the atmos-
phere, primarily in the form of CH3I.1-3 The possible role of
iodine in tropospheric chemistry, and in particular, the marine
boundary layer, has been described previously.4,5 More recently,
halogen chemistry has been implicated in the sporadic rapid
ozone loss observed in the Arctic spring after sunrise.6,7 Iodine
may lead to catalytic ozone depletion via the reaction sequence
Solomon et al.8 pointed out that the trend in low-altitude,
midlatitude ozone loss observed over the last decade could be
explained by the presence of small concentrations of iodine (e1
pptv) interacting with chlorine and bromine. This trend could
reflect the increasing chlorine abundance in the stratosphere over
this time period. If all of this ozone depletion is attributed to
the IO + ClO cycle, the rate coefficient for the reaction
IO + ClO f products
(5)
has to be large (≈10-10 cm3 molecule-1 s-1), the reaction
products need to be atomic iodine and chlorine or lead to them
in the atmosphere, and the stratospheric abundance of iodine
has to be ∼1 pptv. Lower values of iodine and/or the rate
coefficient, k5, will lead to a proportionately smaller contribution
to ozone depletion.
The chemistry of iodine also merits concern due to the
possible anthropogenic emissions of iodine-containing com-
pounds. If a significant fraction of these compounds reach the
stratosphere, either by rapid transport or by direct release into
the upper atmosphere, then the effectiveness of their interaction
with chlorine or bromine in ozone destruction must be known.
For example, when the ozone depletion potential of a compound
such as CF3I is evaluated, the reactions of iodine need to be
considered.
We have used pulsed laser photolysis/pulsed laser-induced
fluorescence (LP-PLIF) to produce and monitor IO radicals. We
have combined a discharge flow source for producing ClO
radicals with UV absorption for determining [ClO] to measure
the rate coefficient for reaction 5 over the temperature range
important in the atmosphere. The I atom yield in reaction 5
was also determined. In the second part of our study we have
measured the IO + BrO reaction rate coefficient and discuss
the atmospheric implications of the kinetic measurements.
I + O3 f IO + O2
X + O3 f XO + O2
IO + XO f f I + X + O2
net: 2O3 f 3O2
(1)
(2)
(3)
(4)
where X represents Cl, Br, I, or OH. Reaction 3 is denoted
with a double arrow to indicate that it may proceed by more
than one elementary step. Because reaction 3 is the rate-limiting
step in this sequence, the rate coefficients for these reactions
must be known to ascertain the effectiveness of these cycles.
Solomon et al.8 recently reevaluated the possible role of iodine
in the catalytic destruction of stratospheric ozone. Previously
it was assumed that iodine could play a significant role only in
tropospheric chemistry, due to the short tropospheric photolysis
lifetime of iodine-containing compounds.4,5 However, Solomon
et al.8 postulated that iodine-containing compounds can be
rapidly transported to the stratosphere by strong convective
transport in the tropics. Once in the stratosphere, these
compounds would readily photolyze to release I atoms which
then catalytically destroy ozone via reactions 1-3. Since the
reservoir species for iodine in the atmosphere (i.e., HI, HOI, or
IONO2) are photolytically unstable, most of the atmospheric
iodine will be in the active forms of I or IO. Therefore, iodine
is particularly effective at destroying ozone, and small amounts
Experimental Section
Measurement of rate coefficients for a bimolecular reaction
under pseudo-first-order conditions requires that the concentra-
tion of one reagent be constant and known. The easiest way to
attain this requirement is to have a large excess of one reagent
over the other. First-order kinetics has the advantage that only
the absolute concentration of the excess reagent is needed to
determine the rate coefficient. The accuracy of the measured
rate constant is directly related to the accuracy with which the
* Address correspondence to this author at: NOAA/ERL, R/E/AL2, 325
Broadway, Boulder, CO 80303. Also affiliated with the Department of
Chemistry and Biochemistry, University of Colorado, Boulder, CO 80309.
† Current address: Department of Biology (EPOB), University of
Colorado, Boulder, CO 80309.
X Abstract published in AdVance ACS Abstracts, July 1, 1997.
S1089-5639(97)00914-6 CCC: $14.00 © 1997 American Chemical Society

5518 J. Phys. Chem. A, Vol. 101, No. 30, 1997
Turnipseed et al.
Figure 1. Schematic of the pulsed laser photolysis/pulsed laser-induced fluorescence system coupled to a discharge flow system for generating
ClO radicals (P ) capacitance monometer, MW ) microwave, HV ) high voltage).
excess reagent concentration is measured. In a conventional
pulsed photolysis system, the excess reagent is usually a stable
gas whose concentration is measured using flow rates and
pressure or UV absorption. However, to study the reaction of
IO + ClO under pseudo-first-order conditions, one of the
unstable radical species must be generated in large excess, its
concentration must be measured, and any of its losses due to
reactions (i.e., self-reaction) must be quantified. In addition,
the temporal profile of the other radical must be measured. Due
to the slow self-reaction9 of ClO and the sensitivity with which
IO can be detected by laser-induced fluorescence,10 we chose
to study this reaction under conditions of [ClO] > 10[IO]0
(pseudo-first-order in IO). The experimental apparatus (Figure
1) combines discharge flow (DF), pulsed laser photolysis/pulsed
laser-induced fluorescence (PLP/PLIF), and UV absorption
spectroscopy. The merging of techniques which are normally
optimized under very different experimental conditions (i.e.,
pressure, flow rate, and velocity) leads to some complications
that are not normally encountered (for example, temperature
gradients and fast loss due to transport from the viewing zone).
These complications and their influence on the measurements
are addressed in the following sections.
ClO Generation. The flow tube consisted of a 2.54 cm
Pyrex tube held at 298 K with a movable injector (0.6 cm o.d.)
down the center of the tube. Typically a large flow of N2 or
He/N2 was added such that the flow rate was 13-60 STP cm3
s-1, total pressure was 5-16 Torr, and linear gas flow velocity
was V ) 500-1000 cm s-1 in the flow tube. Cl atoms were
generated in a side arm of the flow tube by passing a dilute
(0.1-1% Cl2) mixture of Cl2 in He through a microwave
discharge. Within the flow tube, Cl atoms reacted with either
Cl2O, OClO, or O3 which were added through the movable
injector.
Cl + Cl2O f ClO + Cl2
(6)
Cl + OClO f 2ClO
Cl + O3 f ClO + O2
(7)
(8)
Concentrations of Cl2O, OClO, or O3 were always sufficiently
large to react away any Cl atoms, which avoided removal of
IO or its precursors via their reactions with Cl atoms. The
residence time within the flow tube was ∼15 ms, which was
sufficient for complete consumption of the Cl atoms. However,
since ClO concentrations were typically <1014 molecule cm-3
little ClO was lost due to its self-reaction:
,
ClO + ClO f ClOO + Cl
f Cl2 + O2
(9a)
(9b)
(9c)
(9d)
f OClO + Cl
9
M8 Cl2O2
where k9(298 K) ) 2 × 10-14 cm3 molecule-1 s-1 at 5 Torr.9
It should also be noted that channels 9a and 9c produce Cl atoms
which were recycled rapidly to ClO by reaction with the excess
Cl2O, OClO or O3. Thus, not all of reaction 9 leads to ClO
removal, and the loss was quite small.
From the flow tube, the gases passed through a narrow (1.3
cm i.d.) 50 cm preliminary absorption cell held at 298 K. This

Kinetics of the IO Radical. 1
J. Phys. Chem. A, Vol. 101, No. 30, 1997 5519
cell was used to calibrate the path length of the photolysis cell
as described below. The residence time within the preliminary
absorption cell was ∼20 ms.
Atot
Leff
LHg
T
)
(III)
(
)
AHg 298
UV Absorption Spectroscopy. The gas effluent from the
preliminary absorption cell passed into the photolysis cell which
consisted of an 18 cm long jacketed Pyrex tube (2.0 cm i.d.)
with orthogonal windows in the center. ClO concentrations
were measured through the photolysis cell via UV absorption
spectroscopy. The output of a D2 lamp was collimated and
passed through the photolysis cell along the axis of the flow
(counter to the path of the photolysis laser). The beam was
then focused onto the entrance slit of a 0.5 m spectrometer.
The spectrometer employed a 300 grooves/mm grating and a
cooled 1024 element diode array detector, which allowed
measurements over a 165 nm wide region. The ClO concentra-
tion was quantified using the unstructured portion of the
spectrum since the cross section in this region is known to be
independent of temperature and resolution.11 The spectral
subtraction, however, utilized the structured region of the
spectrum. A slit width of 150 µm resulted in a spectral
resolution of 0.8 nm (fwhm). For monitoring ClO, OClO, and
Cl2O, the spectrometer was typically set to observe from 220
to 385 nm. This wavelength range was calibrated using the
emission lines from a low-pressure Hg lamp. Reference spectra
of ClO and OClO needed for spectral subtraction were measured
using the same spectrometer.
where AHg and LHg are, respectively, the measured absorption
(using the Hg lamp) and the path length (50 cm) for the
preliminary absorption cell.
Reference spectra for the species measured in the present
study are shown in Figure 2a and are in good agreement with
those in the literature.9,13 ClO reference spectra were measured
prior to each experiment by titrating either OClO, O3, or Cl2O
with an excess of Cl atoms (reaction 6, 7, or 8). These reference
spectra were recorded under flow, pressure, and temperature
conditions that were identical to the subsequent kinetic experi-
ments. In contrast to the reference spectral measurements, in
the kinetics experiment, ClO was produced in an excess of Cl2O,
OClO (OClO was used only at 298 K), or O3 (which was used
only for the I atom yield experiments). The ClO concentration
was measured after generating IO and measuring its temporal
profile using the PLP/PLIF (described below). The photolysis
laser was blocked, and the D2 beam was passed through the
photolysis cell to measure the ClO/Cl2O, ClO/OClO, or ClO/
O3 spectrum. I0 for this measurement was obtained after turning
off the microwave discharge and shutting off the flow of Cl2O,
OClO, or O3. The absorption due to ClO was extracted from
the measured spectrum by subtracting the scaled ClO reference
spectrum until all of the ClO structured absorption was
completely removed. An example of this subtraction using Cl2O
is shown in Figure 2b, and the residual is shown in Figure 2c.
The abundance of ClO subtracted to get the unstructured residual
was used to determine the [ClO] by using the absorption at 253.7
nm (Atot ) Leffσ[ClO]T, σ ) 4.25 × 10-18 cm2 molecule-1).9
PLP/PLIF System. IO radicals were generated in the
photolysis reactor following the 193 nm photolysis of N2O/
CF3I or N2O/I2 mixtures in N2:
As seen in Figure 1, only part of the optical path was at the
temperature of interest. Also, to minimize the formation of solid
deposits on the entrance and exit windows, pure He or N2 was
flushed in front of the windows, slightly reducing the optical
path length. Therefore, it was necessary to measure the
“effective” absorption path length through the photolysis cell
under each set of experimental conditions (flows, pressure,
temperature). In the simplest case, the path length can be
divided into two regions: (1) the temperature-controlled region
at temperature T and (2) the 298 K region, the sections before
and after the photolysis cell. The temperature gradients at the
entrance and exit of the temperature-controlled region were
measured using a movable thermocouple probe and compensated
for each other. The absorption due to ClO measured through
the photolysis cell was
N2O + hν f N2 + O(1D)
O(1D) + N2 f O(3P) + N2
(10)
(11)
O(3P) + CF3I f CF3 + IO
f products
(12a)
(12b)
(13)
Atot ) AT + A298 ) σTLT[ClO]T + σ298L298[ClO]298 (I)
O(3P) + I2 f I + IO
where Atot ) -ln (I/I0) (I being the light intensity with ClO
present and I0 when it is not) and σ is the absorption cross
section. LT and L298 are the path lengths at temperatures T and
298 K, respectively. Knowing that [ClO]T ) [ClO]298(298 K/T)
and that σ at 253.7 nm is temperature independent, [ClO]T was
described by
CF3I or I2 was added upstream of the photolysis cell near the
end of the flow tube, and N2O was added with the main N2 or
He flow. Additional small flows of either He or N2 were
introduced right in front of the entrance and exit windows
through which the photolysis laser beam was passed. This
procedure, as mentioned earlier, minimized deposits on the
windows. Without flush gases, solid deposits gradually formed
on the windows. These flushes were <15% of the total gas
flow. Due to the large flow velocity in the LIF reactor, the
photolysis laser was passed through the photolysis cell along
the axis of the flow to minimize loss of IO due to transport
from the viewing region.
The IO temporal profiles were measured using a Nd:YAG
pumped dye laser tuned to the band head of the A2Π3/2 (V′ )
2) r X2Π3/2 (V′′ ) 0) transition at 444.8 nm, as described
previously.10 The dye laser beam intersected the photolysis laser
at 90° at the center of the cell. Fluorescence was collected
orthogonal to both laser beams by a convex lens (focal length
of 5 cm), passed through a cutoff filter (λ > 475 nm) to
discriminate against scattered light from the probe laser, and
Atot
[ClO]T )
(II)
σLeff
where Leff ) L298(T/298) + LT. Leff was measured prior to each
experiment by adding O3 to the flow tube and measuring its
absorption at 253.7 nm (Hg line) in the preliminary absorption
cell (all at 298 K). The O3 then passed through the LIF reactor
where its absorption was again measured by the D2 lamp/diode
array spectrometer at the same wavelength. O3 was chosen as
our standard since it is known that its absorption cross section
at 253.7 nm (σ ) 1.16 × 10-17 cm2 molecule-1) does not
significantly change with temperature (<2% change from 298
to 200 K).12 The ratio of the absorptions in the two cells was
used to calculate Leff.

5520 J. Phys. Chem. A, Vol. 101, No. 30, 1997
Turnipseed et al.
purification and were added to the reactor through calibrated
electronic mass flow meters. Their concentrations were cal-
culated using the flow rates and the total pressure, which was
measured by a 50 Torr capacitance manometer. I2 (99.8%) was
introduced to the photolysis cell as a mixture in N2. A small
flow of N2 was passed over I2 crystals and the mixture passed
through an absorption cell equipped to monitor 508 nm light
from a Cd lamp (σ ) 2.5 × 10-18 cm2 molecule-1).14 This
flow was then diluted by the main gas flow. The concentrations
of N2O, CF3I, and I2 are not essential for determining the IO +
ClO rate coefficient. However, they are needed (along with
the measured photolysis fluence) to estimate the concentration
of IO to insure pseudo-first-order conditions and to estimate
contributions of secondary reactions.
A mixture of 9.5% Cl2/He (UHP, 99.9% Cl2 in UHP He)
was further diluted to <1% for generating Cl atoms, and its
flow rate was measured using a mass flow meter. Cl2O was
synthesized as described previously15 and stored at 195 K. Cl2O
was added to the reactor by bubbling a small flow of N2 or He
through the liquid. OClO was produced on-line by passing a
small flow of the 9.5% Cl2/He mixture over glass beads covered
with sodium chlorite (80%). The Cl2 to OClO conversion was
nearly 100%. O3 was produced by passing O2 (UHP, 99.99%)
through an ozonizer and trapping the O3 on silica gel at 195 K.
It was added to the flow tube by passing a flow of N2 or He
through the silica gel trap. Cl2O, OClO, and O3 were all
measured by UV absorption along with ClO in the photolysis
cell.
Figure 2. (a) Absorption spectra of species measured in the present
study. Spectra are scaled to the absorption cross sections for each species
as reported in ref 9. (b) Example of the absorption spectrum of ClO in
an excess of Cl2O at 298 K and the deconvolved ClO and Cl2O spectra
obtained after spectral subtraction of a ClO reference spectrum. (c)
The residual shown was obtained by subtracting ClO, Cl2O, and a small
amount of Cl2 from the spectrum labeled ClO + Cl2O in (b).
Results
ClO Generation. For measuring k5, the total rate coefficient
for reaction 5, ClO {(0.35-9.5) × 1013 molecule cm-3} was
produced via reactions 6 or 7 with Cl2O and OClO concentra-
tions in excess ([Cl2O] ) (5-14) × 1013 and [OClO] ) (3-7)
× 1013 molecule cm-3). IO was produced in a pulsed manner
after ClO formation was complete to keep the concentration of
Cl low ([Cl] < 1010 atom cm-3) and to suppress reactions of
Cl with IO and its precursors:
imaged onto a photomultiplier tube (PMT). A second filter,
which transmitted λ < 600 nm, was used to discriminate against
an emission at wavelengths longer than 600 nm. This emission
was present when the ClO source was on. The output of the
PMT was sent to a gated charge integrator and microcomputer
for analysis. The delay time between the photolysis and probe
laser pulses was varied between 20 µs and 30 ms to obtain the
temporal profile of IO. Even the shortest delay time of 20 µs
allowed for thermalization and completion of any O(1D)
reactions. Removal from the viewing zone of the reactor was
nearly complete in 30 ms. The detection limit for IO,
determined from [N2O] and laser fluence, was typically ∼1 ×
109 molecule cm-3 (S/N ) 1, averaging 100 laser shots).
The temperature in the photolysis cell, where the IO temporal
profiles were measured, was regulated by flowing either cooled
methanol or heated ethylene glycol from a thermostated bath
through the jacket. The temperature at the intersection point
of the photolysis and probe lasers was measured under flow
and composition conditions, identical to those in which the IO
temporal profiles were monitored by inserting a calibrated
thermocouple through a movable tube located opposite the PMT.
Because the pressures and fast flows were significantly different
from the familiar discharge flow system, temperature gradients
were of a concern. When N2 was used as the bath gas, a large
temperature gradient (up to 25 °C) between the solvent reservoir
and the photolysis cell was observed. The use of He with ∼1
Torr of N2 added {to quench O(1D)} nearly eliminated this
gradient. The measured temperatures were found to be constant
to (2 K over this region.
Cl + IO f ClO + I
Cl + I2 f ICl + I
(14)
(15)
Cl + CF3I f ICl + CF3
(16)
Bedjanian et al.16 recently reported temperature independent rate
coefficients of k14 ) (4.4 ( 1.0) × 10-11 cm3 molecule-1 s-1
and k15 ) (2.1 ( 0.3) × 10-10 cm3 molecule-1 s-1, respectively.
A rate coefficient of k16(T) ) 6.25 × 10-11 exp[-1500/T] cm3
molecule-1 s-1 is also reported.17 Excess Cl2O and OClO also
helped negate loss of ClO due to reactions 9a and 9c. With
typical residence times of 15 and 20 ms in the flow tube and
preliminary absorption cell, ClO loss due to its self-reaction
was calculated to be <5% in these regions for the highest ClO
concentrations used.
Within the photolysis cell, the typical residence time at 298
K was ∼70 ms. During this time, the loss of ClO due to the
bimolecular self-reaction was calculated to be <15% at 298 K.
Since this loss was small, the average ClO concentration
measured by UV absorption was nearly identical to the ClO
concentration within the reaction zone, i.e. the middle of the
photolysis cell. At low temperatures the bimolecular channels
are greatly retarded; however, reaction 9d becomes faster. At
200 K and 5 Torr, the effective bimolecular rate coefficient for
Reagents. N2 (UHP, >99.999%), He (UHP, >99.9999%),
N2O (UHP, 99.99%), and CF3I (99%) were used without
reaction 9d is k9d ) 8.7 × 10-15 cm3 molecule-1 s-1 9
. Computer
simulations showed that this rate coefficient results in a

Kinetics of the IO Radical. 1
J. Phys. Chem. A, Vol. 101, No. 30, 1997 5521
describes the loss of O(3P) atoms via transport out of the
detection zone or reactions with impurities. The IO temporal
profiles were fit to eq IV to obtain k′, at various concentrations
a
of CF3I. k12 was obtained from a linear least-squares fit of k′
a
vs [CF3I] yielding a value of k12(298 K) ) 4.7 × 10-12 cm3
molecule-1 s-1, in good agreement with our recent results.19
The intercept of the fit yielded a value of 1000 ( 150 s-1 for
the loss of O(3P) via reaction 19.
The observed IO decays were exponential up to reaction times
of 10-15 ms. The measured values of k18 ranged from 120 to
300 s-1 and were dependent upon the gas flow rate, pressure,
and temperature. IO concentrations were calculated, using the
known [N2O] and photolysis fluence, to be <3 × 1011 molecule
cm-3. These low IO concentrations made the self-reaction of
IO unimportant. The decay profiles of IO were strictly
exponential and showed no evidence for a second-order loss
process, as would be expected if the self-reaction of IO were
occurring to a significant extent. Furthermore, variation of either
[N2O] or the photolysis fluence over a factor of 5 did not affect
the measured value of k18, suggesting that the primary loss of
IO was due to transport out of the reaction zone.
Figure 3. IO temporal profiles measured from the 193 nm photolysis
of N2O/CF3I in an excess of ClO. ClO concentrations are (a) 0, (b)
1.68 × 1013, and (c) 6.19 × 1013 molecule cm-3. The solid lines are
fits to eq V.
calculated ClO loss of ∼6% in the temperature-controlled region
at the highest [ClO] that was used. Note that approximately
half of the total path length through the photolysis cell is in the
temperature-controlled region; therefore, total losses due to
reaction 9 are calculated to be <20% at the lowest temperatures
and highest [ClO] studied. Although it is difficult to detect in
the presence of excess Cl2O, no evidence for Cl2O2 (ClO dimer)
was observed in the UV absorption spectra taken at low
temperatures suggesting that its formation was small. At high
temperatures, the bimolecular channel leading to Cl2 becomes
the dominant loss of ClO; however, the total ClO loss was
always calculated to be <20%.
In 5-15 Torr of N2, all of the O(1D) is quickly quenched to
O(3P). However, when 4 Torr of He and 1 Torr of N2 were
used, up to 40% of the O(1D) reacted with N2O to generate NO
via
O(1D) + N2O f N2 + O2
f 2NO
(20a)
(20b)
However, since the initial [O(1D)] was kept low, the NO
produced was small and did not influence the IO temporal
profile. As stated above, variation of the [N2O] (over a factor
of 5), and consequently the concentration of photolytically
produced NO, resulted in no measurable change in k18.
Therefore, the amounts of NO that were produced were
insignificant.
OClO was used as the ClO precursor only at 298 K. At low
temperatures OClO reacts18 with ClO to form Cl2O3
ClO + OClO
9
M8 Cl2O3
(17)
IO + ClO Rate Coefficient. Initial tests were conducted to
ensure that IO did not react with Cl2, N2O, or Cl2O and that I
atoms did not react with N2O. Addition of Cl2 (1 × 1015
molecule cm-3) or Cl2O (8 × 1014 molecule cm-3) did not affect
the IO temporal profile, and we place upper limits of k21 < 2
× 10-14 and k22 < 3 × 10-14 cm3 molecule-1 s-1 at 298 K for
the reactions
resulting in a large ClO loss. Therefore, reaction 6 was the
only ClO source used for low-temperature measurements.
IO Production. The 193 nm photolysis of N2O/I2 or N2O/
CF3I was used to generate IO radicals via reactions 10-13.
Because of the low vapor pressure of I2 at T < 250 K, CF3I
was used as the IO source in the low-temperature experiments.
The reaction of O(3P) + CF3I is known to produce IO with
∼85% yield with a rate coefficient of k12(T) ) 7.9 × 10-12
conditions ([N2O] ) (1-5) × 1015 molecule cm-3, [CF3I] )
(4-8) × 1014 or [I2] ) 5 × 1013 molecule cm-3 and E ) (0.2-
1.0) mJ pulse-1 cm-2), the IO signal reached a maximum value
within 0.1-1 ms. A typical IO temporal profile in the absence
of ClO is shown in Figure 3, curve a. The profile exhibits a
fast exponential rise, reaction 12a, followed by a slower decay.
The slow decay is due to pump out of the detection region and
is described by
IO + Cl2 f products
IO + Cl2O f products
(21)
(22)
exp[-175/T] cm3 molecule-1 s-1 19
Under typical experimental
.
Maintaining a constant [IO], varying the concentration of N2O
over a factor of 5 did not influence the IO temporal profile.
The 351 nm photolysis of CH2I2 to produce I, in the presence
of 35 Torr of N2O, did not produce detectable concentrations
of IO; on the basis of this observation, we place an upper limit
of 1 × 10-18 cm3 molecule-1 s-1 for the reaction of I + N2O
to produce IO. However, we did observe a reaction between
IO and OClO
IO f loss
(18)
IO + OClO f products
(23)
and the IO temporal profile can be described by the expression
The first-order rate constant for removal of IO in an excess of
OClO yielded k23 ) (1.0 ( 0.5) × 10-13 cm3 molecule-1 s-1
in 5 Torr of N2 at 298 K. Because this reaction is slow, we
could still use low concentrations of [OClO] (<6 × 1013
molecule cm-3) to generate ClO and study reaction 5.
k12a[CF3I][O(3P)]0
18t
(e-k′t - e-k
)
(IV)
(19)
a
[IO]t )
k18 - k′
a
where ka′) k12[CF3I] + k19. Reaction 19
O(3P) f loss
Cl2O absorbs strongly at 193 nm, σ(193) ) 1.3 × 10-18 cm2
molecule-1 20
, and the yield of Cl atoms has been measured to

5522 J. Phys. Chem. A, Vol. 101, No. 30, 1997
Turnipseed et al.
be ∼2;21 therefore, the primary photodissociation channel for
to changes in [Cl2O], [CF3I], and [N2O] indicate that IO
loss was due solely to reaction with ClO and transport out of
the reaction zone. These measurements were carried out at
various temperatures between 200 and 360 K. In a few
experiments at room temperature, k′b, the appearance rate
coefficient, was measured at various concentrations of ClO while
holding the [CF3I] constant. From this appearance rate coef-
ficient we obtained k25(298 K) ) (4.5 ( 0.7) × 10-11 cm3
Cl2O at 193 nm is
Cl2O + hν f 2Cl + O(3P)
(24)
Therefore, O atoms produced by photolyzing Cl2O {(5-10) ×
1013 molecule cm-3} were comparable to that from N2O
photolysis (σN O ) 9 × 10-20 cm2 molecule-1).3 As the Cl2O
2
molecule-1 s-1
.
concentration was increased, [N2O] was decreased so that the
maximum IO signal was equal to or less than that observed in
the absence of Cl2O. In many cases, the N2O was completely
turned off and Cl2O was used as the only source of O(3P). The
Cl atoms produced by photolysis were rapidly consumed by
Cl2O to form ClO (reaction 6).
The measured values of k5 are summarized in Table 1 and
shown as a function of temperature in Figure 5. The Arrhenius
expression that describes the temperature dependence is k5(T)
) (5.1 ( 1.3) × 10-12 exp[(280 ( 60)/T] cm3 molecule-1 s-1
,
where the uncertainty represents 2σ precision from the Arrhenius
plot. The estimation of the systematic error is described in the
Discussion section.
When IO was produced via reaction 12 in an excess of ClO,
its temporal profile exhibited an initial exponential rise followed
by an exponential decay (see Figure 3) due to the following
reactions:
IO + ClO Reaction Rate Coefficient in Excess O3.
Reaction 5 has many possible products, some of them leading
to the formation of I atoms:
O(3P) + CF3I f IO + CF3
f products
(12a)
(12b)
(25)
IO + ClO f I + ClOO
f I + OClO
(5a)
(5b)
(5c)
O(3P) + ClO f O2 + Cl
f ICl + O2
O(3P) f loss
IO + ClO f products
IO f loss
(19)
(5)
f OIO + Cl
f IOO + Cl
f I + Cl + O2
(5d)
(5e)
(5f)
(18)
Note that any ClO lost due to reaction 25 will give ClO back
rapidly via Cl reaction with the excess Cl2O. This minimizes
contributions to the IO loss rate due to Cl reacting with IO.
The IO temporal profile in this mechanism is described by the
expression:
9
M8 IO2Cl
(5g)
The reaction of I with ClO to produce IO is endothermic by
∼8 kcal mol-1 and, hence, too slow to be important in this
system. I atom production can be estimated in our experiments,
albeit indirectly, by conversion of it back to IO via reaction 1.
Under conditions where k1[O3] ∼ k5[ClO], I atoms produced in
reaction 5 regenerated IO at later times by reaction 1. ClO was
generated by the Cl + O3 reaction, and IO was generated by
reaction 12 subsequent to photolytic production of O atoms from
O3. When the mixture of O3 {(1-10) × 1014 molecule cm-3},
ClO {(1-8) × 1013 molecule cm-3}, and CF3I (8 × 1014
molecule cm-3) was photolyzed (0.3 mJ pulse-1 cm-2), the IO
temporal profile showed a fast rise and then an exponential
decay. However, at longer reaction times (>1 ms), reformation
of IO was apparent (see Figure 6). The initial slope increased
with [ClO], and the extent of IO regeneration increased with
the O3 concentration, as expected. The measured temporal
profile of IO was fit using FACSIMILE, a nonlinear least-
squares fitting routine,22 to the mechanism described in Table
2. A total of 30 profiles were measured with varying [ClO]
and [O3] ([O3]/[ClO] ) 4-40). The initial concentration of
IO was calculated from the photolysis fluence, [O3], the
absorption cross section for O3 at 193 nm (σ ) 4.3 × 10-19
cm2 molecule-1), and the net quantum yield for O(3P) in N2
from O3 photolysis at 193 nm.9 These fits yielded a total rate
coefficient of k5 ) k5I + k5nonI ) (1.37 ( 0.26) × 10-11 cm3
molecule-1 s-1, in good agreement with our directly measured
value (see Table 2 for definition of k5I and k5nonI). A rate
constant for the channels which produce I atoms of k5I ) (1.10
( 0.18) × 10-11 cm3 molecule-1 s-1 was also obtained, which
results in a yield of Φ(I) ) 0.8 ( 0.2 for I atoms.
k12a[CF3I][O(3P)]0
b
d
[IO]t )
(e-k′t - e-k′t
)
(V)
k′ - k′
d
b
where k′ ) k12[CF3I] + k25[ClO] + k19 and k′ ) k5[ClO] +
b
d
k18. The IO temporal profiles were fit by a nonlinear least-
squares fitting routine to the equation [IO]t ) A{exp(-k′) -
b
exp(-kd′t)} to determine both k′ and k′. IO formation was
b
d
always ∼10 times faster than its decay due to reaction 5. Once
the production of IO was complete, its temporal profile was
governed by the equation
[IO]t ) [IO]0 exp(-k′t)
(VI)
d
Therefore, only the decay of IO was recorded in most of the
measurements. When only the decay of IO was analyzed using
eq VI, the kd′ obtained was within a few percent of that
determined by analyzing the temporal profile using eq V.
Reactions of CF3 radicals could be problematic if they resulted
in regeneration of IO. However, CF3 most likely reacts with
Cl2 or Cl2O in our system and, as shown by varying [CF3I] and
photolysis fluence, does not affect the measured IO temporal
profiles. The measured k′ values were unaffected (<10%) by
d
changes in [Cl2O], or [OClO], [N2O], [CF3I] and photolysis
fluence over factors of 5 and were exponential over at least
two 1/e lifetimes.
The temporal profiles of IO were measured at various [ClO],
and k′ versus [ClO] yielded the rate coefficient, k5, as the slope
d
(Figure 4a). Plots of k′ versus [ClO] were linear over an order
At higher [O3] all of the I atoms produced can be recycled
rapidly relative to IO loss from reaction 5. Under these
conditions decays of IO were exponential and varied linearly
d
of magnitude in [ClO] with intercepts equal to the measured
value of k18. This observation, as well as the invariance of k′
d

Kinetics of the IO Radical. 1
J. Phys. Chem. A, Vol. 101, No. 30, 1997 5523
Figure 5. Variation of the measured rate coefficient, k5, with
temperature. Symbols are Cl/Cl2O and O/CF3I radical sources at 5 Torr
(b) and 15 Torr (3). Cl/OClO and O/CF3I radical sources (4), Cl/
Cl2O and O/I2 radical sources (0). The line is an Arrhenius fit yielding
k5(T) ) (5.1 ( 1.7) × 10-12 exp[(280 ( 80)/T] cm3 molecule-1 s-1
.
The uncertainty includes 2σ uncertainty from the fit and 18% estimated
systematic error.
Figure 4. (a) Plot of k′ vs [ClO] at 298 K. Symbols indicate different
d
experimental conditions. Using Cl/Cl2O and O/CF3I radical sources,
O ≡ 5 Torr of N2; 0 ≡ 4 Torr of He/1 Torr of N2; ] ≡ 15 Torr of N2.
Using Cl/OClO and O/CF3I radical sources, 4; and using Cl/Cl2O and
O/I2 radical sources, 3 in 5 Torr of N2. The various experiments are
normalized to an average value for k18 ) 220 s-1. The line is the linear
least-squares fit to the data. (b) A plot of k′d vs [ClO] at 298 K ([) in
a large excess of O3 ([O3] ) 4 × 1015 molecule cm-3). The line is the
least-squares fit to the data.
TABLE 1: Summary of the Rate Coefficients Measured for
the IO + ClO Reaction
Figure 6. Temporal profile of IO in an excess of O3 ([O3] ) 7.5 ×
1014 molecule cm-3) and ClO ([ClO] ) 9.2 × 1013 molecule cm-3).
The line is the fit to the data using the mechanism described in Table
2 and gives a rate coefficient of k5 ) 1.4 × 10-11 cm3 molecule-1 s-1
and a yield of 0.8 for I atom production.
Ptot
[ClO] (1013
(k5 ( 2σ)a,b (10-11
T (K) (Torr)
Leff (cm)
molecule cm-3
)
cm3 molecule-1 s-1
)
362
326
298
5.3 36.5 ( 0.2
5.3 35.4 ( 0.2
5.4 32.6 ( 0.4
5.1 32.6 ( 0.4
5.1 34.1 ( 0.3
5.3 34.1 ( 0.3
15.3 34.1 ( 0.3
1.1-6.4
0.6-7.6
1.4-8.8
2.6-9.6
1.4-8.2
1.0-8.9
1.1-5.9
1.03 ( 0.08
1.26 ( 0.05
1.17 ( 0.08c
1.16 ( 0.07d
1.47 ( 0.07e
1.20 ( 0.15
1.43 ( 0.18e
avg at 298 K )
1.29 ( 0.27
1.47 ( 0.10
1.64 ( 0.12
1.84 ( 0.13
1.94 ( 0.13
1.91 ( 0.17
1.94 ( 0.12
2.00 ( 0.15
TABLE 2: Reaction Mechanism Used To Determine the
Yield of I Atoms from Reaction 5
reaction
k(298 K)a
k5I
notes
IO + ClO f I + products
IO + ClO f non-I products
I + O3 f IO + O2
Cl + O3 f ClO + O2
IO f loss
(5)
(5)
(1)
(8)
(18)
(27)
varied in fit
varied in fit
k5nonI
269
249
243
226
213
210
200
5.2 32.7 ( 0.5
5.1 32.4 ( 0.6
5.3 31.0 ( 0.3
5.2 30.1 ( 0.5
5.1 30.2 ( 0.4
16.0 30.2 ( 0.4
5.3 29.1 ( 0.7
0.7-8.4
0.5-7.0
0.7-6.8
0.4-5.3
0.7-6.7
1.2-5.8
0.5-7.3
1.2 × 10-12
1.2 × 10-11
150
b
c
measured
I f loss
150
d
a Units of cm3 molecule-1 s-1 for second-order reactions and s-1
for first-order reactions. b From ref 9. c From ref 10. d Assumed to be
the same as for IO.
a Uncertainties are from the precision of the kd′ vs [ClO] fit. b Unless
otherwise noted, measurements were made using ≈4 Torr of He and 1
Torr of N2 and using O + CF3I to produce IO and Cl + Cl2O to produce
ClO. c Using O + I2 reaction as the IO source in N2. d Using Cl +
OClO as the ClO source in N2. e In N2.
of kd′ vs [ClO] from this experiment is shown in Figure 4, curve
b, and yields a rate coefficient for k5nonI ) (1.88 ( 0.36) ×
10-12 cm3 molecule-1 s-1. Combining this with the total rate
coefficient, k5, (Figure 4, line a), gives a branching ratio of Φ
) 0.14 ( 0.04 at 298 K for the sum of channels that do not
produce I atoms.
These measurements were also attempted at 223 K. However,
due to the slower rate coefficient for the I + O3 reaction, we
were unable to attain conditions where k1[O3] > 10k5[ClO] and
with [ClO], and the IO temporal profile was described by eq
VI where k′ ) (k5nonI[ClO] + k18). ClO and IO were produced
d
via reactions 8 and 12 as in the previous experiments. The
photolysis fluence was maintained at a low level (0.02 mJ
pulse-1 cm-2) to avoid producing too much IO at these high
concentrations of O3 ([O3] ) 4 × 1015 molecule cm-3). A plot

5524 J. Phys. Chem. A, Vol. 101, No. 30, 1997
Turnipseed et al.
measure [ClO] and changes in the IO loss. Therefore, we had
to rely on modeling the measured temporal profile of IO to
obtain information on the yield of I atoms. From 10 temporal
profiles ([O3] ) 1 × 1015, [ClO] ) (1-8) × 1013 molecule
cm-3), we obtained a yield of Φ(I) ) 0.78 ( 0.25 for I atom
production from FACSIMILE simulations. Although the un-
certainties are larger than at 298 K, it appears that the I atom
yield does not change significantly with temperature.
Other sources of systematic error include the spectral subtrac-
tions and the absorption cross section for ClO at 253.7 nm.
Spectral subtractions of ClO from the excess Cl2O were
reproducible at the (2% level, and the combined uncertainty
in the absorption cross section and the effective path length for
the [ClO] measurements is estimated to be <(10%. Another
small source of systematic error involves the measurement of
the reaction zone temperature (2%, which gives an uncertainty
of (1% in [ClO]. Therefore, our total systematic error in k5 is
(18%. On the basis of these values we quote k5 ) (5.1 ( 1.7)
× 10-12 exp[(280 ( 80)/T].
Discussion
The measured values of k5 are listed in Table 1. These values
were insensitive to variation of [N2O], [CF3I], [Cl2O], [OClO]
and varying the photolysis fluence over a factor of 5. The 298
K value of k5 was the same using different chemical sources
for both IO and ClO. These observations strongly indicate that
IO was reacting solely with ClO and its decay was not
influenced by reactions with other precursors or radicals. If
reaction 5 produced I*, electronically excited I atoms, its
reaction with ClO is likely to produce IO and is likely to be
rapid. However, on the basis of known thermochemistry,
reaction 5 cannot lead to I*.
The accuracy of our rate coefficients are critically dependent
upon the accuracy of the ClO concentration in the volume where
IO temporal profiles are measured. The major contributions to
this uncertainty are the possible loss of ClO along the length of
the reactor and inaccuracies in the optical measurements. We
measure the ClO column in the absorption cell while the IO
temporal profile is controlled by ClO concentration in the
reaction zone. As discussed above, under all experimental
conditions, losses due to the ClO self-reaction are calculated to
be <20% and computer simulations of the ClO loss along the
absorption path length indicated that the measured average [ClO]
is representative of the [ClO] in the reaction zone to within
(5% even at the highest [ClO].
Several experimental tests were carried out to verify the above
calculations of [ClO]: (1) We added absorption cells (with an
glass insert to cover the O-rings) to measure [ClO] upstream
and downstream of the photolysis cell. The value of k5
determined at 298 K when [ClO] was measured using a 50 cm
absorption cell upstream of the LIF reactor was k5 ) (1.1 (
0.1) × 10-11 cm3 molecule-1 s-1. When the absorption cell
was placed downstream of the photolysis cell, we measured k5
) (1.7 ( 0.2) × 10-11 cm3 molecule-1 s-1. These values
bracket the rate coefficient obtained by measuring [ClO] directly
through the LIF reactor. The loss of ClO necessary to reproduce
these numbers is consistent with the ClO + ClO reaction being
the main loss process. (2) The rate coefficient for the O +
ClO reaction was measured by following the IO rise to be k25
) (4.5 ( 0.7) × 10-11 cm3 molecule-1 s-1, which is within
15% of the recommended value.9 (3) We used OClO as the
ClO precursor, with OClO in excess. Since each Cl atom,
produced via the ClO self-reaction, reacts with OClO to generate
2ClO, the removal of ClO via reactions 9a, 9b, and 9c
collectively replenish the ClO that was lost. This process
effectively eliminates depletion of ClO via reaction 9. The rate
coefficient measured using the OClO source agreed with that
measured using the Cl2O source. (4) Since reaction 9 is second-
order, the loss of ClO should increase as [ClO]2 at larger
concentrations. If this loss was significant, the fractional loss
of ClO should be larger and display curvature in the kd′ vs [ClO]
plots. As mentioned earlier, the plots of kd′ vs [ClO] were linear
in all experiments. From these observations and the computer
simulations, we estimate that the uncertainty in the measured
[ClO] within the reaction zone due to losses was ≈(5% via
the self-reaction and the first-order wall loss.
We can place the measured value of k5 in the context of other
analogous reactions of halogen monoxides.9,23 The IO + ClO
reaction exhibits a weak negative temperature dependence (E/R
) -280 K-1). The reactions of halogen oxides with other
halogen oxides are thought to occur via the formation of an
intermediate complex (XOOY or OXOY), which can then be
stabilized or decompose to form products. The competition
between decomposition of the complex back to reactants or to
products or stabilization leads to the observed increase in the
rate coefficient at lower temperatures.
The room temperature rate coefficient and activation energy
for reaction 5 are remarkably similar to that for the reaction
between ClO and BrO,24-28 k27(298) ) 1.2 × 10-11 cm3
molecule-1 s-1, E/R ≈ -300 K.
BrO + ClO f Br + ClOO
f Br + OClO
Φ ≈ 0.4 (27a)
Φ ≈ 0.5 (27b)
Φ ≈ 0.1 (27c)
f BrCl + O2
The BrO + ClO reaction is known to proceed through three
different channels,25-28 whose branching ratios change with
temperature. The IO + ClO reaction can also proceed through
analogous pathways. Reaction 5a, 5b, and 5c are exothermic
by 4.6, 5.3, and 49.1 kcal mol-1, respectively at 298 K, using
a ∆fH°(298 K)(IO) < 28.8 kcal mol-1 19
Reaction 5f is slightly
.
endothermic but highly favored entropically. The thermochem-
istry of channels 5d, 5e, and 5g are unknown due to the lack of
thermochemical data on OIO, IOO, and IO2Cl. ClOO is known
to be weakly bound (∼5 kcal mol-1 9
)
and will rapidly
decompose to Cl and O2. On the basis of trends in the halogen
species, IOO is predicted to be less stable than ClOO;29,30
therefore, channels 5a and 5e effectively result in the production
of atomic chlorine and iodine (channel 5f).
Experiments in excess O3 indicated that I atoms were a major
product in this reaction. At 298 K and 7 Torr of N2, a yield of
0.14 ( 0.04 was measured for the sum of channels that do not
produce I atoms (Φ(nonI) ) {k5c + k5d + k5g}/k5). By
comparison with other halogen monoxide reactions, channel 5c
is the most probable pathway.9 The analogous BrCl formation
from reaction 27 has been observed25-27 to have a yield of
≈10%. It is possible that electronically excited ICl could be
formed and dissociate at the low pressures utilized here. At
higher pressures, quenching of ICl* could compete with
dissociation and, possibly, increase the branching ratio for this
channel. Toohey et al.24 have suggested a similar mechanism
in the BrO + ClO reaction to explain their observed decreases
in the branching ratio of Br with increasing pressure.
A
complete pressure dependence of the branching ratio and/or
detection of ICl is necessary to better understand the mechanism
of the IO + ClO reaction.
Formation of OIO, which has been recently observed by flash
photolysis of an I2/O3 mixture,31 cannot be ruled out, but a recent
estimation of the enthalpy of formation of this radical (∆fH )

Kinetics of the IO Radical. 1
40 kcal mol-1 32
J. Phys. Chem. A, Vol. 101, No. 30, 1997 5525
(2) Klick, S.; Abrahamsson, K. J. Geophys. Res. 1992, 97, 12683.
(3) Moyers, J. L.; Duce, R. A. J. Geophys. Res. 1972, 77, 5229.
(4) Zafiriou, O. C. J. Geophys. Res. 1974, 79, 2730.
)
suggests that this channel would be ≈13 kcal
mol-1 endothermic and, thus, unimportant. The lack of any
pressure dependence in the rate coefficient also suggests that
stabilization of IClO2 is not important in our system. However,
all of our measurements were done at <15 Torr; therefore, this
species cannot be completely ruled out at higher pressures.
Mauldin et al.30 observed an unknown product in the BrO +
BrO reaction at 223 K and 300 Torr N2 and tentatively identified
it as Br2O2. Similar observations have been reported suggesting
that formation of IClO2 may also be possible.33
(5) Chameides, W. L.; Davis, D. D. J. Geophys. Res. 1980, 85, 7383.
(6) Barrie, L. A.; Bottenheim, J. W.; Hart, W. R. J. Geophys. Res.
1994, 99, 25313.
(7) The Tropospheric Chemistry of Ozone in the Polar Regions,
Proceedings of the NATO Advanced Research Workshop; Niki, H., Becker,
H. B., Eds.; Springer-Verlag: New York, 1993.
(8) Solomon, S.; Garcia, R. R.; Ravishankara, A. R. J. Geophys. Res.
1994, 99, 20491.
(9) 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, EValuation No. 11; Jet Propulsion Laboratory: Pasadena, CA,
1994.
(10) Turnipseed, A. A.; Gilles, M. K.; Burkholder, J. B.; Ravishankara,
A. R. Chem. Phys. Lett. 1995, 242, 427.
(11) Trolier, M.; Mauldin, R. L., III; Ravishankara, A. R. J. Phys. Chem.
1990, 94, 4896.
(12) Molina, L. T.; Molina, M. J. J. Geophys. Res. 1986, 91, 14501.
Our experiments cannot distinguish between channels 5b and
the channels that produce atomic I and Cl. In the atmosphere,
the production of I + OClO in reaction 5 will not lead to a net
ozone loss. We attempted to monitor OClO by LIF; however,
our detection sensitivity was an order of magnitude too poor to
observe the concentrations of OClO that could be produced
under our experimental conditions.
Atmospheric Implications. The modeling study of Solomon
et al.8 assumed k5 ) 1 × 10-10 cm3 molecule-1 s-1. Using
this rate coefficient, a total iodine abundance of 1 pptv in the
stratosphere was necessary to completely account for the
observed trend in low-altitude, midlatitude O3 loss during the
past two decades. Our measured value at temperatures char-
acteristic of the stratosphere is approximately a factor of 5 lower
than the assumed value. Thus, the interaction of iodine with
chlorine will be five times less efficient at destroying ozone
than predicted if we assume reaction 5 does not produce OClO.
The contribution of this reaction to stratospheric ozone loss is
discussed in detail in the next paper.
(13) Solomon, S.; Burkholder, J. B.; Ravishankara, A. R.; Garcia, R.
R. J. Geophys. Res. 1994, 99, 20929.
(14) Calvert, J. G.; Pitts, J. N., Jr. Photochemistry; Wiley: New York,
1966.
(15) Renard, J. J.; Bolker, H. I. Chem. ReV. 1976, 76, 487.
(16) Bedjanian, Y.; LeBras, G.; Poulet, G. J. Phys. Chem. 1996, 100,
15130.
(17) Kaiser, E. W.; Wallington, T. J.; Hurley, M. D. Int. J. Chem. Kinet.
1995, 27, 205.
(18) Burkholder, J. B.; Mauldin, R. L., III; Yokelson, R. J.; Ravishankara,
A. R. J. Phys. Chem. 1993, 97, 7597.
(19) Gilles, M. K.; Turnipseed, A. A.; Talukdar, R. K.; Rudich, Y.;
Villalta, P. W.; Huey, L. G.; Burkholder, J. B.; Ravishankara, A. R. J. Phys.
Chem. 1996, 100, 14005.
Note Added in Proof. Yuri Bedjanian, Georges Le Bras
and Gilles Poulet have measured the 298 K rate coefficient k5-
(298 K) ) (1.1 ( 0.2) × 10-11 cm3 molecule-1 s-1 using a
mass spectrometric discharge flow system (J. Phys. Chem. A
1997, 101, 4088). They also report branching ratios of (0.55
( 0.03), (0.25 ( 0.02), and (0.20 ( 0.02) for channels 5b, 5f,
and 5c, respectively. These results are in excellent agreement
with those reported in this paper.
(20) Lin, C. L. J. Chem. Eng. Data 1976, 21, 411.
(21) Gilles, M. K.; Ravishankara, A. R. Unpublished data.
(22) Malleson, A. M.; Kellett, H. M.; Myhill, R. G.; Sweetenham, W.
P. FACSIMILE; A. E. R. E. Harwell Publications Office: Oxfordshire, 1990.
(23) Sander, S. P. J. Phys. Chem. 1986, 90, 2194.
(24) Toohey, D. W.; Anderson, J. G. J. Phys. Chem. 1988, 92, 1705.
(25) Turnipseed, A. A.; Birks, J. W.; Calvert, J. G. J. Phys. Chem. 1991,
95, 4356.
(26) Friedl, R. R.; Sander, S. P. J. Phys. Chem. 1989, 93, 4756.
(27) Poulet, G.; Lancar, I. T.; Laverdet, G.; LeBras, G. J. Phys. Chem.
1990, 94, 278.
(28) Sander, S. P.; Friedl, R. R. J. Phys. Chem. 1989, 93, 4764.
(29) Blake, J. A.; Browne, R. J.; Burns, G. J. J. Chem. Phys. 1970, 53,
3320.
(30) Mauldin, R. L., III; Wahner, A.; Ravishankara, A. R. J. Phys. Chem.
1993, 97, 7585.
(31) Himmelmann, S.; Orphal, J.; Bovensmann, H.; Richter, A.; Lad-
staetter-Weissenmayer, A.; Burrows, J. P. Chem. Phys. Lett. 1996, 251,
330.
Acknowledgment. The authors thank Dr. S. Solomon for
useful discussions concerning this work. We also acknowledge
Dr. Bedjanian for both sending us preprints prior to publication
and his insightful comments. This work was funded in part by
NASA’s Mission to Planet Earth Program through the Upper
Atmospheric Research Program.
References and Notes
(32) Chase, M. W. J. Phys. Chem. Ref. Data 1996, 25, 1297.
(1) Reifenhauser, W.; Heumann, K. G. Atmos. EnViron. 1992, 26A,
2905.
(33) Johnsson, K.; Engdahl, A.; Koelm, J.; Nieminen, J.; Nelander, B.
J. Phys. Chem. 1995, 99, 3902.