Temperature and pressure dependence study of the reaction of IO radicals with dimethyl sulfide by cavity ring-down laser spectroscopy

Article abstract of DOI:10.1021/jp0345147

The reaction of IO radicals with dimethyl sulfide was studied using cavity ring-down laser spectroscopy. The reaction rate constant shows both a temperature and pressure dependence. At 100 Torr total pressure, the reaction has reached its high-pressure limit and has a rate constant of (2.5 ± 0.2) × 10^-13 molecule^-1 cm³ s ^-1 at 298 K. On the basis of the Arrhenius plot in the region of 273-312 K, the reaction has a negative activation energy (E_a = -18.5 ± 3.8 kJ mol^-1). The atmospheric implications of these findings are discussed. In light of these new data, DMS oxidation by IO can compete with oxidation by the hydroxyl radical in the marine boundary layer. Quoted uncertainties are one standard deviation from regression analysis.

Full text of DOI:10.1021/jp0345147

J. Phys. Chem. A 2003, 107, 6381-6387
6381
Temperature and Pressure Dependence Study of the Reaction of IO Radicals with Dimethyl
Sulfide by Cavity Ring-Down Laser Spectroscopy
†
Yukio Nakano, Shinichi Enami, Shinji Nakamichi, Simone Aloisio, Satoshi Hashimoto, and
Masahiro Kawasaki*
Department of Molecular Engineering, Kyoto UniVersity, Kyoto 606-8501, Japan
ReceiVed: February 27, 2003; In Final Form: June 4, 2003
The reaction of IO radicals with dimethyl sulfide was studied using cavity ring-down laser spectroscopy. The
reaction rate constant shows both a temperature and pressure dependence. At 100 Torr total pressure, the
-
13
-1
3 -1
reaction has reached its high-pressure limit and has a rate constant of (2.5 ( 0.2) × 10 molecule cm s
at 298 K. On the basis of the Arrhenius plot in the region of 273-312 K, the reaction has a negative activation
energy (E ) -18.5 ( 3.8 kJ mol ). The atmospheric implications of these findings are discussed. In light
-
1
a
of these new data, DMS oxidation by IO can compete with oxidation by the hydroxyl radical in the marine
boundary layer. Quoted uncertainties are one standard deviation from regression analysis.
1
. Introduction
chemically converted to active iodine by photolysis to produce
1
6
the I atom followed by reaction with ozone (O3) to produce
The oxidation of dimethyl sulfide (DMS) is of particular
17
IO. A large fraction of iodine is thought to be in its active
importance to the production of atmospheric aerosols, especially
in the marine boundary layer (MBL) because DMS is the main
sulfur-containing species emitted from the oceans. DMS is a
relatively reduced form of sulfur and is produced biologically.
When oxidized, however, the organic sulfur species become
hygroscopic and may form condensation nuclei, leading to the
production of aerosols and possibly clouds. The effects of
aerosols and clouds remain as the largest uncertainty in climate
forecasting today. For this reason, the understanding of DMS
oxidation is important to the understanding of our atmosphere.
18,19
form (i.e. IO and I), due to the fast rate of photolysis of HOI,
a major reservoir species for iodine. Less is known about iodine
chemistry than the chemistry of the other natural halogen
species, chlorine and bromine. A more comprehensive explana-
tion of IO studies, with special attention given to the IO radical
1
2
20
self-reaction, can be found a recent work by Bloss et al.
3
21,22
At one time, it was thought that DMS oxidation by IO,
4
IO + DMS f I + DMSO
(1)
is an important reaction in the atmosphere. The measured rate
Many aspects of these processes remain incompletely under-
-
11
3
constant in those works was on the order of 10
cm
5
stood, however, including the oxidation of DMS. For example,
a recent work by James et al. comparing measured and modeled
results at a site in Mace Head, Ireland, show that the daytime
oxidation rate of DMS is underestimated by models by over a
factor of 3. Nighttime oxidation of DMS is thought to be
-1 -1
molecule s . Those measurements were found to be in error,
and the current recommended value of this rate constant is 1.2
6
-
14
3
-1 -1
11
×
10
cm molecule
s
at room temperature, with the
temperature dependency of this reaction remaining unknown.
The major obstacles in the earlier studies of reaction 1 include
the following: (1) the uncertainty of the absorbance cross section
of IO radical; (2) the uncertainty of the rate constant for the
self-reaction of IO radical; (3) wall loss of the IO radical. This
recommended value is based on three measurements by Maguin
7
,8
dominated by reaction with nitrate radical (NO3). In the
daytime, DMS is thought to be oxidized by reaction with the
hydroxyl radical (OH).9 This reaction has a recommended
,10
11
-
12
3
-1
room-temperature rate constant of 5.0 × 10 cm molecule
-
1
s . Due to an absence of a known alternative, it is common
for models to include only OH and NO3 as oxidation sources
for DMS.6
2
3
24
25
et al., Barnes et al., and Knight and Crowley, as well as
2
6
an upper limit set by Daykin and Wine. The former three
studies were done at low pressures (2 Torr or lower), and the
study by Daykin and Wine (at 40, 100, and 300 Torr) found no
pressure dependence.
Field measurements12 have been made of iodine oxide (IO)
in the MBL, also at the Mace Head site. However, because of
the small number of measurements made to this point and the
fact that these measurements are close to the current detection
limit of IO in the field, how representative these concentrations
are for the entire MBL is unknown. The authors of ref 12
speculate that IO is ubiquitous in the MBL. The main source
of IO is biogenic,1 in the form of organic iodine compounds
such as CH2I2. Emission of these compounds correlates with
times of high biogenic activity.15 These compounds are photo-
In this work, we examine the reaction of IO radicals with
DMS using cavity ring-down spectroscopy (CRDS). This
2
7
technique was employed in a previous study, performed in
this laboratory, of the reaction of Br atoms and BrO radicals
with DMS. In that work, the reaction of BrO radicals with DMS
was found to have a pressure dependence. In this work, both
the pressure and the temperature effects of reaction 1 are
examined. In the process, we examine the high-resolution
spectrum of the IO radical using laser spectroscopy. The low
detection limit achieved by using the CRDS technique allows
us to spectroscopicly observe IO radical directly and in a
quantitative manner.
3,14
*
To whom correspondence should be addressed. E-mail: kawasaki@
photon.mbox.media.kyoto-u.ac.jp. Fax: +81-75-753-5526.
Present address: California State University at Channel Islands, One
University Drive, Camarillo, CA 93010.
†
1
0.1021/jp0345147 CCC: $25.00 © 2003 American Chemical Society
Published on Web 07/26/2003

6
2
382 J. Phys. Chem. A, Vol. 107, No. 33, 2003
. Experimental Section
Nakano et al.
giving the following reaction sequence:
1
The CRDS apparatus used in this study is described in detail
O + hν (λ ) 266 nm) f O( D) + O
(2)
(3)
(4)
3
2
2
8
elsewhere. The system employs two pulsed lasers, a probe
laser and a photolysis laser. The probe beam was generated by
1
3
O( D) + N f O( P)+ N
3+
2
2
the 355 nm output of a Nd :YAG laser (Spectra-Physics PRO-
30). The beam was directed to an optical parametric oscillator
Spectra-Physics Quanta Ray MOPO-SL), with a tunable output
2
(
3
O( P) + CF I f IO + CF
3
3
range from 440 to 1800 nm. In this experiment, the idler output
from the MOPO was then frequency doubled with a BBO crystal
The rate constant for reaction 3 is 2.6 × 10 molecule-1 cm3
-
11
-
1
11
s
at room temperature, resulting in a lifetime of less than
.2 µs for all experimental conditions in this study. The rate
(Spectra-Physics FD-900) to obtain the desired wavelength
0
range. This beam was used to probe the concentrations of
reactive species in the system. After the photolysis laser beam
traversed the cell nearly collinear to the axis of the ring-down
cavity, the probe laser beam was injected through one of two
high-reflectivity mirrors, which made up the ring-down cavity.
The mirrors (Research Electronic Optics) had a specified
maximum reflectivity of >0.9994 at 435 nm, a diameter of 7.8
mm, and a radius of curvature of 1 m and were mounted 1.04
m apart. Light leaking from one of the mirrors of the ring-down
cavity was detected by a photomultiplier tube (Hamamatsu:
R212UH) through a narrow band-pass filter for 442 nm. The
decay of the light intensity was recorded using a digital
oscilloscope (Tektronix TDS714L) and transferred to a personal
computer. The decay of the light intensity is given by eq I,
constant for reaction 4 was monitored using our system, with
the results being presented in this work. The CF3I concentration
1
4
-3
was varied as (3-18) × 10 molecule cm for determination
of the rate constant of reaction 4. For determination of the rate
14
constant of reaction 1, the CF3I concentration was 8 × 10
-
3
molecules cm . The maximum IO radical concentrations (6 ×
1
1
-3
1
0
molecules cm ) was observed to occur between 400 and
8
00 µs after the initial photolysis laser pulse. We also attempted
the experiment using CH3I instead of CF3I but found that the
yield of IO atoms was less. This is in agreement with previous
work by Gilles et al. That work demonstrated that the
branching ratio of IO production from the O( P) reaction with
CH3I was less than that of CF3I by a factor of about 2 to 1. The
IO radical concentration was monitored at 435.60 nm, corre-
sponding to the A Π3/2 r X Π3/2 (3,0) band head. We measured
2
9
3
Where I0 and I(t) are the intensities of light at time 0 and t,
respectively. τ0 is the empty cavity ring-down time (1.5 µs at
2
2
-
17
2
the IO absorption cross section to be 5.9 × 10
cm
4
35.51 nm). LR is the length of the reaction region (0.46 m). L
-
1
molecule at this wavelength with a resolution of about 0.2
is the cavity length (1.04 m). τ is the measured cavity ring-
down time. n and σ are the concentration and absorption cross
section of the species of interest, and c is the velocity of light.
-
1
cm . The IO concentration profile was measured between 800
and 10800 µs after the initial photolysis laser pulse. The decay
profile was a result of the self-reaction of IO, as well as the
reaction of IO with DMS.
Dimethyl sulfide (DMS, 98%) was purified using the freeze-
pump-thaw method. DMS vapor was collected in an evacuated
glass gas bulb. The bulb was then filled with N2 to produce a
5% mixture of DMS/N2. The amount of DMS in the reaction
chamber was determined by using a mass flow controller. DMS
I(t) ) I exp(-t/τ) ) I exp(-t/τ - σnc(L /L)t) (I)
0
0
0
R
The reaction cell consisted of a Pyrex glass tube (21 mm
i.d.) that was evacuated by a combination of an oil rotary pump
and a mechanical booster pump. The temperature of the gas
flow region was controlled by circulation of a thermostated
mixture of ethylene glycol and water and was controllable over
the range 273-312 K. The difference between the temperature
of sample gas at the entrance and exit of the flow region was
1
4
15
number densities were varied between 5 × 10 and 5 × 10
-
3
molecules cm . The other commercially obtained reagents,
CF3I (97%, Lancaster), N2 (>99.999%), and O2 (> 99.995%),
were used without further purification.
<
1 K. The pressure in the cell was monitored by an absolute
pressure gauge (Baratron). Gas flows were measured and
regulated by mass flow controllers (KOFLOC: 3660). A slow
flow of nitrogen gas was introduced at both ends of the ring-
down cavity close to the mirrors to minimize deterioration
caused by exposure to reactants and products. The total flow
3. Results
In the course of determining the rate constant for the reaction
of IO with DMS, we measured the cross section of IO in the
2
2
A Π3/2 r X Π3/2 (3,0) band head. We calculated the cross
section at this wavelength by plotting the absorbance at this
band against the known IO number density. This plot is shown
in Figure 1. We determined the number density of IO by
measuring the absorbance at 427.2 nm, the band head of the
3
3
-1
rate was about 2 × 10 cm min (STP).
Ozone was produced by irradiating the oxygen gas flow with
the 184.6 nm output of a low-pressure Hg lamp (Hamamatsu
L937) at high pressure (>720 Torr). It was measured upstream
of the reaction tube by monitoring the absorption at 253.7 nm
2
2
A Π3/2 r X Π3/2 (4,0) band. The cross section at this point is
30
-17
known from previous work by Harwood et al., 3.6 × 10
-
17
2 11
(
σ ) 1.15 × 10
cm ) using a separate low pressure Hg
2
2
2
cm . The maximum absorbance for the A Π3/2 r X Π3/2 (3,0)
lamp as the light source. The probe lamp is covered with Vicol
glass so that ozone is only detected and not generated from
-17
2
band is at 435.60 nm, with a cross section of 5.9 × 10 cm .
This is the first measured cross section of this band using laser
1
2
that light source. Typical ozone number densities of 4 × 10
-1
spectroscopy with resolution of 0.2 cm . As can be seen in
Figure 1, the (3,0) band is a relatively structured band compared
to the (4,0) band as reported by Harwood et al. For this reason,
when using a higher resolutions the (3,0) band head has a higher
absorption cross section than the (4,0) band, in contrast to what
has been reported.
Details of production of the IO are in the Experimental
Section. IO was formed by reacting CF3I with O( P). We used
-
3
molecules cm were generated in this way.
3+
30
The 266 nm output of a Nd :YAG laser (Spectra Physics
-18
GCR-250) was used to photodissociate ozone (σ ) 9.7 × 10
2
1
cm ) to give O( D) atoms. The contribution of I atoms from
the photodissociation of CF3I at 266 nm is relatively minor,
-19
2
1
due to its small cross section (σ ) 6.8 × 10 cm ). The O( D)
3
3
atoms are relaxed by N2 collisions to produce O( P) atoms. IO
3
radicals were formed by allowing O( P) to react with CF3I
the rise time of IO production to determine the rate constant

Reaction of IO Radicals with Dimethyl Sulfide
J. Phys. Chem. A, Vol. 107, No. 33, 2003 6383
Figure 3. Decay profile of IO in the absence (open circles) and
1
5
-3
2
2
presence (filled squares) of DMS (1.6 × 10 molecules cm ). Curves
Figure 1. Absorbance of IO at the A Π3/2 r X Π3/2 (3,0) band head
2
2
were fit to eq II. Data were taken at 100 Torr of N diluent and 298 K.
2
plotted against the concentration of IO. The A Π3/2 r X Π3/2 (3,0)
band of IO is shown in the inlay box. An arrow marks the peak
wavelength used in monitoring IO concentration.
the concentration of IO can be given by the following equation:
2
k5
2k5
1
1
1st
)
+
exp(k1 t) -
(II)
1
st
1st
[
IO]t {([IO]0
)
(
)}
k1
k1
Here [IO]t is the concentration of IO, [IO]0 is the initial
concentration of IO, k5 is the rate constant of reaction 5, and t
is the time. The self-reaction of IO shows no pressure or
significant temperature dependencies under the range of condi-
tions of this study. The first-order term, k1 , is the sum of
the pseudo-first-order reaction of IO + DMS, plus loss due to
diffusion.
3
0
1st
1
st
k1 ) k [DMS] + k
(III)
1
d
In the absence of DMS, IO removal is regulated by reaction 5
and by the diffusion rate. Figure 3 shows the decay profile of
IO in our system in the absence of DMS. When we fit these
data to eq II, we were able to calculate k5 to be (9.4 ( 1.8) ×
Figure 2. Rise profile of IO after the initial photolysis pulse (time )
0
µs) by 266 nm output of a Nd3+:YAG laser.
-
11
-1
3
-1
1
0
molecule cm s at room temperature. This is in
1
1
reasonably good agreement with the NASA recommended
for reaction 4. The rise time is shown in detail in Figure 2.
From these data, we determined the rate constant for reaction 4
-11
value of 8 × 10 . It is in excellent agreement with the value
30
-11
-1
3 -1
obtained by Harwood et al., 9.9 × 10 molecule cm s .
We performed experiments to determine the rate constant for
the reaction of IO with DMS, reaction 1. In addition to the plot
with no DMS, Figure 3 also shows the decay profile of IO
-
12
-1
3
-1
to be (7.8 ( 0.3) × 10
molecule cm
s
at room
temperature. This vale is larger than some recently reported
values by Gilles et al.29 and by Holscher et al. However,
complications from the reaction of photolysis of the CF3I were
minimal in our system. This is because the reaction of I + O3
31
15
concentration with a DMS concentration of 1.6 × 10 molecule
-3
3
cm . Due to competition for O( P) with DMS, the maximum
IO number densities decreases when the amount of DMS is
-
12
-1
3
-1
is a relatively slow reaction (1.2 × 10
molecule cm s
at room temperature).11 Also, as previously mentioned CF3I has
a relatively small cross section. Under experimental conditions,
IO production from I + O3 is estimated to be below 2% of the
total contribution. CF3 mostly reacts with O2 to generate CF3O2,
which is considered to be less reactive radical. Therefore, the
contribution of the CF3 reactions can be negligible. The effects
3
increased. The reaction of O( P) with DMS affects to the only
the initial concentration of IO; hence, there is no influence of
1st
this reaction for the determination of k1 . DMS number
14
15
-3
densities ranged from 5 × 10 to 5 × 10 molecules cm .
1st
We plot k1 versus the DMS number density and fit the data
using a linear regression. An example of this is shown in Figure
3
of the reaction of O ( P) with I2 and IO are negligible, because
4
. From this fit, we can obtain the rate constant for the IO +
these concentrations are much lower compared with CF3I.
DMS reaction, k1. At 5 Torr total pressure and room temperature,
-
14
-1
3 -1
We then measured the IO concentration decay as a function
of time. The IO is depleted due to the self-reaction,
we calculate k to be (1.0 ( 0.3) × 10
molecule cm s .
1
11
This value corresponding to the NASA recommendation value,
-
14
-1
3 -1
1
.2 × 10 molecule cm s , is based on the previous low-
23-25
IO + IO f products
(5)
pressure studies
(<2 Torr), as well as the upper limit
2
6
determined by Daykin and Wine. As we increased the total
pressure of the reaction chamber, we observed an increase in
the rate constant. This pressure dependency was observed up
to about 100 Torr total pressure, where the reaction seems to
and the IO reaction with DMS (reaction 1), as well as diffusion
out of the detection region. Hence, the decay of IO can be
analyzed as a sum of the first and second-order kinetics, and

6384 J. Phys. Chem. A, Vol. 107, No. 33, 2003
Nakano et al.
Figure 4. Second-order plot for the reaction of IO radicals with DMS
in 100 Torr of N
fit.
2
diluent at 298 K. The line is a linear least-squares
Figure 6. Arrhenius plot for the reaction of IO radicals with DMS.
Data were taken at 100 Torr of N diluent.
2
TABLE 2: Rate Constant of IO + DMS with 100 Torr of
Diluent at Several Temperatures
N
2
1
3
13
1
0 k
1
10 k
1
temp (K) (molecule cm s-1) temp (K) (molecule-1 cm s-1)
-
1
3
3
2
2
73
83
3.9 ( 1.1
3.1 ( 0.7
298
312
2.5 ( 0.2
1.3 ( 0.4
noting that, as recently shown by Troe,32 the value of 0.6 in the
equation above is an empirically derived parameter. In actuality,
the value can vary by a factor of 2 or more. In the case of our
data, the above equation fit the results reasonably, but other
values of that parameter ranging from 0.3 to 0.8 also fit the
high
equation within error limits. The value for k1 was determined
-
13
-1
3 -1
to be (2.5 ( 0.2) × 10 molecule cm s . The low-pressure
low
rate constant, k1 , determined from the fit to the equation
-
31
-2
6
-1
above, is (8.9 ( 5.2) × 10
molecule cm s .
The pressure dependence observed in this work is in direct
contrast to the previous study of Daykin and Wine.26 In that
work, the authors performed experiments at 40, 100, and 300
Torr total pressure. In those experiments, the photolysis of NO2
produced oxygen atoms, which reacted with I2 to produce IO.
Another major difference between that work and ours is that
the DMS concentration was 1-2 orders of magnitude larger in
the study of Daykin and Wine, while the IO concentration is
about twice as large in their work.
Figure 5. Rate constant of IO radicals with DMS as a function of N
2
diluent pressure at 298 K. The plot includes the NASA/JPL recom-
mended value (open square) at 1 Torr total pressure.11
TABLE 1: Rate Constant of IO + DMS at 298 K with
2
Several Pressures of N Diluent
1
4
14
tot. pressure
10 k1
tot. pressure
10 k1
Torr of N2) (molecule- cm s-1) (Torr of N2) (molecule cm s-1)
1
3
-1
3
(
5
0
00
1.0 ( 0.3
11 ( 4
25 ( 2
150
200
17 ( 9
25 ( 8
We also measured the temperature dependence of reaction
1. The values of the rate constant at various temperatures from
5
1
2
73 to 312 K are listed in Table 2. From these data, we
reach a high-pressure limit. At 100 Torr, our measured rate
constructed an Arrhenius plot shown in Figure 6. Linear least-
squaresd analyses of the data in Figure 6 yield the following
expression:
-
13
-1
3
-1
constant is (2.5 ( 0.2) × 10
molecule cm s , while at
2
0
00 Torr total pressure our measured rate constant is (2.5 (
-
13
-1
3
-1
.8) × 10
molecule cm s . The observed rate constants
+
4.5
-1.0
-16
at room temperature, with varying pressures from 5 to 200 Torr,
are listed in Table 1. The NASA recommended value previously
measured is also included in Figure 5. These data have been
used to fit the pressure dependence of the rate constant to the
following empirically derived equation:
k1 ) 1.2
× 10
×
-
1
3 -1
exp[(2230 ( 460)/T] molecule cm s (V)
We believe this is the first determination of the Arrhenius
expression for the reaction of IO with DMS. This reaction
exhibits a negative activation energy of -18.5 ( 3.8 kJ mol .
This is consistent with what was observed in the reaction of
BrO with DMS. As in that reaction, it suggests a mechanism
involving a complex intermediate species:
-
1
low
k1 [M]
{1+[log(k low[M]/k high)]2}-1
1
1
k1 )
× 0.6
(IV)
(
low
high
)
1
+ (k1 [M]/k1
)
In this equation, k1 is the observed rate constant for reaction 1,
IO + DMS a [IO-DMS] f I + DMSO
(1a)
low
high
k1 and k1
are the low and high pressure limits, and [M] is
the total number density. The rate constants, including the best-
fit curve, are plotted versus total pressure in Figure 5. It is worth
The magnitude of the negative activation energy suggests a
longer lived intermediate than is the case for BrO-DMS. This

Reaction of IO Radicals with Dimethyl Sulfide
J. Phys. Chem. A, Vol. 107, No. 33, 2003 6385
is consistent with the larger pressure effects seen for this
reaction. The large pressure effect of this reaction is evidence
for the termolecular reaction path. This effect is greater for IO
involve a complex intermediate. Indeed, Diaz-de-Mera et al.35
postulate such a mechanism for the reaction of ClO + DMS in
their recent work. We expect that the stability of the intermediate
is greater in the case of IO and BrO than for ClO. The greater
stability translates to a longer lifetime to allow the reaction to
proceed to form DMSO + X:
+
DMS than for BrO + DMS. A recent study by Williams et
3
3
al. shows that the adduct of OH with DMS may be underes-
timated at lower temperatures. On the basis of these studies,
DMS seems to readily form adducts with potential oxidizers.
The relative reactivity of halogen oxides (XO) toward DMS
at pressures relevant to the atmosphere follows the pattern BrO
+
M
q
XO + DMS a [XO-DMS]
8 [XO-DMS] f
X + DMSO (VI)
≈
IO > ClO. This is in contrast to a previously reported study
based on low-pressure studies, which claims that the reactivities
of IO and ClO are similar. All reactions of these halogen oxides
This interpretation of the results is speculative, however, and
clearly more detailed studies of the reaction mechanism are
needed to fully explain the reaction trend.
3
4
with DMS are thought to produce DMSO. As previously
pointed out, on the basis of thermodynamic arguments, it might
be expected the order of reactivity is IO > BrO . ClO. Our
calculated change in enthalpy uses data taken from ref 11. The
change in enthalpy of these reactions is listed below.
4
. Atmospheric Implications
DMS is ubiquitously emitted from the oceans of the world
3
9
from biological sources. It is the most important source of
IO + DMS f I + DMSO
3
9
sulfur from the oceans. Average mixing ratios in the marine
-
1
∆
H (298) ) -134 kJ mol (1)
39
f
boundary layer are around 100 ppt, but concentrations are
highly variable with biological and tidal activity, as well as
climate. Oxidation of DMS eventually produces highly water-
BrO + DMS f Br + DMSO
-
1
∆
H (298) ) -127 kJ mol (6)
soluble species with low vapor pressures, such as sulfuric
f
1
(
H2SO4) and methanesulfonic acid (CH3SO2OH, MSA). These
ClO + DMS f Cl + DMSO
species are thought to form the building blocks of aerosol
production for a large number of atmospheric particles,4
hence the link between oxidation of DMS and aerosol formation.
DMS is removed less efficiently at night, when oxidation is
dominated by nitrate radical (NO3).
A predominate sink of DMS during the day is reaction with
0,41
-
1
∆
H (298) ) -94.5 kJ mol (7)
f
The exothermicity of these reactions seems to agree with our
observed reactivity order for XO + DMS. This thermodynamic
argument is based primarily on the difference in bond dissocia-
tion energy of the halogen oxide. It is worth noting that, for
BrO and IO, this value has an uncertainty of about 10%. There
is also uncertainty related to the measured rate constants. These
errors combined may improve the reaction enthalpy correlation
to the reaction kinetics.
The difference in reactivity between IO and BrO may also
be attributed to steric hindrance. IO may have a lower
probability, with respect to BrO, to be properly oriented for
reaction. There is evidence for this in the apparently much
smaller preexponential factor observed for the IO + DMS
reaction than for the BrO + DMS reaction. This would be
expected given the size of iodine with respect to bromine. This
argument cannot be extended to the ClO case, for which the
preexponential factor has been recently measured35 to be smaller
than that for BrO.
4
2,43
OH.9 Oxidation of DMS is far from completely understood.
,10
6,44
6
As mentioned earlier in this work, one study showed that
oxidation of DMS by OH was underestimated by a factor of 3
in the MBL at the Mace Head site. The rate constant for this
11
-
12
3
-1 -1
reaction of OH with DMS is 5.0 × 10
cm molecule
s
at room temperature, with some large uncertainty in this value
3
3
at lower temperatures. A recent work finds that this rate
constant is underestimated at much lower temperatures, but it
is unlikely that this can be used to explain the discrepancy.
The lack of understanding of DMS oxidation has led to a
search for other sinks, such as chlorine atom. Recently,
45
measurements have indicated the amount of molecular chlorine
in the marine boundary layer is much larger than previously
thought. Chlorine atom reacts rapidly with DMS, with a rate
-
10
3
-1 -1
constant of about 3.3 × 10
cm molecule
s
at room
Ionic curve-crossing models36 provide an alternative explana-
tion of these relative reactivities to the ones presented above.
In that model, relative reactivities of radical-molecule reactions
can be accurately predicted on the basis of the thermodynamics
of the ionic potential energy surface. This model is noted for
its ability to explain the reactivity of OH and chlorine atom
46
6
temperature. The modeling work by James et al. suggests
that DMS oxidation by Cl cannot account for the measurement
results observed because peak concentrations are too short-lived.
4
3
It has been shown that the Cl + DMS reaction can proceed
via an adduct, as well as via hydrogen abstraction. The hydrogen
abstraction path leads to the formation of aerosols, while the
3
7
1
47
(Cl) reactions with a series of alkanes. In that case, the increase
fate of the adduct is unclear. There is some evidence that the
Cl-DMS adduct can react with oxygen to produce DMSO, but
further studies are needed.
in the rate constant for a particular electron acceptor, the radical,
is associated with the ionization energy of the electron donor,
the alkane. In the case of XO + DMS, the halogen oxide is the
electron acceptor; hence, its reactivity in this type of reaction
should be associated with its electron affinity. The electron
affinities for IO, BrO, and ClO are 2.38, 2.35, and 2.38 eV,
respectively. Again, the trend seen in the bond dissociation
model is followed. It is not entirely clear to these authors
whether a difference in these values, about 10 kJ mol , can
cause such a dramatic difference in the reaction rates.
In light of the data presented in this work, it does seem that
the reaction mechanisms of IO + DMS and BrO + DMS
Reaction of DMS with halogen oxide molecules has been
speculated to play a role in the atmosphere for some time.2
Reaction with ClO and BrO (as well as IO) has been deemed
1,39
3
8
1
too slow to compete with OH and Cl oxidation. This is due to
-
15
3
3
a slow rate constants for DMS with ClO (9.5 × 10
cm
cm
-
1
-1
-13
molecule
molecule
s
at 300 K) and with BrO (4.2 × 10
-
1
-1
s-1 at 300 K).
24,27,35
Maximum BrO number
7
densities in the MBL are thought to be on the order of 1 × 10
-
3 48
molecules cm . Arctic air can contain larger amounts of
49
halogen oxides, but this is not considered typical of the MBL.

6386 J. Phys. Chem. A, Vol. 107, No. 33, 2003
Nakano et al.
It is not known to the authors how mixing of this air can effect
DMS oxidation in the MBL in places such as the Mace Head
site.
(6) James, J. D.; Harrison, R. M.; Savage, N. H.; Allen, A. G.; Grenfell,
J. L.; Allan, B. J.; Plane, J. M. C.; Hewitt, C. N.; Davison, B.; Robertson,
L. J. Geophys. Res. 2000, 105, 26379-26392.
(7) Wallington, T. J.; Atkinson, R.; Winer, A. M.; Pitts, J. N. J. Phys.
Recent measurements show that concentrations of iodine
monoxide (IO) in the marine boundary layer are much higher
than previously thought.14 These data show marine boundary
layer IO seems to be predominately produced by the photolysis
and oxidation of CH2I2. While CH3I is the most abundant iodine-
containing species emitted from the oceans, CH2I2 is photolyzed
by longer wavelength light,16 hence acting as the source of IO
in the lower atmosphere more efficiently. There is further
evidence of this in the work of Simpson et al.,50 showing that
CH3I and DMS concentrations are not correlated in the MBL.
Chem. 1986, 90, 4640-4644.
(8) Dlugokencky, E. J.; Howard, C. J. J. Phys. Chem. 1988, 92, 1188-
1193.
(
9) Hynes, A. J.; Wine, P. H.; Semmes, D. H. J. Phys. Chem. 1986,
0, 4148-4156.
10) Barone, S. B.; Turnipseed, A. A.; Ravishankara, A. R. J. Phys.
9
(
Chem. 1996, 100, 14694-14702.
(11) DeMore, W. B.; Sander, S. P.; Golden, D. M.; Hampson, R. F.;
Kurylo, M. J.; Howard, C. J.; Ravishankara, A. R.; Kolb, C. E.; Molina,
M. J. Chemical Kinetics and Photochemical Data for use in Stratospheric
Modeling, JPL Publication 97-4; Jet Propulsion Laboratory: Pasadena, CA,
1997.
(12) Alicke, B.; Habestreit, K.; Stutz, J.; Platt, U. Nature 1999, 397,
572-573.
14
Measurements by Carpenter et al. show that IO concentrations
8
-3
can reach as high as 1.5 × 10 molecules cm . It should be
noted that there are only a few measurements of IO in the MBL
and more measurements need to be made to assess its
significance. The room-temperature rate constant measured in
(
13) Happell, J. D.; Wallace, D. W. R. Geophys. Res. Lett. 1996, 23,
105-2108.
14) Carpenter, L. J.; Sturges, W. T.; Penkett, S. A.; Liss, P. S.; Alicke,
2
(
B.; Hebestreit, K.; Platt, U. J. Geophys. Res. 1999, 104, 1679-1689.
(15) Yokouchi, Y.; Nojiri, Y.; Barrie, L. A.; Toom-Sauntry, D.;
Fujinuma, Y. J. Geophys. Res. 2001, 106, 12661-12668.
(16) Roehl, C. M.; Burkholder, J. B.; Moortgat, G. K.; Ravishankara,
-
13
3
-1 -1
this work is 2.5 × 10
cm molecule s . Under these
conditions, the lifetime of DMS with respect to IO is about 7.4
1
4
h. It should be noted that, for the field measurements, IO
concentrations were below the detection limit of the instrumen-
tation on most days. The detection limits for the quoted
measurements are near 5 × 10 molecules cm . However, even
using the detection limit as a typical maximum of IO, the
lifetime of DMS with respect to IO becomes 22 h. The daily
temporal profile of IO concentration should generally be similar
to that of OH because they are both produced by photolysis of
species (O3 and CH2I2) that absorb light in roughly the same
region of the solar spectrum. Furthermore, recent works by de
A. R.; Crutzen, P. J. J. Geophys. Res. 1997, 102, 12819-12829.
(17) Turnipseed, A. A.; Giles, M. K.; Burkholder, J. B.; Ravishankara,
A. R. Chem. Phys. Lett. 1995, 242, 427-434.
(18) Bauer, D.; Ingham, T.; Carl, S. A.; Moortgat, G. K,; Crowley, J.
N. J. Phys. Chem. A 1998, 102, 2857-2864.
7
-3
(
19) Rowley, D. M.; Mossinger, J. C.; Cox, R. A.; Jones, R. L. J. Atmos.
Chem. 1999, 34, 137-151.
20) Bloss, W. J.; Rowley, D. M.; Cox, R. A.; Jones, R. L. J. Phys.
(
Chem. A 2001, 105, 7840-7854.
(21) Martin, D.; Jourdain, J. L.; Laverdet, G.; Le Bras, G. Int. J. Chem.
Kinet. 1987, 19, 503-512.
(22) Barnes I.; Becker, K. H.; Carlier, P.; Mouvier, G. Int. J. Chem.
Kinet. 1987, 19, 489-501.
5
1
Chen et al. show an underestimation of DMSO yield from
DMS oxidation. Oxidation by halogen oxides is thought to
produce DMSO with a yield of almost unity.1
(
23) Maguin, F.; Mellouki, A.; Laverdet G.; Poulet, G.; Le Bras, G.
Int. J. Chem. Kinet. 1991, 23, 237-245.
24) Barnes, I.; Bastian, V.; Becker, K. H.; Oerath, R. D. Int. J. Chem.
Kinet. 1991, 23, 579-591.
25) Daykin, E. P.; Wine, P. H. J. Geophys. Res. 1990, 95, 18547-
18553.
(26) Knight, G. P.; Crowley, J. N. Phys. Chem. Chem. Phys. 2001, 3,
93-401.
27) Nakano, Y.; Goto, M.; Hashimoto, S.; Kawasaki, M.; Wallington,
T. J. J. Phys. Chem. A 2001, 105, 11045-11050.
28) Ninomiya, Y.; Hashimoto, S.; Kawasaki, M.; Wallington, T. J. Int.
(
5. Conclusions
(
We have measured the pressure and temperature dependencies
of the reaction of IO + DMS. The rate constant was measured
by using cavity ring-down spectroscopy. This allowed us to
measure a small amount of IO, reducing the uncertainties
associated with some other techniques, such as wall loss. At
lower pressures, the rate constant was found to be in reasonable
agreement with previously reported values. The reaction shows
a pressure dependence, however. Under atmospheric conditions
relevant to the marine boundary layer, we show that IO can be
an important oxidizer of DMS. For this reason, modeling studies
and field measurement campaigns should be performed to
provide better understanding of this critical topic.
3
(
(
J. Chem. Kinet. 2000, 32, 125-130.
(29) Gilles, M. K.; Turnipseed, A. A.; Talukdar, R. K.; Rudich, Y.;
Villalta, P. W.; Huey, L. G.; Burkholder, J. B.; Ravishankara, A. R. J. Phys.
Chem. 1996, 100, 14005-14015.
(30) Harwood, M. H.; Burkholder, J. B.; Hunter, M.; Fox, R. W.;
Ravishankara, A. R. J. Phys. Chem. A 1997, 101, 853-863.
(31) Holscher, D.; Fockenberg, C.; Zellner, R. Ber. Bunsen-Ges. Phys.
Chem. 1998, 102, 716-722.
(32) Troe, J. Int. J. Chem. Kinet. 2001, 33, 878-889.
(
33) Williams, M. B.; Campuzano-Jost, P.; Bauer, D.; Hynes, A. J. Chem.
Acknowledgment. This work is supported by a Grant-in-
Aid for the priority research field “Radical Reactions” from the
Ministry of Education of Japan. S.A. thanks the Japan Society
for Promotion of Science for a fellowship.
Phys. Lett. 2001, 344, 61-67.
34) Barnes, I.; Bastian, V.; Becker, K. H.; Overath, R. D. Int. J. Chem.
(
Kinet. 1991, 23, 579-591.
(35) Diaz-de-Mera, Y.; Aranda, A.; Rodriguez, D.; Lopez, R.; Cabanas,
B.; Martinez, E. J. Phys. Chem. A 2002, 106, 8627-8633.
(
(
36) Donahue, N. M. J. Phys. Chem. A 2001, 105, 1489-1497.
37) Smith, I. W. M.; Ravishankara, A. R. J. Phys. Chem. A 2002, 106,
References and Notes
(1) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower
4798-4807.
Atmosphere; Academic Press: San Diego, CA, 1999.
(38) Gilles, M. K.; Polak, M. L.; Lineberger, W. C. J. Chem. Phys. 1992,
(
2) Aneja, V. P. J. Air Waste Manage. Assoc. 1990, 40, 469-476.
9
6, 8012-8020.
(3) Charlson, R. J.; Lovelock, J. E.; Andreae, M. O.; Warren, S. G.
(
39) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics;
John Wiley and Sons: New York, 1998.
40) Charlson, R. J.; Lovelock, J. E.; Andreae, M. O.; Warren, S. G.
Nature 1987, 326, 655-661.
41) Andreae, M. O.; Crutzen, P. J. Science 1997, 276, 1052-1058.
Nature 1987, 326, 655-661.
4) Houghton, J. T.; Ding, Y.; Griggs, D. J.; Noguer, M.; van der
Linden, P. J.; Xiaosu, D.; Maskell, K.; Johnson, C. A. Climate Change
001: The Scientific Basis; Contribution of Working Group I to the Third
Assessment Report of the Intergovernmental Panel on Climate Change
IPCC); Cambridge University Press: New York, 2001 (ISBN: 05210
4956 2001).
5) Ravishankara, A. R.; Rudich, Y.; Talukdar, R.; Barone, S. B. Philos.
Trans. R. Soc. London, Ser. B 1997, 352, 171-181.
(
(
2
(
(
(42) Wallington, T. J.; Atkinson, R.; Winer, A. M.; Pitts, J. N., Jr. J.
Phys. Chem. 1986, 90, 5293-5396.
(43) Butkovskaya, N. I.; Le Bras, G. J. Phys. Chem. 1994, 98, 2582-
2591.
1
(

Reaction of IO Radicals with Dimethyl Sulfide
J. Phys. Chem. A, Vol. 107, No. 33, 2003 6387
(
44) de Bruyn, W. J.; Harvey, M.; Cainey, J. M.; Saltzman, E. S. J.
Atmos. Chem. 2002, 41, 189-209.
45) Spicer, C. W.; Chapman, E. G.; Finlayson-Pitts, B. J.; Plastridge,
R. A.; Hubbe, J. M.; Fast, J. D.; Berkowitz, C. M. Nature 1998, 394, 353-
56.
46) Stickel, R. E.; Nicovich, J. M.; Wang, S.; Zhao, Z.; Wine, P. H. J.
Phys. Chem. 1992, 96, 9875-9883.
47) Kinnison, D. J.; Mengon, W.; Kerr, J. A. J. Chem. Soc., Faraday
Trans. 1996, 92, 369-372.
48) Grenfell, J. L.; Harrison, R. M.; Allen, A. G.; Shi, J. P.; Penkett,
S. A.; O’Dowd, C. D.; Smith, M. H.; Hill, M. K.; Robertson, L.; Hewitt,
C. N.; Davison, B.; Lewis, A. C.; Creasey, D. J.; Heard, D. E.; Hebestreit,
K.; Alicke, B.; James, J. J. Geophys. Res. 1999, 104, 13771-13780.
(49) Tuckermann, M.; Ackermann, R.; Golz, C.; Lorenzen-Schmidt, H.;
Senne, T.; Stutz, J.; Trost, B.; Unold, W.; Platt, U. Tellus, Ser. B 1997, 49,
533-555.
(
3
(
(50) Simpson, I. J.; Colman, J. J.; Swanson, A. L.; Bandy, A. R.;
Thornton, D. C.; Blake, D. R.; Rowland, F. S. J. Atmos. Chem. 2001, 39,
191-213.
(
(51) Chen, G.; Davis, D. D.; Kasibhatla, D.; Bandy, A. R.; Thorton, D.
C.; Hubert, B. J.; Clark, A. D.; Blomquist, B. W. J. Atmos. Chem. 2000,
37, 137-160.
(