Heterogeneous reaction of HOI with sodium halide salts

Article abstract of DOI:10.1021/jp0044678

The interaction of gaseous HOI with crystalline grains of NaCl and sea-salt and thin films of NaBr crystals has been studied in a wall coated tubular flow reactor coupled to a quadrupole mass spectrometer over a concentration range (0.2-8) × 10¹² molecules cm^-3 at 278 and 298 K. On a fresh surface, the uptake coefficients determined were independent of temperature with γ = 0.034 ± 0.009, γ = 0.016 ± 0.004, and γ = 0.061 ± 0.021 for NaBr, NaCl, and sea-salt, respectively. No increase in reactivity was observed on addition of water vapor between 0 and 23% relative humidity at 278 K. It was also shown that the reactivity of the salt surface decreased with time of exposure to HOI and that steady-state uptake was slower on aged salt surfaces. Products of the reactions released into the gas phase were IBr, ICl, and IBr + ICl for the reaction of HOI on NaBr, NaCl, and sea-salt surfaces, respectively. The atmospheric implications of the results obtained are briefly discussed.

Full text of DOI:10.1021/jp0044678

J. Phys. Chem. A 2001, 105, 5165-5177
5165
Heterogeneous Reaction of HOI with Sodium Halide Salts
Juliane C. Mo1ssinger and R. Anthony Cox*
Centre for Atmospheric Science, Department of Chemistry, UniVersity of Cambridge,
Lensfield Road, Cambridge, CB2 1EW, U. K.
ReceiVed: December 12, 2000; In Final Form: February 28, 2001
The interaction of gaseous HOI with crystalline grains of NaCl and sea-salt and thin films of NaBr crystals
has been studied in a wall coated tubular flow reactor coupled to a quadrupole mass spectrometer over a
concentration range (0.2-8) × 1012 molecules cm-3 at 278 and 298 K. On a fresh surface, the uptake
coefficients determined were independent of temperature with γ ) 0.034 ( 0.009, γ ) 0.016 ( 0.004, and
γ ) 0.061 ( 0.021 for NaBr, NaCl, and sea-salt, respectively. No increase in reactivity was observed on
addition of water vapor between 0 and 23% relative humidity at 278 K. It was also shown that the reactivity
of the salt surface decreased with time of exposure to HOI and that steady-state uptake was slower on “aged”
salt surfaces. Products of the reactions released into the gas phase were IBr, ICl, and IBr + ICl for the
reaction of HOI on NaBr, NaCl, and sea-salt surfaces, respectively. The atmospheric implications of the
results obtained are briefly discussed.
Introduction
An interest in the tropospheric chemistry of iodine has been
ICl, and IBr18 via the reactions
HOI + NaCl f ICl + NaOH
HOI + NaBr f IBr + NaOH
(R5)
(R6)
developed due to its potential involvement in tropospheric ozone
loss.1-3 The principal known source gases for iodine in the
troposphere are alkyl halides emitted from macroalgae and
phytoplankton, which can be found in coastal regions as well
as in the open oceans.4-8 Photolysis of alkyl halides produces
an iodine atom, which can then react with ozone to form the
IO radical, which has recently been observed in three different
locations in the remote marine boundary layer.9,10 The IO radical
can then react further with HO2, NO2, and IO leading to the
formation of a number of inorganic gas-phase species with the
major component predicted to be hypoiodous acid, HOI.2,11
The interhalogens formed are only slightly soluble and can be
released back into the gas phase, where they readily photolyze19
IX + hν f X + I (X ) Cl or Br)
(R7)
The halogen atoms produced could then lead to enhanced ozone
loss in the marine boundary layer.
Only one other study of the interaction of HOI with salt
surfaces has been published to date. Allanic et al.20 investigated
the reaction of HOI on NaCl and KBr surfaces at room
temperature using a Knudsen cell. They reported maximum
gamma values for the reactive uptake coefficients of γ ) (4 (
2) × 10-2 and γ ) (5 ( 2) × 10-2 for the reaction of HOI on
NaCl and KBr respectively, measured at the shortest reaction
times. They found that the uptake coefficient decreased with
exposure of the salt sample to HOI. The only product observed
in their study of the reaction of HOI on NaCl was I2 but both
I2 and IBr were observed in the reaction of HOI on KBr.
To investigate further the possible release of interhalogens
via the reaction of HOI with halide salts, a comprehensive study
of the heterogeneous interaction of gaseous HOI with NaBr,
NaCl, and sea-salt crystals was conducted using a wall coated
flow reactor. The kinetics of HOI uptake on solid NaBr, NaCl,
and sea-salt crystals was measured at 298 and 278 K. Uptake
measurements were also conducted on a solid NaOH surface
and in the presence of water vapor (up to 23% relative humidity
at 278 K), to explore the reactivity of the condensed phase
product, and the effect of adsorbed water on the reactivity of
the solid surfaces. The implications of the resulting uptake
coefficients are discussed in terms of the atmospheric chemistry
of iodine.
12
HOI is formed by the reaction of IO with HO2
IO + HO2 f HOI + O2
(R1)
It can then regenerate IO by photolysis13,14 contributing to ozone
destruction via the following reaction cycle12
HOI + hν f OH + I
I + O3 f IO + O2
(R2)
(R3)
IO + HO2 f HOI + O2
(R4)
HOI can also partition between the gaseous and the con-
densed phase and undergo heterogeneous reaction with atmo-
spheric aerosol particles. Its uptake into the condensed phase
may then lead to removal by wet deposition. Thus, HOI could
be the reservoir species responsible for the apparently rever-
sible attachment of I131 to aerosol particles, which has been
noted during the transport of radioactive material following
the accidental release at Chernobyl in 1986.15,16 In addition,
it has recently been proposed that HOI, analogous to
HOBr,17 could also be involved in the activation of halogens
from sea-salt aerosols due to formation of interhalogen species,
Experimental Section
A conventional flow system with mass spectrometric detection
was used to study the uptake kinetics and gaseous reaction
* To whom correspondence should be addressed. Fax: +44-1223-336362.
E-mail: jcm34@hermes.cam.ac.uk, rac26@cam.ac.u
10.1021/jp0044678 CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/03/2001

5166 J. Phys. Chem. A, Vol. 105, No. 21, 2001
Mo¨ssinger and Cox
TABLE 1: Experimental Conditions for Gases Used
detection
concentration/
molecule
O2
C2H5I/C3H7I
HOI
ICl
IBr
I2
limit/molecules cm-3
molecules cm-3
% error
5 × 1010
1 × 1011
2 × 1010
1 × 1010
3 × 1010
6 × 1010
(0.2-20) × 1013
(0.2-20) × 1013
(0.2-8) × 1012
(2-40) × 1011
(2-6) × 1011
2
5
46
27
27
18
(7-10) × 1011
the flow reactor. The insert was then dried quickly by pumping
out the reactor leaving a slightly opaque thin film of NaOH or
NaBr.
HOI was prepared in situ in the double sliding injector via
the reaction of O atoms with C2H5I or C3H7I23
O2 f 2 O(3P)
(R8)
(R9)
CxHyI + O(3P) f HOI + CxHy-1 > 80%
CxHyI + O(3P) f IO + CxHy < 10%
(R10)
where x ) 2 or 3 and y ) 5 or 7 with k ) 2.38 × 10-11 cm3
molecules-1 s-1 24 and k ) 5.18 ×10-11 cm3 molecules-1 s-1 25
for the reaction with C2H5I and C3H7I, respectively.
A double sliding injector (see Figure 1) was used in which a
flow of O atoms, produced in a microwave discharge of O2-
He mixtures, passed up the inner injector. The O atoms were
then mixed a few centimeters before the end of the sliding
injector with a flow of excess C2H5I or C3H7I in He added via
the outer injector, which was coated with halocarbon wax. The
yield of HOI is in excess of 80%. IO and I2 were present as
small impurities, as detected by mass spectrometry. The resulting
HOI concentrations were quite stable in the flow tube. For the
study of the reactions of I2, ICl, and IBr on NaCl, NaBr, sea-
salt, and NaOH the double sliding injector was exchanged for
a single sliding injector as described previously.22 Samples of
C3H7I (Lancaster 98%), C2H5I (Lancaster 98%), I2 (Breckland
Scientific 98%), ICl (Lancaster 98%), and IBr (Acros 98%) were
purified by repetitive freeze pump thaw cycles. Known pressures
of the gases were then mixed with an appropriate pressure of
Helium (Messer 5.0) to give the desired percentage mixture
(0.01-2%) and stored in bulbs on an all glass manifold. The
mixtures were left to mix overnight. They were kept in the dark
and usually used the next day. The O2-He mixture was either
used straight from the gas cylinder (O2 4.5 in He 5%, Messer)
or diluted further to the desired percentage mixture of 0.5% in
a bulb on the glass manifold.
Gas concentrations were measured using a molecular beam
sampling quadrupole mass spectrometer (Hiden Analytical
model HAL IV). HOI, IBr, I2, ICl, C2H5I, and C3H7I were
monitored at m/z ) 144, 206, 254, 162, 156, and 171
respectively, using an ionization energy of 70 eV and with a
current of 1000 µA. On addition of a flow of a dilute mixture
of any of the gases listed above, the mass spectrometer signal
was found to be linearly dependent on concentration and
calibration curves were obtained for IBr, I2, ICl, C2H5I, and
C3H7I. Decomposition of some of the ICl and IBr into I2 and
Cl2/Br2 was taken into account in the calibrations. The concen-
tration of HOI was determined by subtracting the amount of I2
formed from the amount of C2H5I or C3H7I lost assuming that
no other iodine containing species were produced when the
microwave cavity was switched on. The minimum detectable
concentrations, uncertainties in the measured concentrations and
typical concentrations used are listed in Table 1. The minimum
Figure 1. Experimental setup.
products. The apparatus (Figure 1) is similar to that used
previously for the study of the interaction of HNO3 with
NaCl and the uptake of HI and HBr on ice,21,22 except for the
double sliding injector, which was used for the in situ prep-
aration of HOI. The jacketed flow tube was approximately 1m
in length with an internal diameter of 1.90 cm. A cylindrical
Pyrex inner tube 70 cm long and 1.57 cm i.d. was coated
internally with the substrate of interest and then inserted inside
the flow reactor. This provides the surface for the kinetic
measurements.
Two types of surfaces were used in this work. The first type
was composed of granular crystallites, which was made using
commercial NaCl grains (Breckland Scientific, 99%+) with a
diameter of approximately 0.5 mm or sea-salt grains (Com-
munity Food Ltd, 99%+) of 0.2-0.7 mm diameter, obtained
from solar evaporation of seawater free from any additives or
processing aids, deposited in a layer approximately one grain
thick. The preparation of this surface type has been described
in detail elsewhere.21 The second type was made in situ by
evaporation of a saturated solution of NaBr (Lancaster 99%+)
in 50%:50% distilled water/methanol (BDH, 99.8%) or from a
1 molar NaOH solution (Ampules Convol, BDH). To avoid
formation of Na2CO3, the distilled water used to make up the
NaOH solution was boiled prior to use. Once the NaOH solution
or the saturated solution of NaBr was made up the inner Pyrex
tube insert was immediately coated with it and inserted into

Reaction of HOI with Sodium Halide Salts
J. Phys. Chem. A, Vol. 105, No. 21, 2001 5167
TABLE 2: Diffusion Coefficients Used for Brown
Corrections
and 5%. The error in the slope to determine ksurface was up to
10%, the error for the diffusion coefficients used was 11%. An
error of 20% and 30% is assigned to the limited reproducibility
of the dried solution surfaces and the salt grain surfaces,
respectively. This leads to an overall error on the uptake
coefficient of 18%-35%. The correction for radial diffusion
using the Brown corrections26 described above varied between
5 and 60%. They were small when uptake was slow (γ < 10-2),
diffusion coeff
diffusion coeff
molecule
temp/K
in He/Torr cm2 s-1
in H2O/Torr cm2 s-1
HOI
ICl
IBr
I2
253
278
278
278
224 ( 24.627
288.2 ( 31.7
268.4 ( 29.5
250.6 ( 7.6
33.0 ( 3.627
56.6 ( 6.2
52.0 ( 5.7.
48.0 ( 5.3
but became much larger as γ approached 0.1. When kdiffusion
<
reactant concentration used during the experiments was limited
to 2 × 1011 molecules cm-3 due to the detection limit of 2 ×
1010 molecules cm-3 for HOI. The initial concentration of HOI
has to be sufficiently large to be able to detect changes in the
HOI mass signal.
Experiments were performed at 298 and 278 K using a flow
of Helium (50-500 sccm) as carrier gas, which was injected
through a side inlet at the upstream end of the flow reactor.
In the experiments performed at different relative humidity
(0-23%), a continuous flow of water vapor was added by either
bubbling 150 sccm of Helium through a Pyrex vessel containing
high purity water (Acros Organics, water for HPLC) at low
pressure or by drawing the water vapor directly from the vessel
through another side inlet of the flow reactor. The linear flow
velocity varied between 1047 and 2560 cm3 s-1 and the total
pressure of the flow tube was 1.2-3.5 Torr.
For any given experiment, an initial flow of the reactant was
established with the end of the sliding injector downstream from
the surface of interest. The injector was then moved to an
upstream position so that a certain length of the surface was
exposed to the reactant. Any uptake on the surface leads to a
drop in concentration of the reactant and hence in its signal on
the mass spectrometer, and an increase in the product signal, if
a product is released into the gas phase. By varying the length
of surface exposed the concentration time profile along the flow
tube can be determined. Normally, the decay of the reactant
was first order and the first-order rate constant ksurface can be
obtained from a plot of ln[signal] vs contact time. ksurface can
then be used to calculate the uptake coefficient, γ, assuming
that the surface area of the substrate is equivalent to the
geometrical surface area of the inner Pyrex tube, using the
equation
ksurface the loss of the reactant molecule is limited by the gas-
phase diffusion in the flow tube rather then by the reactive
uptake onto the surface.28 Under such conditions only a lower
limit for the uptake coefficient can be determined. For all
calculations of γ, the geometric surface area of the Pyrex insert
was assumed to be equal to the actual surface area available
for reaction. Irregularities in the surface may lead to a larger or
smaller surface area, as will be discussed later, which would
introduce an additional error for the uptake coefficients calcu-
lated using equation (E1) above.
Results
I. Interaction of HOI with NaCl, NaBr, Sea-Salt and
NaOH Surfaces. I. (a). Uptake of HOI on Fresh Salt Sur-
faces. Figure 2a shows the loss of HOI (1.6 × 1012 molecules
cm-3) on NaCl and the production of ICl. The initial HOI sig-
nal was established ∼2 min after the microwave cavity was
switched on. After ∼4 min the sliding injector was pulled down
to expose the first few centimeters of the surface to HOI and
the initial HOI signal decreased to a new steady state value. It
can be seen from the increase in ICl signal, that ICl was released
promptly into the gas phase. The sliding injector was then
repeatedly pulled down to expose different length of the NaCl
surface.
On exposure to the NaBr surface HOI (1.1 × 1012 molecules
cm-3) was continuously lost to the surface and the product of
the reaction, IBr, was released promptly into the gas phase. This
is shown in Figure 2b. The sliding injector was pulled back
after about 6 min for the first time, and the HOI signal decreased,
whereas the IBr signal increased. After 10 min, the microwave
cavity was switched off to check the position of the baseline.
At 13 min, the cavity was switched back on and the experiment
was repeated.
γ ) ksurface 2r/ω
(E1)
When HOI (1.0 × 1012 molecules cm-3) was lost to the sea-
salt surface (Figure 2c)), IBr was observed to be the first product
of the reaction reaching the gas phase. The IBr production was
found to be time dependent in most cases. At 34 min and 43
min, it can be seen that the initial IBr signal started to decay
away almost instantly, while the signal due to ICl production
started to increase until it reached a steady-state value. After
51 min, the sliding injector was pushed back up past the sea-
salt surface and the HOI signal increased to its initial value,
while the ICl signal decreased to its baseline. On fresh sur-
faces, initial uptake was time dependent, falling from an initial
high value to the time independent γ (see Figure 3). However,
all values for the uptake coefficient, γ, of HOI on the salt
surfaces have been measured after the uptake state had settled
to a constant value, as the initial time dependent uptake was
not very reproducible. In addition, the maximum change in
ksurface due to this initial time dependence was a factor of
2, which results in a value for the uptake coefficients within
the error limits for the uptake coefficient determined under
steady-state conditions. The initial time dependent uptake might
become larger when lower concentrations of HOI are used.
where γ ) uptake coefficient, ω ) average molecular speed
and r ) radius of the inner Pyrex tube insert.
Corrections for ksurface need to be applied for the laminar flow
conditions used. The correction for axial diffusion was found
to be insignificant (<2%). The correction for radial diffusion
was made using the method described by Brown.26 The diffusion
coefficients used for the Brown correction are listed in Table
2. The diffusion coefficients for HOI at 253 K were kindly
provided by J. W. Adams27 prior to publication. For ICl, IBr,
and I2 the diffusion coefficients were calculated using kinetic
theory, as described in detail previously.22 Due to a lack of data
for the characteristic energy ꢀ and characteristic length σ for
IBr in He and H2O, the diffusion coefficients for IBr were
estimated using values of ꢀ and σ lying between those for I2
and ICl.
Uncertainties in temperature, pressure, sample purity, small
leaks, gas mixing, slope to determine ksurface, gas flow, reproduc-
ibility of the preparation of surfaces, and the diffusion coef-
ficients used for the diffusion corrections were considered for
error analysis. The error attributed to temperature, pressure,
sample purity, small leaks, and gas mixing varied between 1

5168 J. Phys. Chem. A, Vol. 105, No. 21, 2001
Mo¨ssinger and Cox
Figure 3. Initial reaction of HOI (1.5 × 1012 molecules cm-3) on a
fresh sea-salt surface at 298 K.
Figure 4. First-order loss plot for the reaction of HOI on sea-salt at
298 K. Two consecutive runs.
Figure 4 shows a typical first-order loss plot for HOI on sea-
salt. The two consecutive runs were performed on the same
surface.
The uptake coefficients for the loss of HOI on the salt surfaces
were found to be independent of temperature between 278 and
298 K with γ ) 0.034 ( 0.009, γ ) 0.016 ( 0.004, γ ) 0.061
( 0.021 for HOI on NaBr, NaCl, and sea-salt, respectively.
The loss of HOI on a clean Pyrex tube insert surface was found
to be insignificant with γ ) 0.0005 ( 0.0001.
I. (b). Surface Aging. It was observed that the reactivity of
the dry salt surfaces was dependent on the time of exposure of
HOI to the dry salt surfaces. With increasing exposure of HOI
to the salt surfaces a decrease in ksurface was observed, i.e., γ
for the reaction decreased. This behavior, called “surface aging”,
has also been observed in other studies.21,20 The measured γ
values on “aged” surfaces, e.g., after ∼30 min of exposure of
the surface to HOI, were determined to be γ ) 0.008 ( 0.003,
γ ) 0.007 ( 0.003, γ ) 0.014 ( 0.004 for NaBr, NaCl, and
sea-salt, respectively. No concentration dependence of this
Figure 2. (a) The uptake of HOI (1.6 × 1012 molecules cm-3) on
NaCl and the production of ICl at 298 K. (b) The uptake of HOI (1.1
× 1012 molecules cm-3) on NaBr and the production of IBr at 298 K.
(c) The uptake of HOI (1.0 × 1012 molecules cm-3) on sea-salt and
the production of IBr and ICl at 298 K.

Reaction of HOI with Sodium Halide Salts
J. Phys. Chem. A, Vol. 105, No. 21, 2001 5169
surface aging effect was observed over the range of HOI
concentration used in this study. However, at lower HOI
concentrations, it should be possible to observe a concentration
dependence of the effect of surface aging.
I. (c). Effect of Water Vapor. Water was added to the reactor
at 278 K up to 23% relative humidity (1.5 Torr water vapor)
for the reaction of HOI on NaCl and NaBr and up to 11%
relative humidity (0.7 Torr water vapor) for the reaction of HOI
on sea-salt. In all cases, no dependence of γ on relative humidity
was observed. With the present system, it was not possible to
conduct experiments with higher relative humidity in the flow
tube. Further increase in water vapor partial pressure would have
required an excessive flow tube pressure, to values incompatible
with the mass spectrometer sampling configuration. A decrease
in orifice size would have resulted in a loss of sensitivity and
thus very high initial HOI concentrations, which would have
been unsuitable for experiments on dry salt surfaces.
I. (d). Product Formation. For the reaction of HOI on NaCl
and NaBr, ICl and IBr, respectively, were the only products
detected in the gas phase. Both the loss of HOI as well as the
production of ICl and IBr were time independent over the time
scale of the experiment. The concentration of ICl produced was
calculated to be (10 + 53/-10)% of the concentration of HOI
lost to a freshly exposed surface as well as to an “aged” NaCl
surface. On a fresh NaBr surface, the amount of IBr produced
was calculated to be (100 ( 53)% of the amount of HOI lost.
After several experiments carried out on the same NaBr surface
the conversion of HOI to IBr decreased to (19 + 53/-19)%.
The large errors are mainly due to the uncertainty in the
concentration of HOI.
Figure 5. Uptake of HOI (1.4 × 1012 molecules cm-3) on NaOH at
298 K.
On exposure of HOI to sea-salt both ICl and IBr were
observed as products. IBr was observed to be the first product
of the reaction reaching the gas phase. Almost instantly, the
initial IBr signal started to decay away while the signal due to
ICl production started to increase until it reached a steady state
value. Thus, IBr and ICl production was time dependent, and
at the shortest exposure time on a fresh surface IBr and ICl
were produced in a 4:1 ratio. With time the ratio declined and
eventually only ICl but no IBr production was observed.
HOI was also found to be taken up on the surface of NaOH,
the most likely product of the reaction of HOI on halide salts,
which is not released into the gas phase (see section Ie) below).
Stoichiometrically, for every IBr or ICl produced there is also
OH- and Na+ produced on the salt surfaces, which may form
the solid reaction product NaOH on the surface. On a medium
aged sea-salt surface (30 + 63/-30)% of HOI lost was
converted to ICl and IBr. Taking into account that HOI can
also be lost to NaOH, (64 ( 63)% of HOI lost could be
accounted for. Within error this could close the mass balance
for the reactions of HOI on sea-salt. Analysis of the data for
the first initial exposure of a fresh sea-salt surface to HOI, i.e.,
no NaOH present, showed that under these conditions (81 (
63)% of the HOI lost was converted to ICl and IBr.
I. (e). Interaction of HOI with NaOH. The other possible
product of the reaction of HOI on all three salt surfaces was
NaOH, which is not likely to be released into the gas phase.
To assess the possible interaction of HOI with NaOH, which is
assumed to be deposited on an aged surface, the uptake of HOI
on an NaOH surface has also been investigated.
Figure 6. Decomposition reaction of HOI and production of I2 within
the outer sliding injector.
same reasons as outlined in section Ia. The uptake of HOI (1.4
× 1012 molecules cm-3) on NaOH is shown in Figure 5.
Between 278 and 298 K, no temperature dependence of the
reaction was observed. On a fresh surface, the uptake coefficient
was determined to be γ ) 0.016 ( 0.004. After repeat exposure
of HOI to the surface, the value for the uptake coefficient
decreased to γ ) 0.0022 ( 0.0006. No gas-phase products were
observed from this reaction. Addition of water vapor at 278 K
to 11% relative humidity increased the value for ksurface
dramatically. Uptake was found to be diffusion-limited with γ
g 0.1. As the deliquescence point of NaOH occurs at 7%
relative humidity,29 a solid NaOH surface will change to a liquid
layer. Thus, the results suggest that reactivity of HOI on a liquid
surface is much greater than on a solid surface.
The initial reaction rate of HOI upon exposure to a fresh
NaOH surface was time dependent, and the uptake coefficient
decreased to a steady-state value after the initial exposure. The
uptake coefficients reported for the uptake of HOI on an NaOH
surface were measured under steady-state conditions for the
I. (f). Decomposition Reaction of HOI. During the course of
the optimization of the HOI source, it was noticed that the HOI
signal decreased with increasing length of exposure to the inner
wall of the outer sliding injector. At the same time, the I2 signal
increased (Figure 6). This behavior might be explained by the

5170 J. Phys. Chem. A, Vol. 105, No. 21, 2001
Mo¨ssinger and Cox
Figure 8. Loss of ICl (3.1 × 1011 molecules cm-3) and production of
Figure 7. Second-order loss plot for the decomposition of HOI.
IBr on a NaBr surface at 298 K.
following reaction scheme suggesting a decomposition reaction
of HOI on the inner wall of the outer sliding injector
2 HOI(g) f 2 HOI(ad)
2 HOI(ad) f I2O + H2O
I2O f I2 + 0.5 O2
(R11)
(R12)
(R13)
To test the suggested reaction scheme, the data were plotted as
a second-order loss plot for the loss of HOI on the walls of the
outer injector (Figure 7). No decomposition after dilution in
the flow tube was observed.
II. Investigation of the Interaction of ICl, IBr, and I2 with
NaCl, NaBr, Sea-Salt and NaOH Surfaces. To obtain some
information about the behavior of the products of the HOI +
NaX reaction, the interaction of ICl and IBr with NaX as well
as of that of the largest impurity present, I2, was also studied.
No detectable loss of either ICl, IBr, or I2 on a clean Pyrex
insert was observed.
II. (a). ICl, IBr, and I2 on NaCl and NaBr. No interaction of
I2, ICl, or IBr was observed with the NaCl surface. I2 and IBr
did not interact with the NaBr surface either. In contrast, ICl
was lost rapidly to the NaBr surface with production of IBr.
The initial rate upon exposure of ICl to a fresh NaBr surface
was rapid but time dependent and decreased to a steady state
value after approximately 4 min. The uptake coefficient reported
for the uptake of ICl on a NaBr surface was measured under
steady-state conditions only for the same reasons as discussed
Figure 9. Physical adsorption and desorption of IBr (3.0 × 1011
molecules cm-3) on sea-salt at 298 K.
7% of the initial IBr concentration was lost. No Br2 or any other
gaseous products of the reaction were observed. When the
sliding injector was moved back to its initial position so that
the surface was no longer exposed to IBr, all the IBr initially
adsorbed onto the salt surface was desorbed from the surface
again. This physical adsorption process of IBr (3.0 × 1011
molecules cm-3) on sea-salt is shown in Figure 9. The steady-
state uptake coefficient was determined to be γ ) 0.0006 (
0.0002.
in section Ia. The uptake of ICl (3.1 × 1011 molecules cm-3
)
on NaBr is shown in Figure 8. After 15 min, the first few
centimeters of the NaBr surface were exposed to ICl resulting
in a decrease of the ICl signal and an increase of the IBr signal.
This was repeated several times until, after 47 min, the sliding
injector was pushed back past the NaBr surface, and the ICl
and IBr signals returned to their initial values. The amount of
ICl converted to IBr on a fresh surface was determined to be
(100 ( 34)%. However, the conversion decreased to (70 ( 34)%
after three consecutive runs on the same NaBr surface indicating
some loss of reactivity with exposure to ICl. The initial steady-
state uptake coefficient was determined to be γ ) 0.0068 (
0.0018.
ICl (3.6 × 1011 molecules cm-3) was observed to react with
sea-salt as can be seen in Figure 10. After 17, 22, and 29 min
increasing lengths of the sea-salt surface were exposed to ICl,
which lead to production of IBr. At 34 min, the sliding in-
jector was pushed back up past the sea-salt surface. It was
observed that only (30 + 34/-30)% of the amount of ICl lost
to the surface was converted to IBr, whereas (70 ( 34)%
desorbed again as ICl. The reaction was found to be time
dependent. After several consecutive experiments on the same
sea-salt surface (in this case at 40 min), IBr production was no
longer observed and (97 ( 34)% of the ICl lost to the surface
II. (b). ICl, IBr, and I2 on Sea-Salt. I2 did not interact with
the sea-salt. On exposure of IBr to the sea-salt surface about

Reaction of HOI with Sodium Halide Salts
J. Phys. Chem. A, Vol. 105, No. 21, 2001 5171
Discussion
I. Formation of ICl and IBr from the Reaction of HOI
on Sea-Salt. It has been shown (results section I, parts a and d)
that the uptake of HOI on sea-salt resulted in production of
both IBr and ICl. On initial exposure to HOI, IBr was observed
to be the first product reaching the gas phase; IBr production
then decayed away while ICl grew and was subsequently
released continuously to the gas phase, as long as the sea-salt
surface was exposed to HOI. Clearly, bromine release is favored
over chlorine considering that the composition of sea-salt is 99.5
wt % Cl- compared to 0.2 wt % Br-.
The uptake coefficient for HOI on a fresh NaBr surface was
determined to be larger than the uptake coefficient for HOI on
a fresh NaCl surface. This suggests that the reaction of HOI
with Br- on sea-salt may be more rapid than with Cl-, leading
to preferential release of IBr. However, it was shown (results
section IIb) that ICl displayed a stronger chemical and physical
interaction with sea-salt than IBr. Therefore, even if ICl is
formed competitively with IBr, release of ICl may be slower
than IBr leading to an apparently slower reaction
Figure 10. Uptake of ICl (3.6 × 1011 molecules cm-3) on sea-salt:
Conversion of ICl to IBr as well as physical adsorption and desorption
of ICl from the surface.
HOI + Cl- f ICl(ad) + OH- f ICl(g) + OH- (R14)
HOI + Br- f IBr(g) + OH-
(R15)
Reaction of ICl with a NaBr surface (see results section IIa)
was also observed to form IBr. Thus, even if HOI forms ICl on
the sea-salt surface, ICl may be converted to IBr as long as
Br- is available for reaction. ICl is then only released into the
gas phase when most of the Br- on the surface has been depleted
HOI + Cl- f ICl(ad) + OH-
ICl(ad) + Br- f IBr(g) + Cl-
(R16)
(R17)
This is similar to the behavior observed by Fickert et al.30
for the interaction of HOBr and conversion of BrCl to Br2 in
sea-salt solutions. Their observations confirmed experimentally
the assumptions made by Huff and Abbatt31 and Kirchner et
al.,32 who speculated that BrCl was converted to Br2 on frozen
salt solutions as long as there is sufficient Br- available for
surface reaction.
Figure 11. Loss of I2 (7.6 × 1011 molecules cm-3) on a NaOH surface.
When IBr was exposed to either a NaCl or a sea-salt surface,
it did not produce any ICl and showed only a very weak physical
interaction with the sea-salt surface. This is consistent with the
higher thermodynamic stability of Cl- compared to Br- in NaX
crystals. The formation of ICl from IBr via the reaction R(18)
is therefore unlikely
desorbed from it again. γinitial was determined to be γ ) 0.0012
( 0.0004.
II. (c). ICl, IBr, and I2 on NaOH. The rates of reaction for I2,
ICl, and IBr on NaOH were all observed to be initially time
dependent but decreased quickly to a steady-state value. This
is shown for the reaction of I2 (7.6 × 1011 molecules cm-3) on
NaOH in Figure 11. No gas-phase products were observed. The
values for gamma were determined to be γ ) 0.0045 ( 0.0012,
γ ) 0.0034 ( 0.0009, γ ) 0.0030 ( 0.0008 for I2, ICl, and
IBr, respectively. Between 278 and 298 K, no temperature
dependence of the reaction was observed.
On addition of water vapor (11% relative humidity at 278
K), ksurface was measured to be much faster and the gamma
values increased to γ ) 0.016 ( 0.004 for I2 and γ g 0.1 for
ICl and IBr. As discussed in section Ie above this result is
consistent with the change from the solid NaOH surface to a
liquid layer at the higher relative humidity.
IBr(g) + Cl-(s) N ICl(g) +Br-
(R18)
(s)
II. Decrease in Surface ReactivitysThe Effect of Surface
“Aging”. A decrease in surface reactivity of dry salt surfaces
with time of exposure to a reactant molecule has already been
observed in many previous studies.20,21,33 In the present study,
a decrease in reactive uptake coefficient was observed with
increasing time of exposure of HOI to the substrate. This was
attributed to the formation of an involatile product, possibly
NaOH, on the surface thus reducing the number of surface
halogen ions available for reaction.
A summary of all uptake coefficients measured is shown in
Table 3. “No loss” means that it was not possible to detect any
loss of the reactant. All γ values reported are an average of at
least four experiments.
It was also shown that HOI reacted with an NaOH surface
but uptake was generally slower than on halide salts and no
gas-phase products were observed from the reaction. Thus, the
fall off in the reaction rate when halide ions are replaced by

5172 J. Phys. Chem. A, Vol. 105, No. 21, 2001
Mo¨ssinger and Cox
TABLE 3: Summary of All Uptake Coefficients Measured
surface
γ HOI
γ ICl
γ IBr
γ I2
sea-salt/fresh
sea-salt/aged
sea-salt/11% RH
NaBr/fresh
NaBr/aged
NaBr/11%RH
NaCl/fresh
0.061 ( 0.021
0.014 ( 0.004
0.02 ( 0.007
0.034 ( 0.009
0.008 ( 0.003
0.008 ( 0.003
0.016 ( 0.004
0.007 ( 0.003
0.008 ( 0.003
0.016 ( 0.004
0.0022 ( 0.0006
g 0.1
0.0012 ( 0.0004
0.0006 ( 0.0002
No loss
0.0068 ( 0.0018
no loss
no loss
no loss
no loss
no loss
NaCl/aged
NaCl/11%RH
NaOH/fresh
NaOH/aged
NaOH/11%RH
pyrex
0.0034 ( 0.0009
0.0030 ( 0.0008
0.0045 ( 0.0012
g 0.1
no loss
g 0.1
no loss
0.016 ( 0.004
no loss
0.0005 ( 0.0001
OH- can be qualitatively accounted for. In solution at high pH,
HOI is known to react in the following way34
HOI + OH- f H2O + OI-
(R19)
OI- then reacts further by instantaneous disproportionation34
-
3 OI- f 2I- + IO3 Kdis ) 1020
(R20)
With increasing amount of NaOH present, the surface would
become progressively basic and HOI could react to form iodide
and iodate on the surface. Further reaction to form volatile I2
would require an acidic surface and is therefore unlikely34
IO3- + 5I- + 6H+ f 3 I2 + 3H2O
(R21)
This would explain the absence of gas-phase products when
HOI was lost to the surface. This mechanism also allows, within
error, closure of the mass balance for the loss of HOI on an
aged salt surface.
ICl, IBr, and I2 were also observed to undergo reaction with
an NaOH surface. Again, no product formation was observed.
In basic solution, the following reactions have been observed34-36
Figure 12. Different behavior of NaNO3 (taken from Laux et al.40)
and NaOH on the NaCl surface.
on NaCl to have the same effect on the uptake coefficient. An
alternative explanation is given by the fact that both HNO3 and
HOI form involatile reaction products on the NaCl surface,
which may respond differently to the addition of water vapor.
For the reaction of HNO3 on NaCl, the involatile product,
NaNO3, is thought to be formed initially as a second layer on
the NaCl surface resulting in an aging effect similar to that
observed for the reaction of HOI on NaCl. However, the addition
of water is thought to result in the formation of a so-called quasi
liquid layer37 on top of the NaCl surface, which leads to
precipitation of NaNO3 crystals (deliquescence points for
NaNO3, NaCl, and mixtures of NaCl/NaNO3 are 74.3%, 75.3%,
and 67.6% relative humidity respectively38,39). This precipitation
process is thought to rupture the surface structure and expose
more Cl- for reaction with adsorbed HNO3 40. This solid
recrystallization process is less likely to happen for NaOH
(deliquescence points for NaOH and sea-salt are 7% and 71-
75% relative humidity respectively29); therefore, restoration by
adsorbed water of sites available for reaction may be less
efficient, leading to the observed lack of any water vapor effect
on the uptake coefficient. These two processes are shown in
Figure 12. If this mechanism for the behavior of NaOH on NaCl
surfaces is valid, NaOH adsorbed on an NaCl surface must react
differently to the addition of water vapor in comparison to the
reaction of a pure NaOH surface to the addition of water vapor.
The experimental results described in this work (see results
section Ie and IIc) showed that addition of water vapor to 11%
IBr + 2OH- f Br- + H2O + OI-
ICl + 2OH- f Cl- + H2O + OI-
I2 + 2OH- f I- + H2O + OI-
(R22)
(R23)
(R24)
OI- could then react further as described in R20 above.
III. Effect of Water Vapor. If the above reactions known
in solution apply to salt surfaces with adsorbed water, an effect
of relative humidity might be expected, if the amount of surface
adsorbed water was a critical factor. To test the effect of
adsorbed water, the relative humidity was increased in several
experiments.
Mochida et al.33 reported that the addition of water vapor
did not have an effect on the reaction of HOBr on crystalline
NaCl and KBr. They concluded that their negative experimental
results meant that there must have been sufficient quantities of
adsorbed water on the salt samples already to allow the surface
reactions involving ionic species to occur.
Work in this laboratory showed that added water vapor had
a significant effect on the uptake rate of HNO3 on crystalline
NaCl surfaces.21 These surfaces were identical to the NaCl
surfaces used in this study of HOI, where no dependence of
the uptake rate on water vapor was observed. If the assumption
of Mochida et al.33 was correct, it would mean that the reaction
of HOI on NaCl needs a lot less water than the reaction of HNO3

Reaction of HOI with Sodium Halide Salts
J. Phys. Chem. A, Vol. 105, No. 21, 2001 5173
relative humidity, i.e., above the deliquescence point for NaOH
(7% R. H.), led to a dramatic increase in the rate of reaction
for the uptake of HOI, ICl and IBr on a pure NaOH surface.
This supports the suggestion that reactions analogous to those
in the liquid phase are responsible for reaction on a solid surface.
In the case of NaOH adsorbed on a NaCl surface, a large
increase in the rate of reaction would only be expected if (a)
the deliquescence on this surface occurred at the same relative
humidity and (b) it led to newly available surface Cl- ions.
IV. Extent of Reaction of Br- and Cl- in the NaCl, NaBr,
and Sea-Salt Samples. To quantify the amount of Br- and Cl-
available for reaction at the salt surfaces used in this study, the
amounts of Br- and Cl- present on the top surface layer, based
on the geometrical surface area, as well as the total amount of
Cl- and Br- in the bulk sample were calculated.
The fraction of Br- in sea-salt is 0.2 wt % or 0.108 mol %
and assuming a well mixed salt, the number of Br- and Cl-
surface sites available at the top surface layer are 6.91 × 1011
cm-2 and 6.39 × 1014 cm-2, respectively. The total amount of
Br- and Cl- available per unit area for reaction can be calculated
from the total amount of salt present on the Pyrex tube insert
and the geometric area of the surface exposed. The total amount
of Cl- and/or Br- available was calculated to be ∼1022
molecules cm-2, ∼1020 molecules cm-2, and ∼1022 molecules
cm-2 (Cl-)/ ∼1018 molecules cm-2 (Br-) for NaCl, NaBr and
sea-salt surfaces, respectively.
The maximum amount of ICl and/or IBr released to the gas
phase after reaction of HOI with the NaCl, the NaBr, and the
sea-salt surfaces in our experiments was 3 × 1014 molecules
cm-2, 1 × 1016 molecules cm-2, and 3 × 1015 molecules cm-2
(Cl-)/2 × 1015 molecules cm-2 (Br-), respectively. It should
be noted that the amount of ICl and IBr released was time
dependent. Nevertheless, the calculated values showed that in
all cases only a small (,1%) fraction of the total Cl- and Br-
present in the sample had reacted. However, in the case of NaCl
less Cl- reacted than was present on the top surface layer,
whereas for the sea-salt surfaces, in some cases, more Cl- had
reacted than what was present on the top surface layer of the
substrate. In the case of NaBr and sea-salt, in most cases, more
Br- had reacted than was present on the top surface layer. This
might be due to the fact that a) the “real” surface areas of the
NaCl, NaBr, and sea-salt surfaces are different to the assumed
geometrical surface areas and/or (b) that the sea-salt was not
well mixed and Br- was concentrated at the surface.
Figure 13. Possible arrangements of salt grains distributed on the
cylindrical Pyrex insert.
interstitial spaces between hexagonal close packed spherical
granules stacked in layers, is not appropriate to account for the
difference between the geometric surface area and “real” surface
area available for reaction. Instead, the approach outlined below
has been chosen as an attempt to account for the difference
between geometric and “real” surface area available for reaction.
For the approximately one grain thick granular salt surfaces
used in this study the surface area of a simple layer of grains
on a surface depends on the arrangements of the grains on the
Pyrex tube surface. Figure 13 shows some possible arrangements
based on the distribution of the salt grains on the Pyrex insert
observed in photomicrographs of the NaCl crystallites taken in
this laboratory.21 Any of the arrangements illustrated in Figure
13 could be representative of the surface. If all the salt grains
were attached with one face to the Pyrex surface touching each
other without gaps as shown in Figure 13a, the geometric and
the real surface area would be approximately equal, i.e., the
correction factor is 1. Looking at the photomicrographs most
of the NaCl grains appear to be attached to the surface on their
edges or corners as shown in Figure 13b-c. As they are
distributed all around the glass surface, only the top faces of
the crystal cubes have to be considered. If all the salt grains
were attached with one edge to the Pyrex surface touching each
other without gaps as shown in Figure 13b, the real surface
area would be twice the geometric surface area, i.e., the
correction factor is 2.
V. Reactivity of the Salt Surfaces Studied. The uptake
coefficient for the reaction of HOI on pure NaCl was signifi-
cantly lower than for HOI on sea-salt. Similarly, de Haan and
Finlayson-Pitts41 observed a lower γ by an order of magnitude
for the uptake of HNO3 on pure NaCl compared to synthetic
sea-salt. Behnke and Zetsch42,43 also reported that reactions
generating chlorine atoms are accelerated on sea-salt aerosol
relative to NaCl aerosols. A possible explanation for this
observation is the increase in the number of defect sites in sea-
salt, caused by Br- dislocations and the presence of mineral
salt such as MgCl2.
VI. Geometric Surface Area vs “Real” Surface Area. The
correct determination of the actual surface area of salt surfaces
available for reaction has been a matter of much debate.41,44-47
Leu et al.47,46 applied a model to account for the surfaces of
the salt granules beneath the top layer. Other studies41,44,45,37
claimed that the model is not applicable in all cases and depends
not only on the type of surface, but also on the reactant used.
For the approximately one grain thick granular salt surfaces used
in this study, the model applied by Leu,46,47 which corrects for

5174 J. Phys. Chem. A, Vol. 105, No. 21, 2001
Mo¨ssinger and Cox
However, so far in both cases, the assumption was made that
there are no gaps between the salt particles on the Pyrex surface.
This does not appear to be the case looking at the photomicro-
graphs. The porosity, θ, i.e., the fraction of the geometric surface
that is void can be calculated from the bulk density, Fb (1.13 g
cm-3), and the true density, Ft (2.165 g cm-3), of NaCl
solutions and deliquescent salt aerosols, all interactions take
place in solution. It should also be noted that HOCl and HOBr
are both stronger acids (HOCl > HOBr > HOI with acid
dissociation constants being Ka ) 3 × 10-8, 2 × 10-9, 2 ×
10-11 respectively49), and are more thermally stable molecules
compared to HOI.
Huff and Abbatt31 investigated the reaction of HOCl on frozen
bromide-ice, chloride-ice, and bromide/chloride-ice surfaces
as a function of temperature and pH of the original solution.
At 233 K, a time dependent loss of HOCl on the bromide-ice
surface was observed with Br2 and BrCl being formed as the
products of the reaction. The reaction rate was independent of
bromine concentration (0.1-1%) at pH < 4. The uptake
coefficients reported were γ ) 0.051 ( 0.0013 at pH 2 and
0.014 ( 0.004 at pH 4-10. These values for the reactive uptake
coefficient for HOCl are within error the same as the uptake
coefficient for HOI in this study on a fresh NaBr surface and
an “aged” NaBr surface, respectively. Surface aging is thought
to occur due to poisoning of the dry salt surface possibly by
NaOH (see results section Ib and discussion section II, which
may have the same effect as changing the pH by adding NaOH
to the NaBr solution prior to freezing of halide-ice surfaces.
Huff and Abbatt31 observed no reaction of HOCl with the
chloride-ice surface or the bromide/chloride-ice surface at 233
K, but they observed a loss of HOCl to the bromide/chloride-
ice surface at 248 K with production of Br2 and BrCl. The
reaction was observed to be slower (γ ) 0.013 ( 0.004) than
in the experiments with the bromide-ice films. Huff and
Abbatt31 explain this 5-fold change of their gamma value on
the bromide/chloride-ice surface compared to the bromide-
ice surface with the fact that the chloride physically changes
the reactive surface and that thus less Br- is available for
reaction on the surface. The value of their γ is comparable to
the uptake coefficient of HOI on an aged sea-salt surface in
this study, where the amount of Br- available for reaction has
also been reduced.
Mochida et al.33 investigated the reaction of HOBr with solid
crystalline NaCl and KBr substrates. They observed γ to be
time and dose dependent and they quoted values of γ at a
vanishing low flow rate of HOBr of e 0.18 ( 0.04 for KBr
substrates and e (6.5 ( 2.5) × 10-3 for NaCl substrates. The
extrapolated value for γ for the reaction of HOBr on NaCl in
their study is comparable to the value for the uptake coefficient
of HOI on an aged NaCl surfaces in this study. On the other
hand, the extrapolated value for γ for the reaction of HOBr on
KBr is much larger than the values of γ for the reaction of
HOCl, HOBr, or HOI on a NaBr/KBr surface of any other study
discussed in this section. This might be explained by the low
concentrations used in Mochida et al.’s33 study (109-1011
molecules cm-3) compared to all other studies (1011-1013
molecules cm-3), but it is not clear why the gamma value should
be so different from the one they obtained for NaCl substrates
using the same HOBr concentrations.
θ ) 1-(1.13 g cm-3/2.165 g cm-3) ) 0.48
(E2)
This means that, on average, for every salt particle on the
surface, there is approximately one void area equivalent to that
of one particle on the surface. If the salt cubes are attached to
the Pyrex surface on their corners (Figure 13c), all six faces of
the cube would be exposed to HOI, but there is only half the
number of salt grains distributed on the geometrical surface area
if the voids are considered. The maximum correction factor used
to obtain the “real surface area” is thus 3.
These simple arrangements do not take into account possible
changes in the surface area or reactivity due to defect sites, such
as steps and edges, on the surface. The variation of the correction
factor from 1 to 3 stresses the importance to work with well
characterized substrates. A mean correction factor of 2 ( 1 may
be appropriate to apply to the observed γ values for application
to the dry sea-salt aerosol.
VII. Comparison with Previous Studies of the Uptake of
HOI on Salt Surfaces. Only one other study of the interaction
of HOI with salt surfaces has been published to date. Allanic et
al.20 reported values for the uptake coefficient of HOI on NaCl
and KBr of γ ) (4 ( 2) × 10-2 and γ ) (6 ( 2) × 10-2
respectively, which are in good agreement with the values
obtained on fresh salt surfaces in this study, considering the
experimental uncertainties. Allanic et al.20 also found that
gamma for HOI on KBr decreased with increasing residence
time in the Knudsen cell reactor, which is consistent with the
surface aging effects reported here.
However, Allanic et al.20 did not observe production of ICl
from the reaction of HOI on NaCl but found I2 to be the main
product. On the KBr surface, I2 was again the main product
with minor amounts of IBr. They concluded that the decomposi-
tion of HOI to I2 was the dominant process, which determined
the rate of uptake. In the present study, both ICl and IBr were
observed as products of the reaction on NaCl and NaBr, and I2
was not detected from either reaction. It should be noted that I2
is one of the major impurities produced as a side product during
the in-situ production of HOI. It is therefore difficult to
distinguish small changes in the I2 signal from its large
background signal. As described in the results section IV
production of I2 due to decomposition of HOI was only observed
within the outer sliding injector, where the concentration of HOI
was much higher than that in the flow reactor, hence favoring
a second-order surface decomposition reaction. Allanic et al.20
did not report the concentration of HOI used in their study, but
if the concentrations were higher decomposition may well have
been the controlling reaction.
For HOBr uptake on NaCl substrates Mochida et al.33
observed Br2 and BrCl as products of the reaction with a delay
in release of BrCl to the gas phase. They suggested that the
production of Br2 originated from the bimolecular self-reaction
of HOBr. For the reaction on KBr, they observed Br2 as the
sole product. Mochida et al.33 also reported a decrease in Br2
yield from 100% to 50% with increasing HOBr flow rate, which
they interpreted as a competition reaction between the hetero-
geneous reaction of HOBr with NaCl and KBr and the self-
reaction of HOBr. A heterogeneous bimolecular self-reaction
was also suggested as a possible reaction for HOI by Allanic et
VIII. Comparison with the Uptake of HOBr and HOCl
on Salt Surfaces. One study for the reaction of HOCl on frozen
salt surfaces31 and four studies of the uptake of HOBr on salt
surfaces,33 frozen salt surfaces,32 aqueous salt solutions,30 and
deliquescent salt aerosols50 have been reported in the literature.
There are significant differences between the four types of
substrates used. The surface of the frozen salt surfaces is
generally constantly renewed by water vapor added to the
experimental system to inhibit evaporation of the halide-ice
surface.31 The dry salt surfaces undergo an “aging process” e.g.,
for HOBr uptake.33 For uptake measurements on aqueous salt

Reaction of HOI with Sodium Halide Salts
J. Phys. Chem. A, Vol. 105, No. 21, 2001 5175
TABLE 4: Summary of Uptake Coefficients for HOI, HOCl, and HOBr on Different Types of Salt Surfaces
author
molecule
surface
NaBr
NaCl
sea-salt
NaBr
NaCl
sea-salt
KBr
NaCl
frozen NaBr
frozen NaBr
frozen sea-salt
salt solutions
frozen NaBr
frozen NaCl
frozen sea-salt
NaCl aerosol
pH
conca
aged?
g
this work
this work
this work
this work
this work
this work
allanic
allanic
huff
huff
huff
HOI
HOI
HOI
HOI
HOI
HOI
HOI
HOI
HOCl
HOCl
HOCl
HOBr
HOBr
HOBr
HOBr
HOBr
-
high
high
high
low
low
low
-
yes
yes
yes
no
no
no
no
no
-
0.008 ( 0.003
0.008 ( 0.003
0.014 ( 0.004
0.034 ( 0.009
0.016 ( 0.004
0.061 ( 0.021
0.06 ( 0.02
-
-
-
-
b
-
-
-
0.04 ( 0.02
low
high
low
lowc
-
high
high
high
-
0.051 ( 0.0013
0.014 ( 0.004
0.013 ( 0.004
-
-
fickert
-
g0.01
kirchner
kirchner
kirchner
abbatt
high
high
high
high
-
0.00327 ( 0.0048
0.00124 ( 0.0047
0.00142 ( 0.00017
e0.0015
-
-
-
-
high
-
a Low means less than 1 × 1012 molecules cm-3, high means up to 1013 molecules cm-3
.
b A dash (-) means not applicable or information not
available c The uptake was limited by diffusion, and Br2 release was only observed at low pH not at high pH. The data obtained by Mochida et
al.33 was not included and is discussed in the text.
al.20 (see discussion section VII) and, in this study, for very
high concentrations of HOI present (∼1013-1014 molecules
cm-3) in the outer sliding injector. It is surprising that the HOBr
would undergo a self-reaction at the low concentrations used
(109-1011 molecules cm-3) by Mochida et al.33 Abbatt48
reported in his study of the reaction of HOBr on ice surfaces
and HBr/HCl doped ice surfaces that second order kinetics were
Fickert et al.30 studied the activation of Br2 and BrCl via
uptake of HOBr onto aqueous salt solutions containing Br- and
Cl-. They found the reaction to be diffusion-limited with an
accommodation coefficient of R > 0.01. The uptake coefficient,
γ, must thus be larger or equal to 0.01 and this value is
comparable to the γ value obtained for HOI on aged sea-salt in
this study as well as to the value obtained by Huff and Abbatt31
for the uptake of HOCl on frozen sea-salt reported above.
Fickert et al.30 also reported that at a pH of less than 6.5 at
least 90% of HOBr taken up was released as gas-phase Br2.
No Br2 release was observed following HOBr uptake onto
nonacidified solution. Similar to what has been reported in the
study by Abbatt and Waschewsky,50 this confirmed the acid
catalyzed halogen activation mechanism from sea-salt aerosols
predicted by recent models of halogen chemistry in the marine
boundary layer.17,51
Overall, the three hypohalous acids, HOCl, HOBr, and HOI
show a very similar behavior on reaction with salt surfaces. In
all cases, the reaction of the hypohalous acid was observed to
be faster on Br- containing surfaces than on Cl- containing
surfaces. For the reaction of both HOBr and HOI on dry salt
surface, a surface aging effect was reported.20,33 In general, for
reactions on dry salt surfaces, frozen salt surfaces and deli-
quescent aerosols a decrease in the values of γ in a more basic
surface environment was observed. Within error the γ values
for the uptake of HOI, HOCl, and HOBr on the salts determined
on fresh surfaces and/or low pH and/or low reactant concentra-
tions (< 1 × 1012 molecules cm-3) were comparable. Similarly,
the γ values determined on an aged surface and/or high pH
observed only at concentrations > 5 × 1012 molecules cm-3
,
which he attributed to the self-reaction of HOBr.
Mochida et al.33 also speculated that the buildup of involatile
bases such as NaOH and KOH may be responsible for a slow
decrease of γ with time, which was observed at the higher flow
rates of HOBr. This speculation is in agreement with the
assumptions made in this work to explain the effect of surface
aging (see results section Ie and discussion section II).
Kirchner et al.32 studied the reaction of HOBr with NaCl,
NaBr, and sea-salt doped ice surfaces with γ ) (1.24 ( 0.47)
× 10-3, γ ) (3.27 ( 0.48) × 10-3 and γ ) (1.42 ( 0.17) ×
10-3 respectively. They observed Br2 and BrCl as products of
the reaction. In comparison with all other data discussed here
for HOCl, HOBr and HOI uptake on salt surfaces, their gamma
values are surprisingly low and only comparable to the value
reported by Abbatt and Waschewsky50 for the uptake of HOBr
on an unbuffered NaCl aerosol. The very high HOBr concentra-
tions (1013 molecules cm-3) used in their study might be an
explanation for the low gamma values observed, as the surface
would become less reactive very quickly.
Abbatt and Waschewsky50 investigated the heterogeneous
interaction of HOBr with deliquescent NaCl aerosols (75%
relative humidity) at room temperature. They reported a value
for γ e 1.5 × 10-3, γ g 0.2, and γ g 0.2 for the uptake of
HOBr on unbuffered aerosols, aerosols at pH 3 and buffered
aerosols at pH 7.2, respectively. Abbatt and Waschewsky50
explained the strong dependence of the HOBr reactivity on
acidity with the rapid depletion of hydrogen ions by the
relatively high concentrations of HOBr ((2-10) × 1012 mol-
ecules cm-3) used. They reported a lower reactivity at high pH,
which is in agreement with the observations made by Huff and
Abbatt,31 Mochida et al.,33 Fickert et al.,30 and in this study.
However, their reported upper limit for γ e 0.0015 is only
comparable to the γ values measured by Kirchner et al.32 on a
frozen NaCl, NaBr and sea salt surface and the reported lower
limit for γ g (6.5 ( 2.5) × 10-3 for the uptake coefficient for
the uptake of HOBr on a dry KBr surface, determined by
Mochida et al.33
and/or high reactant concentrations (> 1 × 1012 molecules cm-3
)
were comparable. This is summarized in Table 4. One exception
is the very high value (γ g 0.2) obtained for the uptake of HOBr
on deliquescent NaCl aerosols at low pH by Abbatt and
Waschewsky.50 This suggests that the γ values obtained for HOI
on the dry salt surfaces in this study are likely to be a lower
limit for application to marine aerosol particles. To provide a
more definite reactive uptake coefficient for atmospheric condi-
tions the uptake of HOI should also be measured on deliquescent
sea-salt aerosols.
Atmospheric Implications. The lifetime of HOI with respect
to photolysis at midday in the marine boundary layer, for Mace
Head, Ireland, July 1997 was calculated to be ∼3 min.52 At the
same location and time, the surface-to-volume ratio of the fine
fraction of the sea-salt aerosol was determined to be between
50 and 70 µm2 cm-3 52 Assuming a value of 60 µm2 cm-3 for
.

5176 J. Phys. Chem. A, Vol. 105, No. 21, 2001
Mo¨ssinger and Cox
the sea-salt aerosol surface-to-volume ratio and γ ) 0.02 for a
loss of HOI on a medium aged aerosol surface, the lifetime of
HOI with respect to loss on the aerosol surface was calculated
to be ∼4 h using the following equations
into the gas phase with some time delay, which was explained
by either conversion of ICl to IBr on the sea-salt surface or a
stronger physical interaction of ICl compared to IBr with the
sea-salt surface. The effect leading to decreased surface reactiv-
ity was attributed to formation of an involatile product of the
reaction of HOI on salt, possibly NaOH, on the salt surfaces.
The uptake coefficient for HOI on the salt surfaces did not show
a dependence on water vapor. It was speculated that this might
result from the fact that NaOH, adsorbed on the salt surface,
does not recrystallize to free additional surface sites for reaction
with Cl-, as suggested for NaNO3 formed in the reaction of
HNO3 on NaCl.40 It was also shown that more Br- was used
for the reaction of HOI on NaBr and sea-salt than is available
on the top surface layer of the substrate, but much less than is
available in total. A correction factor of 2 ( 1 was calculated
according to the possible arrangements of salt grains on the
surface, which made it possible to correct for the difference in
geometric surface area compared to the “real” surface area. The
uncertainty in the correction factor highlighted the importance
of using well characterized substrates. It was shown that salt
aerosols are not likely to be a significant sink for IX and that
Cl- and Br- may be released from salt aerosols in the marine
boundary layer.
$ S
4 V
khet ) γ
(E3)
1
khet
τ )
(E4)
where γ ) uptake coefficient for HOI, S/V ) the surface-to-
volume ratio of the sea-salt aerosol [cm2 cm-3], $ ) the average
molecular speed [cm s-1], khet ) rate constant for loss on the
aerosol [s-1] and τ ) the atmospheric lifetime of HOI with
respect to loss to sea-salt aerosols [s].
Thus, during the day photolysis is the major loss process for
HOI. During night time the uptake on sea-salt becomes the
dominant loss process.
The experimental results of this study show clearly that the
uptake of HOI on sea-salt leads to the formation of the
interhalogen species, ICl and IBr, via the reactions
HOI + NaCl f ICl + NaOH
HOI + NaBr f IBr + NaOH
(R5)
(R6)
Acknowledgment. The authors thank the U. K. Natural
Environmental Research Council for support for this research.
We also thank C. J. Percival for help with the double sliding
injector setup and with some of the data analysis of preliminary
experimental results.
The interhalogens formed are only slightly soluble and can be
released back into the gas phase, where they readily photolyze19
IX + hν f X + I (X ) Cl or Br)
(R7)
Note Added in Revision
These reactions are important as two halogen atoms are released
into the gas-phase per HOI consumed. The halogen atom
concentration in the atmosphere is therefore increased and gas-
phase iodine is recycled.
During the course of this work we learned of a similar study
on HOI/alkali halide salt interactions conducted at the Max-
Planck Institute for Chemistry in Mainz, Germany, which has
recently been submitted to PCCP. The results of the two studies
are in good agreement and show that IBr and ICl are released
as reaction products of the interaction of HOI with salt surfaces
and not I2 as previously reported in the literature.
If HOI concentrations in the marine boundary layer were
sufficiently large and the heterogeneous reaction with Cl- and
Br- in the sea salt aerosol was rapid enough, reaction R5-7
might provide a general mechanism for release of reactive Cl-
and Br- as well as a recycling mechanism for IX from the marine
aerosol (where IX ) total gaseous iodine species). Thus, the
marine aerosol would not act as a sink for IX. The reversible
nature of the interaction of gaseous HOI with the salt aerosols
could offer a reasonable explanation for the observed behavior
of radioactive iodine released to the atmosphere in the accident
at the nuclear power plant in Chernobyl in 1986. I131 partly
remained in gas phase reservoirs and partly in the condensed
phase for several days, after release, in an apparent reversible
process, leading to enhanced dry deposition in of gaseous I131
over a wide region.15,16
References and Notes
(1) Chameides, W. L.; Davis, D. D. J. Geophys. Res. 1980, 85, 7383-
7393.
(2) Davis, D.; Crawford, J.; Liu, S.; McKeen, S.; Bandy, A.; Thornton,
D.; Rowland, F.; Blake, D. J. Geophys. Res. 1996, 101, 2135-2147.
(3) Jenkin, M. E.; Cox, R. A.; Candeland, D. E. J. Atmos. Chem. 1985,
2, 359-375.
(4) Carpenter, L. J.; Sturges, W. T.; Penkett, S. A.; Liss, P. S.; Alicke,
B.; Hebestreit, K.; Platt, U. J. Geophys. Res. 1999, 104, 1679-1689.
(5) Klick, S.; Abrahamsson, K. J. Geophys. Res. 1992, 97, 12 683-
12 687.
(6) Laturnus, F.; Wiencke, C.; Kloeser, H. Mar. EnViron. Res. 1995,
41, 169-181.
(7) Schall, C.; Heumann, K. G. Fresenius J. Anal. Chem. 1993, 346,
717-722.
(8) Singh, H. B.; Salas, L. J.; Stiles, R. E. J. Geophys. Res. 1983, 88,
3684-3690.
(9) Alicke, B.; Hebestreit, K.; Stutz, J.; Platt, U. Nature 1999, 397,
572-573.
(10) Allan, B. J.; McFiggans, G.; Plane, J. C. J. Geophys. Res. 2000,
105, 14,363-14,239.
(11) Jenkin, M. E. The Photochemistry of Iodine Containing Compounds
in the Marine Boundary Layer; AEA-EE-0405, UKAEA, Harwell Labora-
tory, 1992.
However, the general mechanism of IX recycling and Br-
and Cl- release described above is also strongly dependent on
the nature of the salt aerosol. Marine aerosols vary significantly
in size and are liquid in most cases. In addition, the HOI
concentrations used in this study are several orders of magnitude
higher than the atmospheric concentration of HOI (∼1 ppt). The
measurements performed in this study using high HOI concen-
trations on solid surfaces are therefore likely to provide a lower
limit for the reactive uptake coefficient of HOI on submicron
marine aerosol particles.
(12) Wayne, R. P. Chemistry of Atmospheres, 2nd ed.; Oxford Science
Publications: New York, 1991.
Conclusions
(13) Bauer, D.; Ingham, T.; Carl, S. A.; Moortgat, G. K.; Crowley, J.
N. J. Phys. Chem. A 1998, 102, 2857-2864.
(14) Rowley, D. M.; Mo¨ssinger, J. C.; Cox, R. A. J. Atmos. Chem. 1999,
34, 137-151.
This study has shown that reaction of gaseous HOI with NaCl,
NaBr, and sea-salt occurs with release of gaseous ICl, IBr, and
ICl + IBr, respectively. On the sea-salt surface ICl was released

Reaction of HOI with Sodium Halide Salts
J. Phys. Chem. A, Vol. 105, No. 21, 2001 5177
(15) Pacyna, J. M.; Johansen, O.; Salbones, J.; Semb, A. Air RadioactiV-
ity at Selected Stations in Norway after the Chernobyl Reactor Accident;
NILU, 1986.
(16) Clark, M. J.; Smith, F. B. Nature 1988, 332, 245-249.
(17) Vogt, R.; Crutzen, P. J.; Sander, R. Nature 1996, 383, 327-330.
(18) Vogt, R.; Sander, R.; Glasow von, R.; Crutzen, P. J. J. Atmos. Chem.
1999, 32, 375-395.
(19) Seery, D. J.; Britton, D. J. Phys. Chem. 1964, 68, 2263-2266.
(20) Allanic, A.; Rossi, M. J. J. Geophys. Res. 1999, 104, 18 689-
18 696.
J.; Nyholm, R.; Trotman-Dickenson, A. F., Eds.; Pergamon Press: Oxford,
1973; Vol. 2; pp 1191-1401.
(35) Johnson, D. W.; Margerum, D. W. Inorg. Chem. 1991, 30, 4845-
4851.
(36) Troy, R. C.; Kelley, M. D.; Nagy, J. C.; Margerum, D. W. Inorg.
Chem. 1991, 30, 4838-4845.
(37) Beichert, P.; Finlayson-Pitts, B. J. J. Phys. Chem. 1996, 100,
15 218-15 228.
(38) Tang, I. N.; Munkelwitz, H. R. Atmos. EnViron. 1993, 27, 467-
473.
(21) Davies, J. A.; Cox, R. A. J. Phys. Chem. A 1998, 102, 7631-
7642.
(39) Tang, I. N.; Munkelwitz, H. R. J. Appl. Meterol. 1994, 33, 791-
796.
(22) Percival, C. J.; Mo¨ssinger, J. C.; Cox, R. A. PCCP 1999, 1, 4565-
4570.
(40) Laux, J. M.; Fister, T. F.; Finlayson-Pitts, B. J.; Hemminger, J. C.
J. Phys. Chem. 1996, 100, 19 891-19 897.
(23) Monks, P. S.; Stief, L. J.; Tardy, D. C.; Liebman, J. F. J. Phys.
Chem. 1995, 99, 16 566-16 570.
(41) de Haan, D. O.; Finlayson-Pitts, B. J. J. Phys. Chem. A 1997, 101,
9993-9999.
(24) Loumis, R. A., Klaassen, J. J., Lindner, J., Christopher, P. G.; Leone,
S. R. J. Chem. Phys. 1997, 106, 3934-3947.
(42) Behnke, W.; Zetzsch, C. J. Aerosol Sci. 1990, 21, S229-S232.
(43) Behnke, W.; Scheer, V.; Zetzsch, C. Production of Photolytic
Precursors of Atomic Cl from Aerosols and Cl- in the Presence of Ozone;
Kluwer Academic Publishers: Dordrecht, 1995.
(44) Fenter, F. F.; Caloz, F.; Rossi, M. J. J. Phys. Chem. 1994, 98, 9801-
9810.
(25) Gilles, M. K.; Turnipseed, A. A.; Talukdar, R. K.; Rudich, Y.;
Villalte, P. W.; Huey, L. G.; Burkholder, J. B.; Ravishankara, A. R. J. Phys.
Chem. 1996, 100, 14 005-14 015.
(26) Brown, R. L. J. Res. Nat. Bur. Stand. 1978, 83, 1-8.
(27) Adams, J. W.; MPI-Mainz, Germany, private communication 2000.
(28) Rudich, Y.; Talukdar, R. K.; Imamura, T.; Fox, R. W.; Ravishan-
kara, A. R. Chem. Phys. Lett. 1996, 261, 467-473.
(29) Twomey, S. J. Appl. Physics 1953, 24, 1099-1102.
(30) Fickert, S.; Adams, J. W.; Crowley, J. N. J. Geophys. Res. 1999,
104, 23 719-23 727.
(45) Fenter, F. F.; Caloz, F.; Rossi, M. J. J. Phys. Chem. 1996, 100,
1008-1019.
(46) Leu, M.-T.; Timonen, S. R.; Keyser, F. L. J. Phys. Chem. 1995,
99, 13 203-13 212.
(47) Leu, M.-T.; Timonen, R. S.; Keyser, F. L. J. Phys. Chem. A 1997,
101, 278-282.
(31) Huff, A. K.; Abbatt, J. P. D. J. Phys. Chem. A 2000, 104, 7284-
7293.
(48) Abbatt, J. P. D. Geophys. Res. Lett. 1994, 21, 665-668.
(49) Greenwood, N. N.; Earnshaw, A. The Halogens. In Chemistry of
the Elements, 1st ed.; Pergamon Press: Oxford, 1994.
(50) Abbatt, J. P. D.; Waschewsky, G. C. G. J. Phys. Chem. A 1998,
102, 3719-3725.
(32) Kirchner, U.; Benter, T.; Schindler, R. N. Ber. Bunsen-Ges. Phys.
Chem. 1997, 101, 975-977.
(33) Mochida, M.; Akimoto, H.; van den Bergh, H.; Rossi, M. J. J. Phys.
Chem. A 1998, 102, 4819-4828.
(51) Sander, R.; Crutzen, P. J. J. Geophys. Res. 1996, 101, 9121-9138.
(52) McFiggans, G.; Plane, J. M. C.; Allan, B. J.; Carpenter, L. J.; Coe,
H.; O’Dowd, C. J. Geophys. Res.-Atmos. 2000, 105, 14 363-14 369.
(34) Adams, C. J.; Downs, A. J. Chlorine, Bromine, Iodine, Astatine.
In ComprehensiVe Inorganic Chemistry, 1st ed.; Bailar, J. C.; Emeleus, H.