Kinetics and products of the IO + BrO reaction

Article abstract of DOI:10.1021/jp004494y

The kinetics and products of the IO + BrO reaction were discussed. The reaction was studied using the technique of laser flash photolysis with kinetic ultraviolet (UV) absorption spectroscopy. No direct measurement of I or Br was performed. IBr was observed as a minor product (k_lb/k_l < 0.2). No evidence of OBrO formation was observed (k_ld/k_l < 0.15). The results were compared with previous studies of the IO + BrO reaction, and the atmospheric implications were briefly discussed.

Full text of DOI:10.1021/jp004494y

J. Phys. Chem. A 2001, 105, 7855-7864
7855
Kinetics and Products of the IO + BrO Reaction
David M. Rowley,† William J. Bloss,‡ R. Anthony Cox,* and Roderic L. Jones
Centre for Atmospheric Science, UniVersity Chemical Laboratories, UniVersity of Cambridge, Lensfield Road,
Cambridge, CB2 1EW, U.K.
ReceiVed: December 14, 2000; In Final Form: April 12, 2001
The kinetics and products of reaction 1 between gas-phase BrO and IO radicals have been studied using the
technique of laser photolysis with time-resolved UV-vis absorption spectroscopy. The O + IBr reaction 8,
used as one source of IO and BrO radicals, was found to produce predominantly IO radicals at 295 K. The
rate coefficient for reaction 8 is correlated with the branching ratio for IO production (channel 8a). Using a
value of 0.7 for the branching ratio for IO production (k8a/k8), k8 was found to be (3.6 ( 2.4) × 10-11 cm3
molecule-1 s-1 at 295 K, with no significant temperature dependence between 210 and 333 K. Sensitivity
tests using a kinetic model showed that, in addition to (1), decay traces were also sensitive to the rate coefficient
for reaction 12: I + BrO f Br + IO. This rate coefficient was found to be (1.3 ( 1.2) × 10-11 molecule-1
cm3 s-1 at 295 K, with no significant temperature dependence between 235 and 333 K. The title reaction (1),
IO + BrO f products was found to have a rate coefficient (8.5 ( 1.4) × 10-11cm3 molecule-1 s-1 at 295 K.
Reaction 1 exhibited a negative temperature dependence between 210 and 333 K, adequately described by k1
) (6.7 ( 0.8) × 10-12 exp ((760 ( 30)/T) molecules-1 cm3 s-1. No pressure dependence to k1 was found
between 100 and 760 Torr. All errors are 2σ. Five potential products exist for the IO + BrO reaction: IO +
BrO f I + Br + O2 (1a), IO + BrO f IBr + O2 (1b), IO + BrO f OIO + Br (1c), IO + BrO f OBrO
+ I (1d), and IO + BrO f IBrO2 (1e). No direct measurement of I or Br formation was perfomed. IBr was
observed as a minor product (k1b/k1 < 0.2). OIO formation was observed and shown to result from the IO +
BrO reaction for the first time in this work. No evidence for OBrO formation was observed (k1d/k1 < 0.15).
No evidence for the formation or existence of IBrO2 was observed. The results obtained here, in conjunction
with other published work, were used to constrain the branching ratio, R, for OIO production in the IO + IO
reaction, giving {0.30 ( 0.05 e R e 0.46 ( 0.08}. This constraint allowed the absorption cross section of
OIO to be constrained giving (1.29 ( 0.22) g σOIO × 1017 g 0.87 ( 0.15) molecules-1 cm2 at the (5,1,0)
peak at 549 nm, 295 K, and 760 Torr. Results are compared with previous studies of the IO + BrO reaction,
and the atmospheric implications are briefly discussed.
1. Introduction
abundance of BrO radicals is in general higher than that of IO
radicals,7 the IO + BrO reaction may occur significantly even
Gas-phase halogen monoxide free radicals can participate in
reaction cycles which destroy atmospheric ozone. The origins
and role of iodine containing species have been discussed in
the accompanying paper1 in which a study of the IO + IO
reaction is described. In this paper, the coupling between iodine
and bromine chemistry is considered in a study of the IO +
BrO reaction.
In the low stratosphere, typical BrO abundances of over 10
of parts per trillion (pptv) are routinely observed.2,3 In contrast,
measurements aimed at detecting stratospheric IO initially
reported an upper limit of 0.2 pptv for the IO abundance4,5
although a recently reported measurement has indicated strato-
spheric IO abundances in the range 0.65-0.80 pptv.6 At such
concentrations the IO self-reaction makes a negligible contribu-
tion to ozone depletion; however, coupling between iodine
species and members of other chemical families can lead to
significant iodine-catalyzed ozone loss. The IO + BrO reaction
is one such potential coupling reaction. Since the stratospheric
at low IO concentrations; consequently, if the IO + BrO reaction
produces I or Br atoms, ozone loss can occur even if IO
concentrations are too low for the IO self-reaction to proceed
rapidly.
The potential for very low concentrations of total iodine to
destroy ozone in the stratosphere was originally described by
Solomon et al.8 In a modeling study, these authors found that
1 pptv of total iodine in the low stratosphere would be
partitioned primarily into active (ozone destroying) forms I and
IO. Moreover, Solomon et al. showed that, at a level of 1 pptv,
iodine catalyzed ozone loss would dominate halogen catalyzed
ozone loss in the low stratosphere. In carrying out this modeling
study, Solomon et al. assumed that the IO + XO (X ) Cl, Br)
reactions both proceeded with rate coefficients of 1 × 10-10
molecule-1 cm3 s-1, and produced solely I + X + O2 products.
Occurrence of the IO + BrO reaction would thus lead to ozone
loss via
IO + BrO f I + Br + O2
I + O3 f IO + O2
(1a)
(2)
* To whom correspondence should be addressed. E-mail: rac26@
cam.ac.uk. FAX: + 44 1223 336362.
† Present address: Jet Propulsion Laboratory, California Institute of
Technology, Pasadena, CA 91109. E-mail: wbloss@jpl.nasa.gov.
‡ Present address: University College London, Chemistry Department,
Christopher Ingold Laboratories, 20 Gordon Street, London WC1H 0AJ,
UK. E-mail: d.m.rowley@ucl.ac.uk.
Br + O3 f BrO + O2
(3)
Net
2O3 f 3O2
10.1021/jp004494y CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/27/2001

7856 J. Phys. Chem. A, Vol. 105, No. 33, 2001
Rowley et al.
TABLE 1: Enthalpies of Formation of IO + BrO Reaction
the temperature dependence of reaction 1, using pulsed laser
photolysis within a discharge flow system to generate the
reactants, coupled with LIF detection of IO and UV absorption
detection of BrO. Gilles et al. monitored the loss of IO radicals
in the presence of excess O3, and thus obtained the rate
coefficient for the non-I atom producing channels of reaction
1, as channels producing an I atom led to IO regeneration via
reaction 2 and were therefore effectively suppressed:
Gilles et al. obtained a value of (6.05 ( 0.57) × 10-11
molecules-1 cm3 s-1 for (k1b + k1c + k1e) at 298 K and total
pressures in the range 7-15 Torr. This result, in conjunction
with the measurement of k1 reported by Laszlo et al., indicates
that non-I atom producing channels dominate the IO + BrO
reaction at 298 K. Gilles et al. also found that non-I atom
Moieties
species
∆fH° (298K)/kJ mol-1
ref
IO
BrO
I
116 ( 5
120 ( 6
107
41
38
Br
112
40.9
76.7
150
42
IBr
OIO
OBrO
17
18
The IO + BrO reaction may also be of significance in the
troposphere. IO and BrO radicals have both been observed in
the marine boundary layer9,10 and, while no simultaneous
observation of IO and BrO has been reported, precursors to both
species, such as the photolabile iodobromocarbon CH2IBr, have
been observed in the marine boundary layer.11 The observation
of elevated BrO concentrations correlates with reduced ozone
concentrations in the marine boundary layer.12 This can be
explained in part by the occurrence of the relatively slow BrO
+ BrO reaction 4a; however, reaction 1a could also contribute
to the observed ozone loss, even with low IO abundances, if
k1a were significantly greater than k4a.
channels exhibited a negative temperature dependence, (k1b
+
k1c + k1e) increasing to (8.82 ( 0.50) × 10-11 molecules-1 cm3
s-1 at 204 K. Bedjanian et al.21 measured the overall rate
coefficient k1 and placed limits on branching ratios for the IO
+ BrO reaction at 298 K and 1 Torr total pressure, using the
discharge flow/mass spectrometry technique. These authors
report k1 ) 8.5 × 10-11 molecules-1 cm3 s-1, and that I atom
producing channels have a combined branching ratio ((k1a
+
k1d)/k1) < 0.3. Bedjanian et al. also report that OIO production
dominated the IO + BrO reaction, with a branching ratio of
0.65 e k1c/k1 e 1.
BrO + BrO f 2 Br + O2
(4a)
(k4a,295K ) 2.49 × 10-12 molecule-1 cm3 s-1 (ref 13))
Two of the potential products of reaction 1 have recently been
observed in the atmosphere. Tentative observations of atmo-
spheric OIO have recently been reported.22 However, our
laboratory studies have shown that OIO is a product of the IO
self-reaction1 and measurements of atmospheric IO at the several
pptv level which accompanied the OIO observations indicate
that this reaction could be the source of the observed OIO.
Measurements of OBrO in the nighttime stratosphere at
concentrations of up to 20 pptv have also been reported,23
although very recent measurements do not confirm these
observations.24 In the laboratory, OBrO formation has been
observed and attributed to the reaction between between BrO
and O3,13,25 and also as a product of the self-reaction of
vibrationally excited BrO.26 However, none of the currently
understood bromine chemistry can explain the presence of
significant amounts (>0.01 pptv) of OBrO in the stratosphere.27
Assessment of the atmospheric significance of the IO + BrO
reaction throughout the atmosphere, and in particular the
potential for iodine chemistry to contribute to stratospheric ozone
loss, requires both the rate coefficient and branching ratio of
reaction 1 to be established under appropriate conditions. In
this paper, we report an investigation of the kinetics and products
of the IO + BrO reaction as a function of temperature between
210 and 333 K, and as a function of pressure between 100 and
760 Torr.
By analogy with other halogen monoxide self- and cross-
reactions,14 five potential channels can be envisaged for the IO
+ BrO reaction:
IO + BrO f I + Br + O2
f IBr + O2
(1a)
(1b)
f OIO + Br
f OBrO + I
(1c)
(1d)
(1e)
IO + BrO + M f IBrO2 + M
The potential for the IO + BrO reaction to affect atmospheric
ozone levels depends critically upon the reaction products:
Occurrence of reaction channels 1a or 1b followed by IBr
photolysis, will lead to catalytic ozone destruction, via reactions
2 and 3. The potential impact of channels 1c and 1d on ozone
depends on the fate of the OXO species (X ) I, Br) produced:
if OXO photolyses to form O(3P) + XO, in a manner analogous
to OClO,15 then channels 1c and 1d will lead to a null cycle
with respect to atmospheric ozone loss. Conversely, if OXO is
photostable, as has been argued for OIO,16 some ozone loss is
expected from the halogen atom coproduced in 1c and 1d, and
the role of the IO + BrO reaction will depend on the subsequent
fate of the OXO species. Similarly, the atmospheric implications
of formation of IBrO2 would depend on the structure and
photochemical properties of this species. Thermodynamic data
(Table 1) indicates that reactions 1a and 1b are exothermic, by
17 and 195 kJ mol-1, respectively. Calculated values for the
enthalpies of formation of OIO17 and OBrO18 indicate that
channel (1c) is exothermic by 47 kJ mol-1, while channel 1d is
endothermic by 21 kJ mol-1. The thermodynamic properties of
the hypothesized IBrO2 molecule are unknown.
2. Experimental Details
The IO + BrO reaction was studied using the technique of
laser flash photolysis with kinetic UV absorption spectroscopy,
utilizing a CCD detection system. The apparatus has been
described in detail elsewhere13,28 and briefly in the accompany-
ing paper;1 no further details are given here.
Radical Generation. Three chemical systems were employed
to generate IO and BrO radicals, using different precursor gas
mixtures: I2/Br2/N2O/N2, IBr/N2O/N2, and CF3I/Br2/N2O/N2,
referred to hereafter as systems A, B, and C, respectively.
Reaction initiation was achieved via 193 nm laser photolysis
of N2O to form O(1D) + N2, followed by rapid collisional
quenching of the excited oxygen atoms to O(3P). IO and BrO
The IO + BrO reaction has been the subject of three previous
laboratory studies. Laszlo et al.19 measured the overall rate
coefficient at 298 K and 760 Torr using the flash photolysis/
UV absorption technique, and found k1 ) (6.9 ( 2.7) × 10-11
molecules-1 cm3 s-1. Subsequently, Gilles et al.20 investigated

IO + BrO Reaction
J. Phys. Chem. A, Vol. 105, No. 33, 2001 7857
TABLE 2: Precursor Concentration Ranges
systema N2O × 10-16 I2 × 10-14 Br2 × 10-15 IBr × 10-14 CF3I × 10-16
A
B
C
0.57-9.9
1.4-4.8
2.0-15
9.8
2.5-24
9.8
7.9-19
3.7-8.7
a See text for details. Concentrations in molecules cm-3. N2 added
to 1 atm total pressure.
radicals were generated by reaction of the O(3P) atoms with
the halogen species:
N2O + hν f O(1D) + N2
O(1D) + N2 f O(3P) + N2
(5)
(6)
system A,
I2 + Br2 T 2IBr
(7)
Figure 1. Post-photolysis absorption spectrum recorded using system
C (Br2/CF3I). The spectrum is a postrelative (It) to prephotolysis (Io)
spectrum, showing changes in absorption bought about by the photolysis
flash and subsequent chemistry. Thus, the negative baseline above 380
nm arises from the consumption of Br2 in the postphotolysis reactions.
O(3P) + IBr f IO + Br
f BrO + I
(8a)
(8b)
O(3P) + Br2 f BrO + Br
(9)
equilibrium between I2 and Br2 was attained was provided by
tests in which the contact time for I2 and Br2 was reduced to
approximately 30% of that used in the subsequent kinetic
experiments, and the concentration of Br2 was reduced by a
factor of 15. Under these conditions, the presence of bromine
reduced the I2 concentration by 85%, indicating that under the
normal experimental conditions, equilibrium between I2, Br2,
and IBr was effectively complete.
system B,
system C,
O(3P) + IBr f IO + Br
f BrO + I
(8a)
(8b)
O(3P) + CF3I f IO + CF3
O(3P) + Br2 f BrO + Br
(10)
(9)
The use of different precursor mixtures facilitated the
investigation of different aspects of the IO + BrO reaction
system: System A was used to investigate the kinetics of the
IO + BrO reaction, system B was used to constrain the
branching ratio for the O + IBr reaction, and system C was
used to investigate the products of reaction 1, and to test
qualitative understanding of the secondary chemistry occurring.
System A was used for kinetic measurements as the I2 and Br2
concentrations could be independently adjusted to attain the
initial [IO]:[BrO] ratio which maximized the sensitivity of the
IO and BrO decay traces to the IO + BrO reaction rate
coefficient; such control was not possible with system B. System
C was not used for kinetic experiments as the presence of CF3
and CF3I, the reactions of which with IO and BrO are not well
characterized, could introduce errors arising from secondary
chemical interactions.
In system A, equilibrium between I2, Br2, and IBr was achieved
prior to photolysis. Prior to equilibration, Br2 concentrations
were approximately 30-fold in excess over I2, thus following
equilibration [I2] was low and reaction 11 made a negligible
contribution to the generation of IO:
O(3P) + I2 f IO + I
(11)
In system B, I2 and Br2 were present in equilibrium with IBr,
thus reactions reactions 9 and 11 also occurred; however, the
relative concentrations of I2, Br2, and IBr were such that reaction
8 dominated IO and BrO production.
N2O (Distillers MG, 99%), N2 (Distillers MG, 99.996%), and
O2 (Distillers MG, 99.996%) were used as supplied. I2 (Breck-
land Scientific, resublimed) and IBr (Lancaster, 98%) were
entrained in a flow of N2 passed through bulbs containing
halogen flakes interspersed with glass wool. Br2 was entrained
in a flow of nitrogen passed through a bromine bubbler, which
was held in an ice bath to ensure a known and convenient vapor
pressure. CF3I (Fluorochem, 99%) was taken directly from a
pressure cylinder. Concentrations of precursor species entering
the reaction cell are given in Table 2. In all cases, precursor
concentrations were such that O atoms reacted rapidly and
exclusively with halogen species. Concentrations of N2 and CF3I
were calculated from their known flow rates, and those of I2,
Br2, and IBr from flow rates and the appropriate vapor pressures.
For system A, at typical I2 and Br2 concentrations prior to
mixing of 3 × 1014 and 1 × 1016 molecules cm-3, respectively,
I2 was calculated to be overwhelmingly (98.9%) converted to
IBr, using the equilibrium constant for reaction 7 of 20.19 at
298 K,29 if iodine and bromine had equilibrated. Evidence that
Radical Monitoring. IO and BrO radicals were monitored
via their characteristic structured UV-vis absorption spectra
arising from the characteristic A2Π r X2Π vibronic transition.
Spectrograph settings (150 g/mm grating and 50 µm entrance
slit width) corresponded to coverage of the wavelength range
315-445 nm at a resolution (fwhm) of 1.65 nm, and permitted
simultaneous detection of IO and BrO. A typical postphotolysis
absorption spectrum, recorded using the CF3I/Br2 system C, is
shown in Figure 1.
Concentrations of IO and BrO radicals were determined by
fitting reference cross sections over the wavelength ranges 400-
445 and 315-360 nm, respectively, using the technique of
differential spectroscopy. The magnitude of the vibronic features
manifest in the absorption spectra of species such as halogen
monoxide radicals is dependent upon the instrumental resolution
used, and the cross sections used to quantify such species must
therefore be recorded at the appropriate resolution, or smoothed
from higher resolution measurements. Reference IO cross
sections were recorded at 1.65 nm fwhm resolution, following

7858 J. Phys. Chem. A, Vol. 105, No. 33, 2001
Rowley et al.
3. Results
Products of the O + IBr Reaction. The branching ratio θ
for IO production in reaction 8, is defined by θ ) k8a/(k8a
+
k8b) and was measured using system B, N2O/IBr/N2. Peak IO
and BrO concentrations observed immediately after photolysis
were found to be in the ratio 83%:17%, or 4.75:1. These values
do not allow direct calculation of the branching ratio for reaction
8, however, as IO and BrO are also formed from the reaction
of O atoms with I2 and Br2, respectively, which were present
in equilibrium with IBr. While the I2, Br2, and IBr concentrations
present in the reaction cell could be calculated, the rate
coefficient for the O + IBr reaction is not known. Thus, the
flux of O atoms reacting with IBr could not be ascertained,
precluding determination of the branching ratio for reaction 8
other than in terms of k8.
In system B, in which I2 and Br2 are present in equal
concentrations in equilibrium with IBr, the observed initial ratio
of [IO]:[BrO], R, relates θ and k8 as shown in eq ii, where K8
is the equilibrium constant for reaction 8:
Figure 2. Differential fit of BrO and IO reference cross sections to a
post-photolysis absorption spectrum, with residuals (offset). Fits
independently quantify [IO] and [BrO] at each time point
(k11 + k θ K )
x
8
8
R )
(ii)
(k + k (1 - θ) K )
x
9
8
8
Using the observed value of 4.75 for R, eq ii indicated that the
value of θ is relatively insensitive to the value of k8, varying
between 0.6 and 0.8 for k8 ) 2 × 10-11 to 1.4 × 10-10
molecules-1 cm3 s-1 (the range encompassed by the O + Br2
and O + I2 reaction rate coefficients, respectively.31) The O +
IBr reaction rate coefficient is expected to lie between the O +
I2 and O + Br2 reaction rate coefficients (as has been observed
for O + BrCl and O + ICl, by (e.g.) Lowenstein et al.32),
indicating a value of ca. 5 × 10-11 molecules-1 cm3 s-1 for k8,
corresponding to a value of 0.7 for θ. This value was used in
the reaction system model used to determine the kinetics of the
IO + BrO reaction. The dependence of the results obtained upon
the value assumed for θ is discussed below.
Kinetics of the IO + BrO Reaction. IO and BrO decay
traces recorded using system A, I2/Br2/N2O/N2, were analyzed
to determine kinetic data using a model of the reaction system
constructed in FACSIMILE.33 The model incorporated simula-
tions of photolysis, flowout, the temporal distortion imparted
to decay traces by CCD oversampling and the chemical reaction
scheme given in Table 3. Within the model, the initial O atom
concentration, [O(1D)]t)0, and the values of k1 (IO + BrO), k12
(I + BrO), and k8 (O + IBr) were optimized by minimizing the
sum of squares of residuals between observed and calculated
IO and BrO decay traces. The sensitivity of the kinetic results
so obtained to the secondary chemical reactions incorporated
in the model, and the choice of kinetic parameters optimized,
is discussed in the following section.
Figure 3. IO and BrO concentration-time profiles (decay traces),
optimized fits and residuals (offset), obtained using system A (I2/Br2/
N2O/N2).
photolysis of CF3I/N2O/N2 mixtures. These cross sections were
calibrated by carrying out parallel experiments performed at
higher (1.13 nm fwhm) resolution, using the IO cross sections
reported at this resolution in the accompanying paper.1 Wave-
length-dependent BrO cross sections have been measured by
Wahner et al.30 Gilles et al.20 have recently remeasured the
differential BrO cross section around 338 nm, and found it to
be approximately 9% larger than that reported by Wahner et
al. Gilles et al. noted that making a corresponding correction
to rate coefficients for the BrO self-reaction measured via UV-
absorption detection of BrO using the cross sections of Wahner
et al. brought such results into agreement with those performed
using other techniques. The BrO cross sections used in this work
were those measured by Wahner et al., increased in magnitude
by 9% in line with the findings of Gilles et al., and smoothed
to the resolution used in this study. A typical differential fit to
the IO and BrO absorption features in a postphotolysis absorp-
tion spectrum is shown in Figure 2.
Reaction System Model: SensitiVity Tests. Sensitivity tests
were performed using the reaction model in order to determine
which of the reactions occurring had a significant impact upon
the form of the IO and BrO decay traces (i.e., which reaction
rate coefficients could be determined from IO and BrO
concentration-time data). Sensitivity tests were then used to
establish the dependence of the values of rate coefficients
determined upon the correct implementation of the secondary
chemical reactions (i.e., the sensitivity of the optimized rate
coefficients to other reactions in the model.)
The CCD was operated at a charge-transfer rate of 5 × 105
Hz, corresponding to recording an absorption spectrum of the
reaction mixture every 2 µs, for a period of 1 ms following
photolysis. Fitting of IO and BrO cross sections to successive
absorption spectra determined IO and BrO decay traces such
as those shown in Figure 3, obtained using system A.
The first test was performed by measuring the deviation
between simulated decay traces generated using the reaction
scheme for system A implemented as shown in Table 3 (control

IO + BrO Reaction
J. Phys. Chem. A, Vol. 105, No. 33, 2001 7859
TABLE 3: Secondary Chemical Reaction Scheme, and
Appropriate Rate Coefficients (for 295 K/760 Torr),
Employed in the Reaction System Model
reaction
no.
295 K rate coeffa/
reaction
cm3 molecule-1 s-1 ref
O(1D) + N2 f O(3P) + N2
O(3P) + IBr f IO + Br
O(3P) + IBr f BrO + I
O(3P) + Br2 f BrO + Br
O(3P) + IO f I + O2
O(3P) + BrO f Br + O2
IO + BrO f products
I + BrO f IO + Br
Br + IO f BrO + I
IO + IO f products
BrO + BrO f products
I + Br2 f IBr + Br
Br + I2 f IBr + I
6
2.61 × 10-11
3.5 × 10-11
1.5 × 10-11
1.4 × 10-11
1.4 × 10-10
4.1 × 10-11
6.9 × 10-11
1.9 × 10-11
5 × 10-11
42
8a
8b
9
31
42
42
1
12
13
14
39
39
1
28
44
38
45
1
8.18 × 10-11
2.98 × 10-12
3.4 × 10-13
1.38 × 10-14
4.6 × 10-11
1.7 × 10-10
2.1 × 10-10
3 × 10-12
15
16
17
Br + IBr f Br2 + I
I + IO f I2O
Figure 5. Sensitivity test 2. The sensitivity of the optimized values of
k1 and k12 to the rate coefficient of selected secondary reactions included
in the reaction model.
I + I2O f I2 + IO
1
Br + BrO + M f Br2O + M
Br + Br2O f Br2 + BrO
I + I f I2
13
46
47
48
2 × 10-10
2.47 × 10-13
Similar tests indicated that the decay traces were relatively
insensitive to the branching ratios for the IO + IO and IO +
BrO reactions. Decay trace sensitivity to k1 (the IO + BrO
reaction) was found to maximize when the initial BrO concen-
tration was approximately three times that of IO.
The second test determined the dependence of the optimized
rate coefficients upon the secondary chemical reactions. The
rate coefficient for each secondary reaction, in turn, was
perturbed by ( 50%, and this perturbed reaction model used to
generate IO and BrO decay traces. The original model (as given
in Table 3) was then used to re-analyze these traces, determining
an optimized value for k1 only. The difference between this
optimized value of k1 and the value used in the perturbed model
(6.9 × 10-11 molecules-1 cm3 s-1, as measured by Laszlo et
al.19) gave an indication of the sensitivity of the optimization
of k1 to each secondary reaction in the model.
Br + Br f Br2
7.6 × 10-14
a Rate coefficients for reactions 1, 8, and 12 were optimized when
fitting to experimental IO and BrO concentration profiles; the values
for k1, k8, and k12 listed above are those used for sensitivity tests. I and
Br atom recombination rate coefficients are reported at 760 Torr of
N2.
The returned value of k1 was found to exhibit moderate
sensitivity to many of the secondary chemical reactions incor-
porated in the model, typically varying by 20% for each 50%
perturbation of a secondary reaction rate coefficient. Noting that
the IO and BrO decay traces exhibited strong dependence upon
both k1 and k12, the second sensitivity test was repeated
optimizing both k1 and k12 to fit the artificial IO and BrO decay
traces. In this case, the sensitivity of k1 to the secondary reactions
was much reduced; typically, k1 varied by less than 5% for each
50% perturbation of a secondary reaction rate coefficient. The
sensitivity of k12 was somewhat greater, varying by up to 20%
for the same degree of perturbation. The results from these dual-
optimization tests are plotted in Figure 5, which shows the
percentage change in the reoptimized k1 and k12 for the 50%
perturbation of each secondary reaction rate coefficient.
The insensitivity of the returned values of k1 and k12 to
secondary chemistry, as determined by the dual-optimization
analysis, was verified by investigating the nature of the error
surface defined by the fits to the IO and BrO decay traces.
Control IO and BrO decay traces were generated using the
nominal values for these rate coefficients, and compared with
those obtained when the values of k1 and k12 were altered by
some factor. The deviation between traces was defined as the
sum of the RMS residuals for each species (IO and BrO) over
the 1 ms post-photolysis time-period. This quantity was
expressed as a function of k1 and k12, varying both rate
coefficients by a factor of 3, and generating a “goodness of fit”
error matrix. This is shown in Table 4, and the presence of a
single minimum value for the residual indicates that a unique
Figure 4. Sensitivity test 1. The sensitivity of the form of the IO and
BrO decay traces to perturbations of the rate coefficient of each reaction
occurring.
decay traces), and those generated when the rate coefficient for
each reaction was perturbed, in turn, by an arbitrary factor of
( 50%. The deviation between perturbed and control decay
traces was evaluated as the RMS residual for each species, IO
and BrO, calculated over the 1 ms postphotolysis time period
corresponding to our experimental conditions. The results of
this test (plotted in Figure 4) show clearly that the IO and BrO
decay traces exhibit similar patterns of sensitivity, displaying
the greatest dependence upon the rate coefficient for the I +
BrO reaction 12 followed by the IO + BrO reaction 1.
Sensitivity to the other reactions taking place was considerably
lower: In particular, the traces displayed very little sensitivity
to the rate coefficient for reaction 13, Br + IO, due to the
conversion of Br atoms to I atoms, predominantly through the
halogen/interhalogen atom-molecule reactions.
I + BrO f IO + Br
Br + IO f BrO + I
(12)
(13)

7860 J. Phys. Chem. A, Vol. 105, No. 33, 2001
Rowley et al.
a
a
TABLE 4: Sensitivity Test for k1 and k12
TABLE 5: Pressure Dependence of k1
factor Change in k1
pressure/Torr
k1/10-11 molecules-1cm3 s-1
factor change in k12
0.333
0.5
1
2
3
100
220
360
480
600
760
8.30 ( 0.98
7.54 ( 1.48
8.14 ( 0.25
8.24 ( 1.30
8.82 ( 0.65
8.50 ( 1.40
3
2
1
0.5
11.4
7.0
4.0
22.2
17.0
11.4
4.8
1.4
16.4
16.0
12.2
5.6
0
8.1
15.0
14.6
8.5
1.7
4.3
13.1
16.9
11.2
6.0
8
10.7
0.3333
a Uncertainties are 2σ.
a Change in the RMS residual on decay traces resulting from
deviations in k1 and k12 values from optimized values.
TABLE 6: Values of k1, k12, and k8 Obtained as a Function
of Temperaturea
IO + BrO
I + BrO
O + IBr
temperature/K
k1
(2σ
k12
(2σ
k8
(2σ
333
296
283
273
260
250
235
222
210
6.81
8.49
0.39
1.40
0.23
0.73
0.49
2.92
4.75
1.27
2.91
1.59
1.30
1.06
1.48
1.00
1.38
1.17
0.63
0.30
0.06
1.27
0.01
0.04
0.05
0.16
0.20
1.27
0.02
4.62
3.57
5.38
7.91
7.48
7.54
14.98b
6.52
4.00
1.08
2.48
0.57
0.67
1.12
5.04
2.98
1.62
0.76
11.03
10.08
13.70
16.82
19.44
18.67
21.17
a Units: molecules-1 cm3 s-1 × 10-11
.
b Excluded from subsequent
analysis.
Kinetic Results Pressure and Temperature Dependence.
Experiments were conducted over a range of pressures, between
100 and 760 Torr. At ambient temperature, k1, k12, and k8 were
found to be pressure-independent over this range. The invariance
of k1 with pressure is shown in Table 5.
Figure 6. Relationship between value of k8 obtained by optimization
and value of the branching ratio for reaction 8, θ, used in the
optimization model.
Experiments were conducted as a function of temperature,
between 210 and 333 K. The rate coefficients of the secondary
reactions included in the optimization model were adjusted to
their appropriate temperature-dependent values as given in the
references indicated in Table 3. Exceptions to this were the
branching ratios for the IO + IO and IO + BrO reactions, and
the rate coefficients for reactions 13 and 15-17. These
parameters, the temperature dependencies of which are not
known, were fixed at their 295 K values. The insensitivity of
the value obtained for k1 upon these parameters (Figure 5)
indicated that the use of the model to determine the value of k1
at temperatures other than ambient was valid; however, the
optimized value of k12 was approximately 4-fold more sensitive
to the secondary chemical scheme than k1, implying somewhat
reduced confidence in the temperature-dependent values of this
rate coefficient. The value determined for k9 was solely
dependent upon the initial IO and BrO concentrations, and thus
was not susceptible to incomplete implementation of the
temperature dependence of the secondary chemistry. However,
the analysis did not take account of possible variation in the
branching ratio θ of reaction 8 with temperature.
pair of values for k1 and k12 best described the IO and BrO
decay traces. Hence k1 and k12 were both independently
determinable from the IO and BrO decay traces.
In addition to k1 and k12, the value of k8 (O + IBr) was
optimized in the reaction model. The value obtained for k8 was
determined by the initial [IO]:[BrO] ratio, as all O atoms were
consumed rapidly (with a total pseudo-first-order half-life
ranging from 3 to 25 µs dependent upon conditions), and thus
did not impact upon the subsequent form of the IO and BrO
decay traces, which determined the values of k1 and k12.
Kinetic Results Ambient Conditions. Figure 3 shows IO and
BrO decay traces, optimized fits and residuals (offset) typical
of those used to determine the values of k1, k12, and k8. At
ambient temperature (295 ( 3 K) and at 760 Torr total pressure,
these rate coefficients were found to be
k1 (IO + BrO) ) (8.5 ( 1.4) × 10-11 molecules-1 cm3 s-1
k12 (I + BrO) ) (1.3 ( 1.2) × 10-11 molecules-1 cm3 s-1
k8 (O + IBr) ) (3.6 ( 2.4) × 10-11 molecules-1 cm3 s-1
The values obtained for k1, k12, and k8 as a function of
temperature are given in Table 6. The value of k1 (IO + BrO)
was found to increase with decreasing temperature, in a manner
which could be described by an Arrhenius expression:
Errors are 2σ and represent precision only: Values were
obtained from the analysis of 33 experiments conducted over
the precursor concentration ranges indicated in Table 2, each
the coaddition of 50 or 100 laser shots.
k1 ) (6.7 ( 0.8) × 10-12 exp((760 (
30)/T) molecules-1 cm3 s-1 (iii)
The value obtained for k8 was entirely dependent upon the
branching ratio θ () k8a/k8) employed in the reaction system
model (the value of 0.7 was used for θ to obtain the above
value for k8). These experiments were thus limited to establish-
ing that k8 and θ are related as shown in Figure 6; independent
determination of either k8 or θ was not possible from our data.
Error limits correspond to (2σ and represent precision only.
The values of k1 are plotted as a function of temperature in
Figure 7, together with the temperature-dependent results
obtained by Gilles et al.20 and the 298 K results of Bedjanian
et al.21 and Laszlo et al.19

IO + BrO Reaction
J. Phys. Chem. A, Vol. 105, No. 33, 2001 7861
Figure 8. IO, BrO, and OIO decay traces, recorded using system C
(CF3I/Br2/N2O/N2). The OIO trace was obtained by differential fitting
to an uncalibrated OIO absorption spectrum.
Figure 7. Temperature-dependent values for k1 obtained in this work
(filled circles) compared with the results of Gilles et al.20 (open
triangles), Laszlo et al.19 (open squares), and Bedjanian et al.21 (filled
squares). The line is an Arrhenius fit to k1 (this work).
(c) OIO + Br. Growth and decay of the distinctive absorption
bands of OIO between 500 and 600 nm were observed in the
post-photolysis spectra of the IO + BrO reaction system.
Differential spectral fitting to an arbitrary OIO absorption
spectrum was used to obtain absorption-time profiles for OIO,
revealing behavior qualitatively similar to that observed in the
IO self-reaction:1 OIO was a transient species, forming and
decaying on a time scale similar to that of the IO radical. IO
and BrO concentration profiles, and an OIO absorption profile,
are shown in Figure 8, recorded using the CF3I/Br2 system C.
OIO is known to be a product of the IO self-reaction,1 which
also occurred in the IO + BrO reaction system. Moreover the
ratio of OIO:IO was similar to that observed in our studies of
the IO self-reaction. However, in the IO + BrO system, the
fraction of IO radicals reacting with BrO was much greater (3-4
times) than that undergoing self-reaction; therefore, the observa-
tion of similar OIO:IO ratios indicates that OIO is also a product
of the IO + BrO reaction. Further analysis which constrains
the branching ratio for OIO production from the IO + BrO and
IO + IO reactions and hence its absorption cross section is
described below.
(d) OBrO + I. No evidence for the formation of OBrO was
observed in the course of this work. On the basis of the estimated
cross section for OBrO of 1.5 × 10-17 molecules-1 cm2 at 505.7
nm,35 we determine an upper limit for channel 1d of k1d/k1 e
0.15.
(e) IBrO2. No spectroscopic evidence was observed for the
presence of any absorbing species other than the XO radicals,
the products of the IO self-reaction, OIO from IO + BrO and
the halogen and interhalogen molecules, at any of the temper-
atures investigated in the course of this work (210-333 K).
No evidence for the formation of an IBrO2 species was observed.
The pressure independence of the IO + BrO reaction rate
coefficient between 760 Torr (this work) and 1 Torr21 indicates
that termolecular IBrO2 formation does not occur. Similarly,
no evidence for IBrO2 formation has been observed in previous
studies of the IO + BrO reaction.19,20
The value of k8 (O + IBr) displayed no significant variation
with temperature (within the precision of the measurements)
over the range studied, leading to an unweighted mean value
of (4.6 ( 1.8) × 10-11 molecules-1 cm3 s-1 (210-333 K; (
2σ). k12 was also found to be temperature independent at 235
K and above; over this range, the unweighted mean value
obtained was (1.3 ( 0.4) × 10-11 molecules-1 cm3 s-1 ((2σ).
At lower temperatures (210 and 222 K), the value of k12 obtained
was significantly smaller. The temperature-dependent values of
k8 and k12 are given in Table 6.
Products of the IO + BrO Reaction. Each of the five
potential product channels of the IO + BrO reaction is
considered in turn.
(a) I + Br + O2. It was not possible to detect I or Br atoms
directly using our experimental system. The addition of ozone
to the reaction system, leading to the reconversion of the X (X
) Br, I) atoms to XO radicals, and hence to a reduction in the
observed rate coefficient for the IO + BrO reaction, would have
allowed quantification of halogen atom production. However,
when preliminary experiments were conducted in which O3 was
added to the precursor gas mixture, powdery yellow-white
deposits formed upon the surfaces of the gas manifold and
reaction cell. These deposits were similar in appearance to those
formed when O3 and molecular iodine were brought into
contact,1 and are likely to have been the same higher oxides of
iodine, I2O4, I2O5, and/or I4O9, formed in this case following
reaction between O3 and IBr. The formation of these deposits
on the optical surfaces of the system precluded the conduct of
useful experiments in the presence of ozone; it was not therefore
possible to quantify channel 1a directly.
(b) IBr + O2. Investigation of the production of IBr from
the IO + BrO reaction was complicated by the presence of this
species at significant levels in the precursor gas mixtures
employed. Experiments were conducted using system C, Br2/
CF3I, in which the production of IBr, monitored spectroscopi-
cally, was compared with the modeled flux through the IO +
BrO reaction. The IBr concentrations observed indicated a
branching ratio for this channel of the order of 0.2 (i.e., k1b/k1
≈ 0.2); however, several other potential sources of IBr existed,
notably the I + Br2 and possibly the Br + CF3I34 reactions.
The magnitude of the contributions of these reactions to the
observed IBr could not be calculated with certainty, thus the
value obtained determines an upper limit for production of IBr
from the IO + BrO reaction, k1b/k1 e 0.2.
Yield of OIO from IO + BrO and IO + IO. The branching
ratio for production of OIO from reaction 1 was constrained by
considering the results obtained in this work in conjunction with
those reported elsewhere. Bedjanian et al.21 have determined
the branching ratio for the production of OIO from the IO +
BrO reaction to be equal to or in excess of 0.65; i.e. 0.65 e
k1c/k1 e 1. Bedjanian et al. also placed upper limits for the
formation of I + Br + O2, IBr + O2 and OBrO + I of k1a/k1 e
0.5, k1b/k1 e 0.05 and k1d/k1 e 0.2, respectively. Gilles et al.20

7862 J. Phys. Chem. A, Vol. 105, No. 33, 2001
Rowley et al.
measured the rate coefficient for the non I atom producing
channels of reaction 1, i.e., kobs ) k1b + k1c, assuming k1e ) 0.
Comparison of the value obtained by Gilles et al. at 298 K, kobs
) (6.05 ( 0.57) × 10-11 molecules-1 cm3 s-1, with the value
of k1 obtained in this work, (8.5 ( 1.4) × 10-11 molecules-1
cm3 s-1, indicates that channels 1b and 1c, IBr + O2 and OIO
+ Br, account for 55 to 93% of the IO + BrO reaction, i.e.,
0.55 e (k1b + k1c)/k1 e 0.93, where the range corresponds to
the 2σ uncertainty level in the kinetic results. The range common
to this analysis and the results of Bedjanian et al. thus determines
0.65 e k1c/k1 e 0.93.
The branching ratio for OIO production from the IO + BrO
reaction was used to quantify the production of OIO from the
IO self-reaction, and so to determine the absolute absorption
cross sections of OIO. The branching ratio for the production
of OIO from the IO self-reaction and the IO + BrO reaction,
denoted R and â respectively, are defined by
Figure 9. Values of R (branching ratio for production of OIO from
IO + IO) as a function of time after photolysis, calculated for â
(branching ratio for production of OIO from IO + BrO) ) 0.65 (open
squares) and 0.93 (filled circles).
IO + IO f OIO + I
f other products
(14a)
(14b)
(iv)
R ) k14a/k14
â ) k1c/k1
TABLE 7: Comparison of Literature Values for k1 with the
Results of This Work
(v)
k1 (298 K) × 1011,
authors
date molecules-1cm3 s-1
techniqueb
LFP/UVA
DF/LFP/UVA/LIF
DF/MS
The concentration of reactant molecules (IO and BrO) is known
in both the IO self-reaction and IO + BrO reaction systems.
Because the IO + IO and IO + BrO overall rate coefficients
are also known, the flux through each of these reactions can be
evaluated, and the production rate of OIO calculated in terms
of R and â in each system:
Laszlo et al.19
Gilles et al.20
1997
1997
6.9 ( 2.7
6.05 ( 0.57a
8.5 ( 1.5
Bedjanian et al.21 1998
this work
1999
8.49 ( 1.40
LFP/UVA
a Measured only the non-I atom producing channels of reaction 1.
b LFP ) laser flash photolysis, UVA ) UV absorption spectroscopy,
DF ) discharge flow, LIF ) laser-induced fluorescence, MS ) mass
spectrometry.
OIO production rate POIO ) Rk14[IO]2(IO + IO system)
(vi)
production. The increase in R beyond 0.07 s reflects the greater
loss processes for OIO present in the more chemically diverse
IO + BrO reaction system. In practice the time period when
both the IO + IO and IO + BrO reaction rates were within
10% of their maximum values was used to determine R and
hence σOIO; this time period is indicated in Figure 9. The range
of values determined for R and σOIO are
OIO production rate POIO
)
Rk14[IO]2 + âk1[IO][BrO](IO + BrO system) (vii)
The total concentration of OIO formed at any point following
photolysis, neglecting OIO loss processes, is then given by the
integral of POIO with respect to time. Two equations were
therefore derived, one from each system, relating the calculated
OIO production to the observed OIO absorption A in terms of
0.30 ( 0.05 e R e 0.46 ( 0.08
the OIO cross section σOIO
:
and 1.29 ( 0.22 g σOIO
×
1017 g 0.87 ( 0.15 cm2 molecules-1
Aa/(σOIO L) ) Rk [IO] 2
(viii)
∫
14
a
where the range corresponds to 0.65 e â e 0.93, and the errors
quoted correspond to 2σ limits in the determinations. σOIO was
evaluated at 549 nm and the peak of the (5,1,0) band36 at 1.13
nm fwhm resolution. OIO production thus accounts for 30-
46% of the IO self-reaction, in addition to being the major
product of the IO + BrO reaction. The products of the IO self-
reaction are considered in detail in the accompanying paper1 in
light of this result.
Ab/(σOIOL) ) {Rk [IO] 2 + âk1[IO]b[BrO]b} (ix)
∫
14
b
where the subscripts a and b refer to measurements made in
the IO + IO and IO + BrO reaction systems respectively, and
L is the absorption path length. Equations viii and ix were thus
solved to determine R and σOIO in terms of â, which is known
to lie in the range 0.65 e â e 0.93. This analysis neglects the
chemical reactions responsible for the rapid removal of OIO
observed in both reaction systems, and thus was only applied
at very short times following photolysis, when IO and BrO
concentrations maximize and the OIO concentration is domi-
nated by production. The values of R derived by this method
are plotted, as a function of the value of â used and of time
following photolysis, in Figure 9. The values of R at very short
times are obscured by noise in the retrieval of the (initially low)
IO and BrO concentrations, then settle as the IO + IO and IO
+ BrO reaction rates maximize, before increasing as the loss
processes of OIO become of increasing importance relative to
4. Discussion
There have been three recent studies of the IO + BrO reaction
reported in the literature, the results from which are summarized
in Table 7. The rate coefficient obtained for the overall IO +
BrO reaction in this work is in excellent agreement with the
value reported by Bedjanian et al.21 The overall rate coefficient
is also within the range obtained by Gilles et al.,20 who obtained
a value of 6.05 × 10-11 molecules-1 cm3 s-1 for (k1b + k1c) at
298 K, and determined that the I atom production was e35%,

IO + BrO Reaction
J. Phys. Chem. A, Vol. 105, No. 33, 2001 7863
indicating an upper limit for k1 at 298 K of 1.0 × 10-10
molecules-1 cm3 s-1. The rate coefficient obtained in this work
is slightly higher than the value of (6.9 ( 2.7) × 10-11
molecules-1 cm3 s-1 obtained by Laszlo et al.19 using a similar
experimental system; this discrepancy is partially due to the
BrO cross sections used: Laszlo et al. employed the BrO cross
sections measured by Wahner et al.,30 but without the increased
magnitude recently recommended by Gilles et al.20 Making this
adjustment to the value of k1 reported by Laszlo et al. would
increase the value by approximately 10%, bringing the two
studies into better agreement.
The results reported in this paper were obtained using the
BrO cross sections of Wahner et al.,30 increased in magnitude
by 9% in line with the findings of Gilles et al.,20 to analyze
spectroscopic data and determine [BrO]. After our analysis,
Wilmouth et al.37 have published further measurements of the
BrO absorption spectrum, including a comprehensive review
of the BrO absorption cross sections. They report a mean
differential BrO cross section, obtained from a comparison of
literature results normalized for resolution, which is ap-
proximately 13% smaller than that used in this work. Adoption
of such a σ(BrO) would reduce our measured k1, by somewhat
less than 10%, as the XO kinetics in our system are complex.
Such a change would also increase the deduced OIO yield from
both the IO + BrO and IO + IO reactions. Wilmouth et al.
determined an uncertainty of 11% in the literature differential
BrO cross sections, independent of temperature. The IO cross
sections used in this work have an uncertainty of 9%.1 We
therefore estimate that the component of the systematic uncer-
tainty in our kinetic measurements which arises from the cross
sections used is approximately 15%, considering the dependence
of the value obtained for k1 upon absolute concentrations of
both species.
molecules-1 cm3 s-1 at 210 K. This analysis assumes that the
branching ratio of the IO + BrO reaction is invariant over the
pressure range covered by the three most recent studies: 100-
760 Torr (this work), 7-15 Torr (Gilles et al.20), and 1 Torr
(Bedjanian et al.21). The pressure-independence observed for
k1 in this work and the excellent agreement between the 100
Torr value of k1 obtained in this work and that reported by
Bedjanian et al. at 1 Torr support this assumption. Moreover,
the good agreement between the branching ratio ((k1b + k1c)/
k1) derived from the results of Gilles et al. at 7-15 Torr
combined with this work at 100-760 Torr compared to that
obtained by Bedjanian et al. at 1 Torr provides a further
indication that k1 is pressure independent.
The observed ratio of IO:BrO formed immediately following
the reaction of O(3P) with IBr (reaction 8) of 83%:17% is similar
to the value reported by Laszlo et al.19 of 74%:26%. As with
the kinetic results, the difference between the studies can be
attributed, at least in part, to the BrO cross sections used:
Increase of the cross sections used by Laszlo et al. would reduce
the measured [BrO] and bring the two results into closer
agreement. The measurement of k8 in terms of the branching
ratio θ for the O(3P) + IBr reaction reported here is to our
knowledge the first determination of this rate coefficient.
The value obtained for the I + BrO reaction (k12) at 298 K
((1.3 ( 1.2) × 10-11 molecules-1 cm3 s-1) is in good agreement
with other measurements of this rate coefficient of (1.2 ( 0.6)
× 10-11 (ref 19), (1.45 ( 0.20) × 10-11 (ref 38), and (1.9 (
1.1) × 10-11 (ref 39) molecules-1 cm3 s-1. No other measure-
ment of the temperature dependence of k12 has been reported.
5. Atmospheric Implications
The rate coefficient and products of the IO + BrO reaction
have been shown to vary with temperature. Under conditions
typical of the marine boundary layer/lower troposphere, the
reaction proceeds with a rate coefficient of 8.5 × 10-11
molecules-1 cm3 s-1, and produces predominantly OIO + Br
(65-93%), with minor channels which include one or more of
I + Br + O2, IBr + O2, and/or OBrO + I. Occurrence of
reaction 1 in the troposphere thus leads to the production of
tropospheric OIO, the fate of which determines the role of this
reaction. If OIO is photolabile, the formation of OIO from IO
radicals and its subsequent photolysis lead to a null cycle with
respect to ozone loss. Conversely, if, as has been suggested,16
OIO is photostable, OIO may act as a hitherto unrecognized
iodine reservoir in the troposphere and the role played by iodine
species will depend on the homogeneous and heterogeneous
reactivity of this species.
The measurement of k1b + k1c performed by Gilles et al.20
together with the measurement of k1 reported in this work
determine a value for the branching ratio for channels 1b and
1c of the IO + BrO reaction, {k1b + k1c}/k1, which is in very
good agreement with the range determined by Bedjanian et al.21
at 298 K. Taken together, these values lead to a range of 0.65-
0.93 for k1c/k1. Bedjanian et al. also obtained upper limits for
channels 1a, 1b, and 1d, of which the latter has been further
reduced to 0.15 in this work. No evidence for channel 1e has
been observed. The results of all recent studies constrain the
branching ratios for the IO + BrO reaction at 298 K to
IO + BrO f I + Br + O2 e 0.3
f IBr + O2 e 0.05
(1a)
(1b)
In the low stratosphere, typified by temperatures below 210
K, the IO + BrO reaction proceeds with a rate coefficient of
the order of 2.5 × 10-10 molecules-1 cm3 s-1, and produces
predominantly I + Br + O2 (ca. 70%, assuming production of
OBrO and IBrO2 is negligible) with a minor channel forming
OIO + Br (ca. 30%). The potential for the IO + BrO reaction
to contribute to stratospheric ozone destruction is thus likely to
be slightly greater than was assumed by Solomon et al.8
However field measurements reporting generally low strato-
spheric IO concentrations, and the high yield of OClO measured
for the IO + ClO reaction,40,41 indicate that iodine chemistry,
while it may contribute, is currently unlikely to be a major factor
in stratospheric ozone depletion. No evidence for production
of OBrO has been observed at 298 K in the course of this work,
and OBrO production at lower temperatures is unlikely given
the endothermicity of reaction 1d. The IO + BrO reaction is
not the source of the atmospheric OBrO concentrations reported
f OIO + Br
0.65-0.93
(1c)
(1d)
f OBrO + I e 0.15
f IBrO2
(1e)
Comparison of the temperature-dependent results of this work
with those of Gilles et al.20 (Figure 7) indicates that the
branching ratio for the non-I atom producing channels of
reaction 1, 1b, and 1c decreases with decreasing temperature,
from 71% at 295-298K to 36% at 210 K. The branching ratio
of channels 1a and/or 1d therefore increases with decreasing
temperature. Because no evidence for OBrO formation has been
observed, and this channel is calculated to be considerably
endothermic at 298 K, it is possible that channel 1a, I + Br +
O2, dominates the IO + BrO reaction at low temperatures: If
indeed channel 1d does not occur, k1a ) 1.6 × 10-10

7864 J. Phys. Chem. A, Vol. 105, No. 33, 2001
Rowley et al.
by Renard et al.,23 which are in any case unsupported by more
recent observations.6
(22) Stutz, J.; Hebstreit, K.; Hoenninger, G.; Platt, U. Geophys. Res.
Abs. 1999, 1, 520.
(23) Renard, J. B.; Pierre, M.; Robert, C. J. Geophys. Res. 1998, 103,
25383.
Acknowledgment. D.M.R. thanks the UK NERC for the
award of an advanced fellowship. W.J.B. thanks the UK NERC
for the award of a studentship. This work was part-funded by
the EC Environment program under the project “Laboratory
Experiments of Iodine Chemistry in the Stratosphere-LEXIS”
(Project number ENV4-CT95-0013).
(24) Erle, F.; Platt, U.; Pfeilsticker, K. Geophys. Res. Lett. 2000, 27,
2217.
(25) Rattigan, O. V.; Jones, R. L.; Cox, R. A. Chem. Phys. Lett. 1994,
230, 121.
(26) Li, Z. J. Phys. Chem. A 1999, 103, 1206.
(27) Chipperfield, M. P.; Glassup, T.; Pundt, I.; Rattigan, O. V. Geophys.
Res. Lett. 1998, 25, 3575.
(28) Harwood, M. H.; Rowley, D. M.; Jones, R. L.; Cox, R. A. J. Phys.
Chem. A 1998, 102, 1790.
References and Notes
(29) Cole, L. G.; Elverum, G. W. J. Chem. Phys. 1952, 20, 1543.
(30) Wahner, A.; Ravishankara, A. R.; Sander, S. P.; Friedl, R. R. Chem.
Phys. Lett. 1988, 152, 507.
(31) Atkinson, R.; Baulch, D. M.; Cox, R. A.; Hampson, R. F.; Kerr, J.
A.; Troe, J. J. Phys. Chem. Ref. Data 1992, 21, 1125.
(32) Loewenstein, L. M.; Anderson, J. G. J. Phys. Chem. 1987, 91, 2993.
(33) Curtis, A. R.; Sweetenham, W. P. FACSIMILE, AERE Harwell
Publication R 12805, Computer Science and Systems Division, Harwell
Laboratory: Oxfordshire, U.K. 1987.
(1) Bloss, W. J.; Rowley, D. M.; Cox, R. A.; Jones, R. L. J. Phys.
Chem. A 2001, 105, 7840.
(2) Harder, H.; Camy-Peyret, C.; Ferlemann, F.; Fitzednberger, R.;
Hawat, T.; Osterkamp, H.; Schneider, M.; Perner, D.; Platt, U.; Vradelis,
P.; Pfeilsticker, K. Geophys. Res. Lett. 1998, 25, 3843.
(3) Hegels, E.; Crutzen, P. J.; Klupfel, T.; Perner, D.; Burrows, J. P.
Geophys. Res. Lett. 1998, 25, 3127.
(4) Pundt, I.; Pommereau, J. P.; Phillips, C.; Lateltin, E. J. Atmos. Chem.
1998, 30, 173.
(5) Wennberg, P. O.; Brault, J. W.; Hanisco, T. F.; Salawitch, R. J.;
Mount, G. H. J. Geophys. Res. 1997, 102, 8887.
(6) Wittrock, F.; Mu¨ller, R.; Richter, A.; Bovensmann, H.; Burrows,
J. P. Geophys. Res. Lett. 2000, 27, 1471.
(7) Platt, U. Phys. Chem. Chem. Phys. 1999, 24, 5409.
(8) Solomon, S.; Garcia. R. R.; Ravishankara, A. R. J. Geophys. Res.
1994, 99, 20491.
(9) Alicke, B.; Hebestreit, K.; Stutz, J.; Platt, U. Nature 1999, 397,
572.
(10) Fitzenberger, R.; Bosch, H.; Camy-Peyret, C.; Chipperfield, M. P.;
Harder, H.; Platt, U.; Sinnhuber, B. M.; Wagner, T.; Pfeilsticker, K.
Geophys. Res. Lett. 2000, 27, 2921.
(11) Carpenter, L. J.; Sturges, W. T.; Penkett, S. A.; Liss, P. S.; Alicke,
B.; Hebestreit, K.; Platt, U. J. Geophys. Res. 1999, 104, 1679.
(12) Stutz, J.; Hebestreit, K.; Alicke, B.; Platt, U. J. Atmos. Chem. 1999,
34, 65.
(34) Okafu, E. N.; Whittle, E. Int. J. Chem. Kinet. 1975, 7, 287.
(35) Miller, C. E.; Nickolaisen, S. L.; Francisco, J. S.; Sander, S. P. J.
Chem. Phys. 1997, 107, 2300.
(36) Himmelmann, S. J.; Orphal, J.; Bovensmann, H.; Richter, A.;
Ladsta¨tter-Weissenmayer, A.; Burrows, J. P. Chem. Phys. Lett. 1996, 251,
330.
(37) Wilmouth, D. M.; Hanisco, T. F.; Donahue, N. M.; Anderson, J.
G. J. Phys. Chem. A 1999, 103, 8935.
(38) Bedjanian, Y.; LeBras, G.; Poulet, G. Chem. Phys. Lett. 1997, 266,
233.
(39) Gilles, M. K.; Turnipseed, A. A.; Burkholder, J. B.; Ravishankara,
A. R. Chem. Phys. Lett. 1997, 272, 75.
(40) Turnipseed, A. A.; Gilles, M. K.; Burkholder, J. B.; Ravishankara,
A. R. J. Phys. Chem. A 1997, 101, 5517.
(41) Bedjanian, Y.; LeBras, G.; Poulet, G. J. Phys. Chem. A. 1997, 101,
4088.
(42) 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
Modelling. Evaluation No. 12; JPL Publication 97-4; Jet Propulsion
Laboratory: Pasadena, CA, 1997.
(43) Afeefy, H. Y.; Liebman, J. F.; Stein, S. E. Neutral Thermochemical
Data. In NIST Chemistry WebBook; NIST Standard Reference Database
Number 69; Mallard, W. G., Linstrom, P. J., Eds.; National Institute of
Standards and Technology: Gaithersburg, MD, November 1998.
(44) Dolson, D. A.; Leone, S. R. J. Chem. Phys. 1982, 77, 4009.
(45) Haugan, H. K.; Weitz, E.; Leone, S. R. Chem. Phys. Lett. 1985,
119, 75.
(46) Burkholder, J. B. Int. J. Chem. Kinet. 1998, 30, 571.
(47) Jenkin, M. E.; Cox, R. A.; Mellouki, G.; LeBras, G.; Poulet, G. J.
Phys. Chem. 1990, 94, 2927.
(13) Rowley, D. M.; Harwood, M. H.; Freshwater, R. A.; Jones, R. L.
J. Phys. Chem. 1996, 100, 3020.
(14) Poulet, G.; Biggs, P.; Burrows, J. P.; Cox, R. A.; Coutzen, P. J.;
Hayman, G. D.; Jenkin, M. E.; LeBras, G.; Moortgat, G. K.; Platt, U.;
Schindler, R. N. Halogen Oxides: Radicals, Sources and ReserVoirs in
the Laboratory and in the Atmosphere; Wayne, R. P., Ed.; European
Commission: Brussels, 1995. Atmos. Environ. 1995, 29, 2675.
(15) Vaida, V.; Simon, J. D. Science 1995, 268, 1443.
(16) Cox, R. A.; Bloss, W. J.; Jones, R. L.; Rowley, D. M. Geophys.
Res. Lett. 1999, 26, 1857.
(17) Misra, A.; Marshall, P. J. Phys. Chem. A 1998, 102, 9056.
(18) Guha, S.; Francisco, J. S. J. Phys. Chem. A 1997, 101, 5347.
(19) Laszlo, B.; Huie, R. E.; Kurylo, M. J. J. Geophys. Res. 1997, 102,
1523.
(20) Gilles, M. K.; Turnipseed, A. A.; Burkholder, J. B.; Ravishankara,
A. R.; Solomon, S. J. Phys. Chem. A 1997, 101, 5526.
(21) Bedjanian, Y.; LeBras, G.; Poulet, G. J. Phys. Chem. A 1998, 102,
10510.
(48) Sander, S. P.; Watson, R. T. J. Phys. Chem. 1981, 85, 4000.