Br2 production from the heterogeneous reaction of gas-phase OH with aqueous salt solutions: Impacts of acidity, halide concentration, and organic surfactants

Article abstract of DOI:10.1021/jp063165o

This study reports the first laboratory measurement of gas-phase Br ₂ production from the reaction between gas-phase hydroxyl radicals and aqueous salt solutions. Experiments were conducted at 269 K in a rotating wetted-wall flow tube coupled to a chemical-ionization mass spectrometer for analysis of gas-phase components. From both pure NaBr solutions and mixed NaCl/NaBr solutions, the amount of Br₂ released was found to increase with increasing acidity, whereas it was found to vary little with increasing concentration of bromide ions in the sample. For mixed NaCl/NaBr solutions, Br₂ was formed preferentially over Cl₂ unless the Br ^- levels in the solution were significantly depleted by OH oxidation, at which point Cl₂ formation was observed. Presence of a surfactant in solution, sodium dodecyl sulfate, significantly suppressed the formation of Br₂; this is the first indication that an organic surfactant can affect the rate of interfacial mass transfer of OH to an aqueous surface. The OH-mediated oxidation of bromide may serve as a source of active bromine in the troposphere and contribute to the subsequent destruction of ozone that proceeds in marine-influenced regions of the troposphere.

Full text of DOI:10.1021/jp063165o

10456
J. Phys. Chem. A 2006, 110, 10456-10464
Br2 Production from the Heterogeneous Reaction of Gas-Phase OH with Aqueous Salt
Solutions: Impacts of Acidity, Halide Concentration, and Organic Surfactants
†
Elizabeth K. Frinak and Jonathan P. D. Abbatt*
Department of Chemistry, 80 Saint George Street, UniVersity of Toronto, Toronto, Ontario M5S 3H6, Canada
ReceiVed: May 23, 2006; In Final Form: July 17, 2006
2
This study reports the first laboratory measurement of gas-phase Br production from the reaction between
gas-phase hydroxyl radicals and aqueous salt solutions. Experiments were conducted at 269 K in a rotating
wetted-wall flow tube coupled to a chemical-ionization mass spectrometer for analysis of gas-phase components.
From both pure NaBr solutions and mixed NaCl/NaBr solutions, the amount of Br
2
released was found to
increase with increasing acidity, whereas it was found to vary little with increasing concentration of bromide
-
ions in the sample. For mixed NaCl/NaBr solutions, Br
levels in the solution were significantly depleted by OH oxidation, at which point Cl
Presence of a surfactant in solution, sodium dodecyl sulfate, significantly suppressed the formation of Br
2
was formed preferentially over Cl
2
unless the Br
2
formation was observed.
2
;
this is the first indication that an organic surfactant can affect the rate of interfacial mass transfer of OH to
an aqueous surface. The OH-mediated oxidation of bromide may serve as a source of active bromine in the
troposphere and contribute to the subsequent destruction of ozone that proceeds in marine-influenced regions
of the troposphere.
1
. Introduction
concentrated brines, frozen seawater, frost flowers, marine
aerosol, or snow upon which aerosol has deposited. Through
Modeling studies and field measurements have shown ozone
both experiments and models, possible gas-surface interactions
that have been identified so far include ozone-mediated release
of Br2 on frozen seawater and aqueous sodium bromide
depletion near the Earth’s surface can be positively correlated
to elevated concentrations of inorganic bromine.
1
-7
This
relationship is strongly demonstrated in the Arctic springtime
boundary layer and in regions close to salt lakes at lower
latitudes. O3 depletion in the overall marine boundary layer may
also be occurring, but at levels that are harder to observationally
quantify.
1
4
15
-
solutions, NOx-bromide interactions, interactions of HSO5
1
6
with bromide, and, last, the interactions of hydroxyl radical
formed either photochemically in the particle
from the gas phase.
1
7-19
or taken up
18,20
Of all these, only the chemistry involving
gas-phase hydroxyl radical has not been experimentally exam-
ined, largely because of the particularly difficult nature of the
experiments.
There are now well-established mechanisms by which gas-
7
phase, free-radical bromine catalytically destroys O3; it is
known that autocatalytic release of active bromine can occur
from sea salt via the uptake of HOBr;3 and there is a viable
mechanism for the recycling of active bromine once it is
formed. Of particular note is that depletions in the amount of
bromide, relative to the ratio of bromide to sodium in seawater,
have been reported in snow and aerosol samples in the Arctic
This study builds upon previous work involving bromide
oxidation via the generation of aqueous OH, produced either
through photolysis of dissolved nitrate or nitrite or by pulsed
,5
8
,9
17-19,21,22
19
radiolysis.
In particular, the work of Zafiriou is pivotal
in pointing out that condensed-phase OH is believed to
preferentially oxidize bromide over chloride in seawater solu-
tions. However, the question we address in this paper is different
because the oxidative processes involving OH and bromide will
occur at the interface of the solutions under study, as discussed
later. In particular, although OH is known to be taken up
10-13
as well as lower latitudes.
Also, measurements by Newberg
1
0
et al. show that bromide was depleted even in marine aerosol
with low anthropogenic influence. The reported depletions were
most significant in the size fraction < 1.6 µm, which cor-
responded to the fraction with the highest acidity. These results
are indicative of bromine release and cycling in the troposphere,
but the initial processes responsible for liberating the bromine
from these surfaces remain poorly quantified in either absolute
or relative terms.
There are a number of candidates for this initiation process,
each of which may prevail under different environmental
conditions. Gas-phase photolysis of organobromine compounds
has generally been considered to be too slow, in particular for
the case of bromoform. Instead, most schemes focus on sea salt
as the original source of gas-phase bromine, in the form of
2
3
efficiently by aqueous solutions, it has yet to be shown that
such interfacial processes will lead to bromide oxidation; instead,
existing experimental studies in the literature have only focused
on bulk chemistry arising from the in situ production of OH
17-19,21,22
within the aqueous samples.
The likelihood of signifi-
cant surficial chemistry comes from a suite of modeling studies
that show that halide anions and hydroxyl radicals have a strong
2
4-27
affinity for the surface.
The atmospheric motivation for the specific focus on OH as
an initiator for Br2 release comes from the observation that BrO
formation in the boundary layer is usually associated with
sunrise, either on a diurnal scale, as in the Dead Sea region, or
seasonally, as in the Arctic. Whether this dependence arises only
because the autocatalytic cycle involving HOBr uptake is driven
*
To whom correspondence should be addressed. Phone: (416) 946-
7
358. E-mail: jabbatt@chem.utoronto.ca.
†
Present address: University of New Hampshire, Durham, NH.
1
0.1021/jp063165o CCC: $33.50 © 2006 American Chemical Society
Published on Web 08/17/2006

Gas-Phase OH Reaction with Aqueous Salt Solutions
J. Phys. Chem. A, Vol. 110, No. 35, 2006 10457
Figure 1. Schematic of the experimental apparatus.
TABLE 1: Molar Concentrations of Salt Samples and the
Corresponding Abbreviation Used in the Text
center section of Figure 1. The sample insert effectively acts as
the flow volume within a fixed flow tube. The Teflon tube
conveying the gas-phase components to the chemical-ionization
mass spectrometer (CIMS) also does not rotate. The purpose
of rotating the sample insert was to minimize exposed glass
surface within the flow tube and to replenish the sample surface
with bromide as it is oxidized away. The entire glass sample
insert rotated during experiments at a speed of approximately
10 revolutions per minute (rpm). A uniform sample coating was
attained by rinsing the glass insert with a 5% solution of HF
and with water, prior to conducting an experiment. The solution
was confined by a ridge in the glass insert to the cold region
within the flow tube. The temperature of the sample was held
at 269 K using a Neslab LT50 low-temperature bath circulator
by flowing chilled ethylene glycol through the cooling jacket.
The temperature of the flow tube was calibrated using a copper-
[NaBr]
[NaCl]
abbreviation
3
4
2
.9
<0.02
3.4
3.4
CSB
SWSB
DSB
-
-
3
4
.9 × 10
.5 × 10
by sunlight or whether the initiation step also requires sunlight
is not known. However, it is known that dark processes alone
are not able to provide sufficient levels of active bromine to
destroy ozone in their own right.
In this paper we seek to answer the following questions: Is
gas-phase bromine formed from the interaction of gas-phase
OH with solutions containing bromide? If so, what is the yield
of Br2 formed relative to OH lost? What is the dependence of
the production rate on the acidity, bromide content, and chloride
content of the solutions? And what is the impact of an organic
surfactant present in the solution upon the Br2 production rate?
3
constantan thermocouple. A humidified flow of 1100-1600 cm -
-
1
3
-1
(
STP) min N2 combined with 100-250 cm (STP) min Ar
sustained a relative humidity (RH) of approximately 78% with
respect to liquid water at 269 K to ensure samples remained in
2
. Experimental Section
the aqueous phase. For reference, the deliquescence RHs of
Experiments were conducted in a horizontally oriented,
NaCl and NaBr are 75 and 58%, respectively.29,30 The corre-
rotating, wetted-wall flow tube shown schematically in Figure
, which is similar to a flow reactor described previously in
the literature. To characterize the influence of concentration
of bromine production, samples of varying bromide concentra-
tions were prepared. The most concentrated samples were 3.9
M NaBr (Sigma-Aldrich, 99.0%), abbreviated in the text as CSB
sponding bulk flow velocity ranged between 93 and 143 cm/s.
We estimate the relative change in composition of the solution
by either uptake or loss of water vapor is only 0.03%/min. This
slow rate of change is expected to have a negligible impact on
the experiments in this study.
The schematic in Figure 1 is divided into three sections
labeled microwave discharge cavity, rotating wetted-wall flow
tube, and CIMS. The first region highlights the precursors for
the hydroxyl radicals. OH was formed from the following
reaction:
1
2
8
(concentrated sodium bromide). The most dilute samples were
3
.4 M NaCl (ACP Chemicals, 99.0%), which contained an
-4
impurity of 2.5 × 10 M NaBr as reported by the manufacturer.
These samples will be referred to as DSB (dilute sodium
bromide) in the text. Samples were also prepared containing a
-
-
[Cl ]:[Br ] molar ratio of approximately 700:1, which is similar
H + NO f OH + NO
(1)
2
to the ratio of 620:1 found in seawater. Samples of this
concentration will be condensed to SWSB (seawater sodium
bromide) within this text. Table 1 provides a summary of the
sample concentrations in units of moles per liter with the
corresponding abbreviations. The acidity was increased using
H2SO4 (Fisher, 96.0%), and the pH of the solutions measured
using an Orion 520A pH meter calibrated using buffers at pH
3
-1
Using an MKS flow controller, 105-250 cm (STP) min argon
(BOC Gases, UHP) containing trace amounts of hydrogen gas
(BOC Gases, UHP) was flowed through a Beenakker microwave
discharge cavity. The resulting hydrogen atoms were conveyed
into the flow tube through a Teflon tube of 0.318 cm o.d. and
107 cm length. A 3.4 L glass bulb attached to a manifold
1
.00, 4.00, and 7.00.
contained a mixture of approximately 70:1 N (BOC Gases,
2
Following preparation, a 3 mL aliquot of salt solution was
UHP):NO2 (Matheson). The manifold terminated in a metering
placed in the glass sample insert (1.5 cm i.d., by 30 cm length)
in the flow tube as indicated by the gray shaded area in the
valve which delivered the N2:NO2 mixture into a 0.635 cm o.d.
Teflon tube connected to the moveable glass injector tube, also

10458 J. Phys. Chem. A, Vol. 110, No. 35, 2006
Frinak and Abbatt
of 0.635 cm o.d. Given the concentrations used and the speed
of reaction 1, the NO2 then reacted with the hydrogen atoms
per reaction 1 to rapidly form OH at the tip of the injector rod.
Experiments were carried out in two different modessone in
which NO2 was the limiting reagent and one in which NO2 was
present in excesssfor reasons to be discussed below.
During experiments, the glass injector rod was pulled back
to expose OH to the aqueous salt solution in the rotating wetted-
wall flow tube. Any resulting gas-phase products, as well as
the reactants, were carried downstream of the flow tube and
-
monitored using CIMS. SF6 was chosen as the reagent ion
because it enabled us to monitor NO2, OH, Br2, and Cl2
simultaneously.31,32 In particular, a trace amount of SF6 (Scott
Specialty Gases, 99.99%) was carried in an N2 flow through a
radioactive polonium source (2 Po) to generate SF6 reagent
10
-
3
3
ions. The ion source region is described in detail elsewhere.
The total N2 flow through the ion source region was 10.5 sLm.
-
After reaction of the species of interest with the SF6 reagent
ion, ions continued to a declustering region in the mass
spectrometer, followed by mass selection in a quadrupole and
finally detection with a Channeltron electron multiplier operated
in a single negative ion counting mode. Again, further details
of the system may be found in Thornton et al.33
Figure 2. (Top panel) Correlation between Br
2
production (b, dotted
line, right axis) and OH loss (2, solid line, left axis) as a function of
injector position during a typical experiment. The data shown are for
a CSB sample of pH 0.5. (Bottom panel) Comparison of OH loss on
CSB at pH 5.4 (4, dashed line) and pH 0.5 (2, solid line). Bars for
bromine in the top panel and for OH in the bottom panel represent the
standard deviation in the averaged signal at the corresponding injector
position.
Calibrations of the NO2 and Br2 signals were performed to
convert the resulting CIMS count rates, reported as counts per
second (cps), into concentrations. For both species, the con-
centration was quantified by measuring the change in pressure
in the manifold over time. When OH was formed with NO2 as
the limiting reagent, the OH concentration was determined by
the change in the NO2 signal between when the Beenakker
microwave cavity discharge was on and off. This quantification
of OH represents an upper limit as it was assumed the entire
suppression in the NO2 signal was due to a 1:1 conversion of
NO2 to OH. The Br2 signal was calibrated in a similar manner
as the NO2 signal using a 3.4 L bulb filled containing
approximately 300:1 N2:Br2. Gas-phase Br2 was transferred from
a 100 mL bulb containing liquid bromine (ACP Chemicals) after
successive freeze-pump-thaw cycles into a 3.4 L bulb where
it was mixed with N2.
constant signals of SF , NO , and OH with the glass sample
6
2
insert rotating, the samples were exposed to OH. In the mode
with NO as the limiting reagent, the injector tube was pulled
2
back, incrementally exposing the sample at a rate of 2 cm every
10 min. When NO was present in excess, the injector was
2
retracted in one step, exposing the entire 29 cm of sample.
3. Results
The principal qualitative observation reported in this paper
is that gas-phase Br2 formation occurs when OH is exposed to
bromide-containing solutions, providing they are sufficiently
acidic. The top panel of Figure 2 illustrates typical behavior
where the loss of OH signal is plotted coincident with the
formation of Br2. This particular experiment is for Br2 produc-
tion from 3 mL of a CSB sample of pH 0.5. Injector distance
correlates to the length of sample exposed to OH. The OH signal
shown is the count rate as monitored at a mass-to-charge ratio
of 17, and the Br2 signal is shown in the same units as those
detected at a mass-to-charge ratio of 160. Fits to the data are
single-exponential decay and growth, respectively. SWSB
samples of pH 0.5 were also characterized using this method
and exhibited similar behavior. This is direct experimental
evidence that molecular bromine can be formed by the
heterogeneous interaction of gas-phase OH with halide solutions.
It should be noted that blank experiments were conducted to
ensure the Br2 signal was due to reaction with OH. These
experiments consisted of exposing samples with the microwave
off but with H2 and NO2 flowing, or in the absence of NO2 but
with the microwave on. In both cases, no Br2 was measured. In
addition, one SWSB experiment was performed with HCl
acidification in place of H2SO4. Br2 formation was observed,
indicating that the oxidation chemistry was not being initiated
10
Detection limits for NO2 and Br2 were 9.6 × 10 and 5.6 ×
10
3
1
0 molecules/cm , respectively, for an integration time of 30
s and a S/N ) 1. Note that these detection limits were
compromised to a significant degree by our decision to detect
OH, Br2, NO2, and Cl2 simultaneously through the use of the
-
-
SF6 as the reagent ion. In particular, SF6 is known to react
34
with water vapor. To maintain reasonable detection limits and
roughly 50 000 cps of reagent ion signal, we operated the flow
tube at a sufficiently low temperature, 269 K, that the water
vapor effect was not overwhelming. Similarly, since the goal
of the experiment was not to measure uptake coefficients of
OH to the solutions, the pressure in the flow tube was kept
reasonably high, 94 Torr, where the CIMS gives the best
sensitivity but plug flow conditions do not prevail. Typical OH
starting concentrations were quite high, with an upper limit of
1
2
3
2
× 10 molecules/cm .
A typical experiment consisted of adding a 3 mL sample to
the system and cooling it to 269 K over a period of 2 h. The
flow tube was then opened to a mechanical pump, and the flows
of humidified N2, Ar, and the CIMS N2 were established. The
relative humidity of the system was typically 78% with respect
to liquid water with the pressure in the flow tube at about 94
Torr as measured by a Baratron capacitance manometer. Then
the system was opened to the declustering, mass selection, and
detection regions of the mass spectrometer. Upon establishing
-
by the reaction of OH with HSO4 .
In the bottom panel of Figure 2 is the OH decay rate on a
pH 5.4, CSB solution. This behavior is consistent in replicate
measurements. Clearly, the decay is significantly slower with

Gas-Phase OH Reaction with Aqueous Salt Solutions
J. Phys. Chem. A, Vol. 110, No. 35, 2006 10459
the high-pH sample, and there is no observed Br2 formation.
Care should be taken in interpreting kinetic behavior from the
curves because the conditions were not in the plug flow regime;
that is, the role of gas-phase diffusion and mixing in this
experiment are sufficiently undefined at the relatively high
pressures of the flow tube therefore dominating over the kinetics
of the gas-surface interaction. In particular, turbulent mixing
may be an important factor in the transportation of OH to the
wall. Nevertheless, it is clear that OH is not lost on solutions
near neutral pH as efficiently as on acidic solutions than would
occur if molecular diffusion was the only mixing process.
Quantitatively, experiments were conducted to determine the
bromine yield, defined as the concentration of Br2 produced
divided by the concentration of OH initially present. Using a
35
method from gas-phase, free-radical, flow-tube kinetics studies,
Figure 3. Transition from Br
production (dotted line, right axis) in a sample of CSB at pH 0.5.
2 2
production (solid line, left axis) to Cl
these experiments were done with H atoms in excess. Experi-
mentally, we could verify that NO2 was the limiting reagent by
observing the NO2 signal dropping from roughly 30 cps to very
close to zero cps when the microwave discharge containing H2
was switched on. As previously mentioned in the Experimental
Section, the quantification of OH strictly represents the upper
limit to the concentration. However, given the response of the
NO2 signal to the addition of H2, we believe that the vast
majority of the change in the NO2 signal gives rise to OH
formation.
A second issue related to the calculation of the Br2 yield is
whether the OH formed all reacts on the surface of the solution
or whether it can decay away to some degree in the gas phase
also. The gas-phase reactions of potential importance include
the following:
numerical model described above that this effect is on the order
of about 30% loss.
Figure 3 shows data from the exposure to OH of a DSB
-
-
sample at pH 0.5. For comparison, the [Cl ]:[Br ] molar ratio
in this solution was 13 600:1. In this experiment OH was
produced with NO2 as the limiting reagent. However, the injector
rod was pulled back in one step, exposing the entire 29 cm of
sample at once. The data show that despite the very large excess
of chloride ions, Br2 is evolved before Cl2. Only after the Br2
signal has peaked and begun to decrease does the Cl2 peak grow
in, illustrating the extreme preference oxidants have for bromide
over chloride in solution. As discussed in more detail in section
4
.1 (Reaction Mechanism), this observation is consistent with
24-27
OH + NO + M f HNO + M
(2)
both the proposed surface segregation of bromide ions
and
2
3
37,38
the known bulk chemistry.
During this experiment we looked
OH + NO + M f HONO + M
(3)
(4)
for gas-phase BrCl production but saw no signal above the
CIMS detector noise.
OH + OH f H O + O
The second class of experiments involved measuring the Br2
production rate as a function of solution acidity and bromide
content. Given experimental considerations, we chose to perform
these experiments with NO2 in excess; the NO2 signal would
decrease when the microwave discharge was switched from off
to on, but not to zero. The reason for this transition was that
the conditions in excess H were hard to achieve, being dependent
on the somewhat variable H-atom output of the discharge.
Instead, we found that much more reproducible conditions were
maintained between runs using excess NO2. Slight variations
in the conditions were quantitatively accounted for as described
below.
2
OH + OH + M f H O + M
(5)
(6)
2
2
OH + Br f HOBr + Br
2
The observed first-order rate constant for OH loss on the acidic
solution in Figure 2 is on the order of 44 s
-
1,
which is
significantly greater than the estimates of the rate constants for
reactions 2-5 at the point of addition of OH to the flow tube.
However, reaction 6 between OH and Br2 requires further
consideration. We estimate the importance of this reaction using
a simple numerical model that includes OH loss at the wall
giving rise to 0.5 Br2 molecules per OH molecule lost and gas-
Figure 4 summarizes the Br2 concentrations formed from
exposure to a nominally constant amount of OH, in experiments
performed with NO2 in excess at a concentration of roughly 1
36
phase reaction of OH with Br2. We find that 84% of the overall
-
1
13
3
OH loss of 44 s is due to reaction with aqueous bromide to
produce Br2 with the remainder arising from subsequent reaction
directly with Br2. And so, we feel confident that the primary
loss process for OH is indeed via reaction with the aqueous
surface and that the difference in reactivity on the acidic and
neutral solutions is mainly due to additional OH reactivity on
the former.
× 10 molecules/cm . The concentration of Br2 produced is
shown as a function of pH. The fit is a single exponential to
guide the eye for CSB samples, but similar trends are seen for
the samples of lower bromide concentrations as denoted by the
other symbols. Though the DSB samples exhibit slightly less
Br2 production, there is no detectable difference between the
more concentrated samples. Each data point represents an
average of results conducted from roughly four replicate
experiments, and the error bars indicate the standard deviation
of those measurements. Though Br2 was never measured in
samples of pH 5.4, it is possible that Br2 was released at a
concentration below the instrumental limit of detection of 5.6
The results for the bromine yield experiments at pH 0.5 are
.3 ( 0.1 for SWSB and 0.7 ( 0.5 for CSB samples. The
0
uncertainty is one standard deviation for precision only for four
and five replicate experiments, respectively. For the reasons
outlined above, we consider these to be lower limits to the
bromine yields. An additional factor that arises in this regard is
that Br2 yields can be suppressed to some degree by reaction 6
with OH. For the data in Figure 2, we estimate using the
1
0
3
× 10 molecules/cm . The main point to note is that Br2
production increases dramatically with increasing acidity but
does not depend as strongly on bromide concentration.

10460 J. Phys. Chem. A, Vol. 110, No. 35, 2006
Frinak and Abbatt
pressure decrease with time of the manifold connected to the
NO2 reservoir.
Finally, given the possibility that organic films may reside
3
9,40
on the surface of marine particles,
we attempted to observe
formation of Br2 from pH 0.5, DSB solutions with variable
concentrations of a common surfactant, sodium dodecyl sulfate
(
1
SDS), present. These solutions reproducibly yield (4 ( 2) ×
16
0 molecules of Br2 during 20 min of exposure to OH without
the SDS present. The reported uncertainty is the standard
deviation for five replicate measurements. However, the maxi-
mum Br2 level observed was 12 ( 3% of this value when SDS
concentrations were varied over 2 orders of magnitude from
0
.18 to 2 mM. The low value of this concentration range is
4
1
close to the critical micelle concentration, although full data
on this quantity are not available for the properties of our
solutions at the temperature and pH of the experiments.
Nevertheless, we feel confident in saying that the SDS is able
to significantly hinder the production of bromine, most probably
because OH reacts instead with the surfactant layer present on
the solution surface.
2
Figure 4. Variation in Br production as a function of pH. Markers
correspond to CSB (b), SWSB (O), and DSB (/). Bars indicate standard
deviation determined from replicate measurements. Curve is a single-
exponential fit to CSB data to guide the eye.
4
. Discussion
.1. Reaction Mechanism. The chemistry of OH in bulk
4
sodium bromide solutions has been studied in the laboratory
previously through early work involving the pulsed radiolysis
technique and more recently by investigations involving pho-
tolysis of dissolved OH precursors such as nitrate or hydrogen
1
7-19,21,22
42
peroxide
as well as modeling studies. Oxidation is
initiated by the reaction of OH with bromide, which proceeds
3
8
at approximately the diffusion-controlled rate:
-
-
Br + OH T HOBr
(7)
-
-
HOBr can either dissociate to re-form Br and OH, react with
-
-
Figure 5. Variation in Br
2
production as a function of bromide
a proton to form Br, or react with Br to form Br2 :
concentration. Data shown are for samples of pH 0.5. Bars indicate
standard deviation from replicate measurements.
-
+
HOBr + H f Br + H O
(8)
(9)
2
Figure 5 depicts the variation in Br2 production with bromide
concentration in samples of pH 0.5. The concentration of Br2
is somewhat reduced in the DSB samples, but given the 4 orders
of magnitude change in the bromide concentrations, there is
remarkably little difference between the bromide production
rates for the solutions studied. The results clearly show that a
sample with a similar chloride-to-bromide ratio as seawater
yields the same quantity of Br2 as a sample highly concentrated
in bromide.
-
-
-
-
HOBr + Br f OH + Br
2
Reactions 8 and 9 are keys because both ultimately lead to Br2
formation via rapid chemistry:
-
-
Br + Br T Br
(10)
(11)
(12)
2
Br2- + Br2- T Br3- + Br-
For all of the data in Figures 4 and 5, the experiments were
conducted under nominally the same conditions, with NO2 in
excess, constant flow conditions, and roughly the same flow
rate of H2 through the microwave discharge. However, given
that the production rate of H atoms may have varied from
experiment to experiment, the change in NO2 signal with the
microwave discharge on and off was noted for each experiment.
We used this information to normalize the bromine signals for
variable OH production rates in each experiment; the OH signal
itself was not used in this regard because of variable wall loss
of OH in the Teflon tube connecting the flow tube to the CIMS.
Similarly, because the overall CIMS sensitivity also varied to
some degree from experiment to experiment, either due to the
Br3- T Br + Br
-
2
Thus, the overall reaction for formation of Br2 from OH is
-
+
2OH + 2Br + 2H f Br + 2H O
(13)
2
2
where it is seen that protons are needed for reaction. We see
that the bromine yields of 0.3 ( 0.1 for SWSB and 0.7 ( 0.5
for CSB samples measured for the most acidic solutions are
consistent with the overall stoichiometry of reaction 13.
For the specific reaction between OH and the pure CSB, we
first consider whether the chemistry will occur at the solution
interface or not. Assuming for the time being that there is no
preferential segregation of bromide ions to the surface, the
reactodiffusive length for OH in the solution is calculated to
be extremely short, only 1.5 Å, suggestive of a surface
210
slow decay of the Po source or due to variable SF6 flow rates,
the sensitivity of the instrument to NO2 was measured for each
experiment. This was achieved using the CIMS signal for NO2,
as measured with the microwave discharge off, and from the
concentration of NO2 in the flow tube, as calculated from the
4
3
interaction:

Gas-Phase OH Reaction with Aqueous Salt Solutions
J. Phys. Chem. A, Vol. 110, No. 35, 2006 10461
0
.5
0.5
D
D
L )
)
(14)
(
I
)
(
II
-
)
k
k [Br ]
where we have used a bulk-aqueous-phase diffusion constant
-
5
2
-
of 1 × 10 cm /s, a bulk concentration of 3.9 M Br , and the
10
-1 -1
literature rate constant of 1.1 × 10
M
s
for reaction 7.
Considering that both bromide and OH are believed to partition
to the surface of solutions,2
4-27
we are even more confident
-
-
that OH will interact with Br at the surface, forming HOBr .
This conclusion is consistent with molecular dynamics studies
by Roeselova et al.25 that predict frequent interactions between
surface adsorbed hydroxyl and halide anions.
To evaluate the subsequent chemistry, we apply the steady-
-
state approximation to the concentration of HOBr in the above
chemistry, assuming bulk processes prevail. In particular, we
can write an expression proportional to the rate of formation of
oxidized bromine:
Figure 6. Measured variation in Br
bromide concentration for samples of pH 0.5 as shown as closed circles
b) on the left axis compared to the modeled rate of oxidized bromine
formation on the right axis. Crosses (+) represent a production
mechanism that incorporates reaction 9 between HOBr and Br ,
whereas the open circles (O) represent a mechanism that does not
include reaction 9. See section 4.1 (Reaction Mechanism) for details.
The right axis is a dimensionless value for the normalized rate of
oxidized bromine formation.
2
production as a function of
(
+
-
-
-
(
k [H ] + k [Br ])
8
9
-
rate )
k [OH][Br ] (15)
7
+
-
[
]
k-7 + k [H ] + k [Br ]
8
9
38,44
For experimental concentrations and literature rate constants,
the quantity within the large brackets in eq 15 has a negligible
pH dependence, as depicted in Figure 6 by crosses on the right-
hand vertical axis. The reason for this trend is that reaction 9
proceeds sufficiently fast at high bromide concentrations that a
rapid route for Br2- formation remains at all pH values. Whereas
this expression is consistent with the pH dependence for bromide
oxidation observed in previous kinetics studies,2 we note these
past studies were not performed with such high bromide
concentrations as in our CSB solutions. However, if we assume
that reaction 9 does not proceed, than the quantity in the brackets
displays a distinct pH dependence, shown as open circles in
Figure 6.
1,22
While it is true that the rate constants used to model the
solution kinetics are those for room temperature and measure-
ments were performed at 269 K, the temperature dependence
for such fast ionic processes is usually quite weak. Instead, we
propose the apparent absence of the fast reaction between
Figure 7. Plot of normalized yields of both Br (b) and Cl (0) as a
function of the bromide concentration from a numerical model
simulation of the aqueous-phase chemistry. The model was initialized
2
2
-
-
HOBr and Br is an indication that the chemistry occurring
in the bulk does not prevail in the same manner at the interfaces
of this solution. One possibility is that cations and anions are
strongly associated with each other in this high ionic strength
environment, perhaps in the form of contact-ion pairs. And so,
-
+
-6
with [Cl ] ) 3.9 M, [H ] ) 0.32 M (pH ) 0.5), and [OH] ) 10 M,
and run for 100 s. Room-temperature rate constants were taken from
those referenced in the text and from Ershov et al. In the model, OH
initially interacts with Cl , given its high relative concentration, but
2
Cl is only produced in high yield after the [Cl ]:[Br ] ratio exceeds
37
-
-
it is more likely for the transient HOBr species to interact with
-
-
5
a hydrated proton at the surface, via the surface equivalent of
reaction 8, than with a bromide ion as in (9). This suggestion
is supported by recent surface spectroscopy studies that indicate
protons are also preferentially found at the surface of halide
solutions.45
A previous study has proposed a halogen production surface
mechanism that proceeds without the involvement of a proton.
Specifically, Oum et al.14 observe formation of Cl2 during the
3 × 10 . For reference, the value of this ratio in seawater is 620:1.
relatively high fluxes of OH to the solution surfaces are
removing protons faster than they are able to diffuse from the
bulk to replenish those lost to reaction. In this manner it may
be true that at the lower OH concentrations in the atmosphere,
the acidity dependence displayed in Figure 4 will not be as
strong.
2
54 nm photolysis of aqueous sodium chloride particles, O3,
Turning now to the solutions that contain trace amounts of
sodium bromide in concentrated sodium chloride, the primary
findings are that the bromine yield and the bromine production
rates are similar for the pure CSB solutions, despite orders of
magnitude less bromide in solution. This indicates a very strong
preference for oxidation of bromide over chloride at the
interface, as indicated in earlier work by Zafiriou.19
It is reasonable to ask whether the chemistry is necessarily
the same in the pure bromide solutions as in the mixed chloride/
bromide solutions. In particular, again assuming for the time
being that bulk chemistry prevails and that there is no
and water vapor. Given the strong dependence of our Br2
production rates on acidity, we see no evidence within our
detection limits that indicates an analogous process occurs with
bromine.
A final point we make concerning the acidity dependence is
that the decreased yield of bromine at high pH is consistent
with the decreased loss rate for OH shown in the bottom panel
of Figure 2. Apparently, if the transient HOBr surficial species
cannot find a proton to react with, it dissociates and releases
OH back to the gas phase. Another point to consider is that the
-

10462 J. Phys. Chem. A, Vol. 110, No. 35, 2006
Frinak and Abbatt
preferential concentration of bromide at the surface relative to
chloride, one would expect that the incoming OH would much
more likely interact with a chloride ion than a bromide ion:
the surface during this time. We saw no production of gas-
phase BrCl which may be due to dissolved BrCl being readily
converted to Br2, rapid loss of ClBr via reaction 24, or a lower
-
sensitivity of the CIMS to BrCl.
-
-
OH + Cl T HOCl
(16)
Although bulk-phase chemistry can explain the preference
for oxidation of bromide over chloride with OH, another
intriguing possibility is that the bulk-phase model just presented
is not correct. Given that bromide might partition to the surface
preferentially over chloride, it is also possible that the chemistry
We suggest this because reaction 16 proceeds at the same
diffusion-controlled rate as does reaction 7, but the concentration
of chloride will be much higher than bromide. Similarly, one
would expect that HOCl- would then react with a proton
according to the diffusion-limited reaction 17,
-
is not initiated by reaction of OH with Cl but instead with
-
Br . In this case the subsequent chemistry would then more
-
+
closely resemble that described above for the pure sodium
bromide solutions. The relative rates of the two mechanisms
may be better discerned when experimental measurements of
the ratio of chloride to bromide at the surface of a mixed sodium
halide solution are available. At this point, we can only say
that our observations are consistent with both bulk-phase and
surface-phase chemistry; thus we feel no need to introduce new
chemical pathways to explain the data.
HOCl + H f H O + Cl
(17)
2
but not with chloride given the very slow rate constant for
reaction 18:46
-
-
HOCl + Cl f Cl + H O
(18)
2
2
Once the Cl atom is formed from reaction 17, it will complex
with Cl to form Cl2 , again via a very fast reaction:
-
-
4.2. Surfactant-Coated Solutions. We believe the observa-
tion that the SDS surfactant substantially reduces the Br2
production rate is direct experimental evidence that organic
monolayers can affect the uptake of OH by aqueous aerosol
surfaces. It is likely that a reaction is occurring between OH
and the SDS surfactant, as was experimentally demonstrated
-
-
Cl + Cl f Cl2
(19)
At this point, a competition arises between processes that might
give rise to gaseous Cl2 production or gaseous Br2 production,
represented by the following reactions. In particular, Cl2 can
form via
4
7
by Cooper and Abbatt, where the loss of OH on solid
ammonium sulfate surfaces was shown to be enhanced when
exposed to a surfactant, 1-hexanol. It is also consistent with
Cl2- + Cl2- f Cl - + Cl
-
31
the later study of Bertram et al. that demonstrated that OH
(20)
(21)
3
readily reacts with a number of organic surfaces. In addition to
Cl3- T Cl + Cl-
recent laboratory measurements of N2O5 aqueous aerosol
2
48-50
kinetics from our laboratory and others,
where the loss rate
-
of gas-phase N2O5 is suppressed with organic surfactants such
as humic substances, SDS, and hexanoic acid present in the
aerosol, we believe this result provides additional evidence that
organic surface layers can have a major impact on the rates of
gas-surface chemistry that may occur in the atmosphere.
Via a set of very fast reactions, Cl2 can also initiate the
-
-
oxidation of Br to either Br2 or Br, which ultimately will
lead to Br2 formation:
Cl2- + Br f ClBr + Cl
-
-
-
(22)
(24)
(24)
4
.3. Atmospheric Implications. To assess the potential
-
-
-
2
-
atmospheric importance of the chemistry we have observed, we
can summarize by noting that the low level of bromide present
in marine aerosol does not disfavor the production of bromine
relative to chlorine with high yields, even for dilute bromide
solutions. Although it is true that we only observe bromine
production from a pH of approximately 3 or below, there is
likely to be some bromine production at higher pH with yields
below the detection limit of our experiment. It is also possible
that the decreased yield at higher pH was due to depletion of
protons at the surface of the solution due to the relatively high
OH fluxes used in this work. At the lower OH concentrations
in the atmosphere the yield may remain high under less acidic
conditions.
ClBr + Br f Br + Cl
-
-
ClBr f Cl + Br
In a simple numerical model using the above chemistry as
well as additional, less important reactions listed in Ershov et
37
al., we have compared the relative rates of formation of Cl2
to those of Br2 formation in a mixed salt solution of 3.9 M
NaCl and varying NaBr concentrations from an initial quantity
of dissolved OH. The results are plotted in Figure 7 as a function
of the molar concentration ratio in solution. It is seen that there
is a very strong preference for the oxidation of bromide rather
than chloride, all the way down to a concentration of roughly
-
5
-
-
-
However, given that a proton is clearly involved in the
oxidation, it is most likely that the OH-mediated release of
bromine will be important on acidic marine particles such as
those that have experienced significant levels of pollution
including HNO3 or SO2. Particles in the remote marine boundary
layer also have their pH lowered from the value prevailing in
ocean water by the aqueous-phase oxidation of background SO2,
but the rate at which this occurs and the degree of resultant
acidity is generally lower than in the polluted case.
10
M Br which corresponds to a [Cl ]:[Br ] molar ratio on
5
the order of 3 × 10 .
Thus, bulk chemistry alone can explain the results displayed
in Figure 3, where we see molecular bromine formed in
preference to chlorine for the DSB solutions with only impurity
levels of sodium bromide present. Similarly, we see from Figure
that molecular chlorine is only formed by exposure to OH
once bromide is depleted in solution. Indeed, we estimate that
roughly 4.5 × 10 bromide ions are present in the solution
exposed to OH. To within uncertainties, this is equal to the flux
of OH to the surface during 25 min of exposure time for the
experiment in Figure 3, assuming all OH present in the flow
3
17
As indirect support of the pH dependence of bromine
activation processes in general, we can consider a study by Tas
4
et al., where the formation of bromine oxide (BrO) was
12
3
tube at a concentration of 2 × 10 molecules/cm is lost on
monitored over the Dead Sea. Production of BrO was found to

Gas-Phase OH Reaction with Aqueous Salt Solutions
J. Phys. Chem. A, Vol. 110, No. 35, 2006 10463
vary diurnally, as expected, with the concentrations exceeding
the instrumental limit of detection at the earliest around 2 h
after sunrise. In particular, the concentration of BrO was found
to increase with decreasing pH of the Dead Sea water. Also,
the delay between sunrise and the appearance of measurable
amounts of BrO increases with increasing pH. The results from
this study suggest that the mechanism for Br2 production is pH-
dependent, as expected with OH-bromide oxidation chemistry,
as well as with the HOBr-mediated autocatalytic cycle.
in solution are sufficiently depleted. Given the rapid interactions
-
-
that occur between OH and both Cl and Br in solution, it is
likely that the chemistry occurs at the surface of these solutions,
even in the absence of any partitioning of halides to the surface.
Indeed, bulk chemistry can explain the very strong preference
for oxidation of bromide over chloride. However, it cannot
explain the observed pH dependence of the bromine production
rates in the concentrated CSB solutions, suggestive that
interfacial chemistry is occurring differently from that in the
bulk. The strong acidity dependence of the bromine yields
indicate that this chemistry will be most important in air that
has been subjected to anthropogenic influence. We show that
this process may be viable as a source of active bromine in the
atmosphere, although future work remains to better assess its
significance relative to other sources. Our finding that a surface
coating of an organic surfactant, SDS, significantly reduces
bromine production reinforces the need to better determine the
degree to which such coatings exist on atmospheric particulates.
This work clearly shows that an organic surfactant can
significantly affect the rate of interfacial mass transfer of OH
radicals to an aqueous aerosol.
To be more quantitative, it will be important to compare the
rate of this chemistry to that of other potential bromide activation
processes, such as the heterogeneous uptake of gas-phase N2O5,
NO3, or O3, plus the condensed-phase processes involving
photolysis of nitrate in solution (to form OH) and the oxidation
-
51
of bromide by HSO5 . For example, Hunt et al. have
investigated the heterogeneous reaction of O3 with deliquesced
NaBr aerosol. Upon extrapolation to atmospheric conditions,
they concluded this reaction may serve as a source of between
0
.003 and 22 ppt Br2 in 10 h of darkness, with the range
determined by how their experiments on pure sodium bromide
particles are extrapolated to seawater composition.
For comparison, we can do an initial calculation to illustrate
the potential importance of OH-mediated bromine release. The
In the future, we intend to extend these studies to the
formation of Br2 from solid surfaces, such as snow and ice that
contain halides, as well as to solid salts. We believe the former
needs to be investigated in order to determine the potential of
this chemistry as a bromine source in the polar boundary layer,
where active bromine levels are observed to increase with polar
sunrise. Similarly, the chemistry on solid salts may be important
in salt lake environments.
1
8
calculation that follows is similar to those of Matthew et al.
and Herrmann et al.,20 which both also claim that this chemistry
may represent an important source of active bromine. We use
-
11
3
the same marine aerosol volume loading of 6.3 × 10
cm /
3
3
cm as that employed by Sander and Crutzen, which was in
turn used by Matthew et al.18 and Herrmann et al. For a
20
-
4
monodisperse distribution of particles with radius 1.5 × 10
3
cm, this corresponds to 4.5 particles/cm and a total aerosol
Acknowledgment. This work was funded by NSERC. We
thank Ingrid George and Cort Anastasio for showing us a
preprint of their work prior to publication and for help with the
numerical model of solution chemistry.
-
6
2
3
surface area of 1.3 × 10 cm /cm . For particles of this size
at atmospheric pressure, there is a significant gas-phase diffusion
mass-transfer limitation, with the resistances to uptake and
diffusion being comparable to each other for uptake coefficients
of roughly 0.1. And so, for the calculations below we use an
uptake coefficient of 0.05 for two reasons. One, there will not
be much limitation to mass transfer by gas-phase diffusion for
this value. Two, assuming that the uptake coefficient is unity
in our work at pH ) 0.5 and, as described above, that the
decrease in the Br2 yield as a function of pH is due to a smaller
uptake coefficient, we estimate the value on the pH 2.9 solutions
to be roughly 0.05.
References and Notes
(1) Foster, K. L.; Plastridge, R. A.; Bottenheim, J. W.; Shepson, P.
B.; Finlayson-Pitts, B. J.; Spicer, C. W. Science 2001, 291, 471.
(
(
(
2) Platt, U.; H o¨ nninger, G. Chemosphere 2003, 52, 325.
3) Sander, R.; Crutzen, P. J. J. Geophys. Res. 1996, 101, 9121.
4) Tas, E.; Peleg, M.; Matveev, V.; Zingler, J.; Luria, M. J. Geophys.
Res. 2005, 110, doi: 10.1029/2004JD005665.
(5) Vogt, R.; Crutzen, P. J.; Sander, R. Nature 1996, 383, 327.
(
6) von Glasow, R.; von Kuhlmann, R.; Lawrence, M. G.; Platt, U.;
Crutzen, P. J. Atmos. Chem. Phys. 2004, 4, 2481.
7) Barrie, L. A.; Bottenheim, J. W.; Schnell, R. C.; Crutzen, P. J.;
Rasmussen, R. A. Nature 1988, 334, 138.
8) Abbatt, J. P. D.; Nowak, J. B. J. Phys. Chem. A 1997, 101, 2131.
And so
(
(
[
OH] γVA
g
rate of oxidized Br production )
(25)
(9) Fan, S.-M.; Jacob, D. J. Nature 1992, 359, 522.
(10) Newberg, J. T.; Matthew, B. M.; Anastasio, C. J. Geophys. Res.
005, 110, doi: 10.1029/2004JD005446.
4
2
(
11) Pszenny, A. A. P.; Moldanov a´ , J.; Keene, W. C.; Sander, R.; Maben,
where V is the mean molecular speed, A is the surface area of
aerosol per unit volume of air, and γ is the reactive uptake
J. R.; Martinez, M.; Crutzen, P. J.; Perner, D.; Prinn, R. G. Atmos. Chem.
Phys. 2004, 4, 147.
(12) Simpson, W. R.; Alvarez-Aviles, L.; Douglas, T. A.; Sturm, M.;
Domine, F. Geophys. Res. Lett. 2005, 32, doi: 10.1020/2004GL021748.
6
coefficient. For an OH concentration of 2 × 10 molecules/
3
cm , the yield of oxidized bromine using the values mentioned
(
13) Toom-Sauntry, D.; Barrie, L. A. Atmos. EnViron. 2002, 36, 2683.
3
3
above is 2.0 × 10 /(cm s), which corresponds to 0.3 pptv/h at
atmospheric pressure. Given that values of a few pptv to tens
of pptv of active bromine are required to be atmospherically
significant, this simple calculation illustrates this chemistry may
be important as not only an initiation step that precedes the
(14) Oum, K. W.; Lakin, M. J.; DeHaan, D. O.; Brauers, T.; Finlayson-
Pitts, B. J. Science 1998, 279, 74.
(15) Finlayson-Pitts, B. J.; Livingston, F. E.; Berko, H. N. Nature 1990,
43, 622.
3
(
(
16) Mozurkewich, M. J. Geophys. Res. 1995, 100, 14199.
17) George, I. J.; Anastasio, C. Atmos. EnViron., to be submitted for
-
autocatalytic HOBr cycle but also as an active bromine source
publication.
(18) Matthew, B. M.; George, I.; Anastasio, C. Geophys. Res. Lett. 2003,
in its own right.
3
0, doi: 10.1029/2003GL018572.
(
(
19) Zafiriou, O. C. J. Geophys. Res. 1974, 79, 4491.
20) Herrmann, H.; Majdik, Z.; Ervens, B.; Weise, D. Chemosphere
5
. Summary
2
003, 52, 485.
In this work we demonstrate that a heterogeneous reaction
occurs between gas-phase OH and sodium halide solutions, to
yield gas-phase Br2. Cl2 is only formed when the bromide levels
(21) Matheson, M. S.; Mulac, W. A.; Weeks, J. L.; Rabani, J. J. Phys.
Chem. 1966, 70, 2092.
(22) Zehavi, D.; Rabani, J. J. Phys. Chem. 1972, 76, 312.

10464 J. Phys. Chem. A, Vol. 110, No. 35, 2006
Frinak and Abbatt
(
23) Hanson, D. R.; Burkholder, J. B.; Howard, C. J.; Ravishankara, A.
R. J. Phys. Chem. 1992, 96, 4979.
24) Ghosal, S.; Shbeeb, A.; Hemminger, J. C. Geophys. Res. Lett. 2000,
7, 1879.
25) Roeselov a´ , M.; Jungwirth, P.; Tobias, D. J.; Gerber, R. B. J. Phys.
Chem. B 2003, 107, 12690.
26) Roeselov a´ , M.; Vieceli, J.; Dang, L. X.; Garrett, B.; Tobias, D. J.
J. Am. Chem. Soc. 2004, 126, 16308.
27) Zangmeister, C. D.; Turner, J. A.; Pemberton, J. E. Geophys. Res.
Lett. 2001, 28, 995.
(38) Finlayson-Pitts, B. J. Chem. ReV. 2003, 103, 4801.
(39) Ellison, G. B.; Tuck, A. F.; Vaida, V. J. Geophys. Res. 1999, 104,
11633.
(40) Gill, P. S.; Graedel, T. E.; Weschler, C. J. ReV. Geophys. 1983,
21, 903.
41) Chatterjee, A.; Moulik, S. P.; Sanyal, S. K.; Mishra, B. K.; Puri,
P. M. J. Phys. Chem. B 2001, 105, 12823.
42) Thomas, J. L.; Jimenez-Aranda, A.; Finlayson-Pitts, B. J.; Dabdub,
D. J. Phys. Chem. A 2006, 110, 1859.
43) Hanson, D. R.; Ravishankara, A. R.; Solomon, S. J. Geophys. Res.
994, 99, 3615.
44) Neta, P.; Huie, R. E.; Ross, A. B. J. Phys. Chem. Ref. Data 1988,
7, 1027.
(
2
(
(
(
(
(
(
(
(
(
(
28) Hanson, D. R.; Lovejoy, E. R. J. Phys. Chem. 1996, 100, 6397.
29) Cziczo, D. J.; Abbatt, J. P. D. J. Phys. Chem. A 2000, 104, 2038.
30) Rockland, L. B. Anal. Chem. 1960, 32, 1375.
1
(
1
31) Bertram, A. K.; Ivanov, A. V.; Hunter, M.; Molina, L. T.; Molina,
(
(
45) Petersen, P. B.; Saykally, R. J. J. Phys. Chem. B 2005, 109, 7976.
46) Grigorev, A. E.; Makarov, I. E.; Pikaev, A. K. High Energy Chem.
M. J. J. Phys. Chem. A 2001, 105, 9415.
32) Huey, L. G.; Hanson, D. R.; Howard, C. J. J. Phys. Chem. 1995,
9, 5001.
33) Thornton, J. A.; Braban, C. F.; Abbatt, J. P. D. Phys. Chem. Chem.
Phys. 2003, 5, 4593.
(
1987, 21, 99.
9
(47) Cooper, P. L.; Abbatt, J. P. D. J. Phys. Chem. 1996, 100, 2249.
(
(48) Badger, C. L.; George, I. J.; Griffiths, P. T.; Abbatt, J. P. D.; Cox,
R. A. J. Phys. Chem. A 2006, 110, 6986.
(49) McNeill, V. F.; Patterson, J.; Wolfe, G. M.; Thornton, J. A. Atmos.
Chem. Phys. Discuss. 2006, 6, 17.
(
(
(
34) Arnold, S. T.; Viggiano, A. A. J. Phys. Chem. A 2001, 105, 3527.
35) Howard, C. J. J. Phys. Chem. 1979, 83, 3.
36) 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. JPL Publ. 1997, 97-4.
(50) Thornton, J. A.; Abbatt, J. P. D. J. Phys. Chem. A 2005, 109, 10004.
(51) Hunt, S. W.; Roeselov a´ , M.; Wang, W.; Wingen, L. M.; Knipping,
E. M.; Tobias, D. J.; Dabdub, D.; Finlayson-Pitts, B. J. J. Phys. Chem. A
2004, 108, 11559.
(37) Ershov, B. G.; Kelm, M.; Gordeev, A. V.; Janata, E. Phys. Chem.
Chem. Phys. 2002, 4, 1872.