Mechanism for the reaction of hydroxyl radicals with dimethyl disulfide

Article abstract of DOI:10.1016/S0009-2614(99)00862-3

Strong infrared chemiluminescence from the reactions of OH and OD radicals with CH₃SSCH₃ was observed in a discharge flow reactor viewed by a Fourier transform spectrometer. The recorded spectra were identical to the H₂O and HDO plus D₂O emission spectra from the OH+CH₃SH and OD+CH₃SD reactions, respectively. These observations strongly suggest that the primary reaction in the OH and OD+CH₃SSCH₃ system generates CH₃SH and CH₃SD molecules with the observed emission arising from the OH+CH₃SH and OD+CH₃SD secondary reactions.

Full text of DOI:10.1016/S0009-2614(99)00862-3

15 October 1999
Ž
.
Chemical Physics Letters 312 1999 37–44
www.elsevier.nlrlocatercplett
Mechanism for the reaction of hydroxyl radicals with dimethyl
disulfide
a
b,)
N.I. Butkovskaya , D.W. Setser
a
Institute of Chemical Physics, Russian Academy of Sciences, 117334 Moscow, Russian Federation
b
Kansas State UniÕersity, Department of Chemistry, Manhattan, KS, 66506, USA
Received 10 May 1999; in final form 14 July 1999
Abstract
Strong infrared chemiluminescence from the reactions of OH and OD radicals with CH3SSCH3 was observed in a
discharge flow reactor viewed by a Fourier transform spectrometer. The recorded spectra were identical to the H2O and
HDO plus D2O emission spectra from the OHqCH3SH and ODqCH3SD reactions, respectively. These observations
strongly suggest that the primary reaction in the OH and ODqCH3SSCH3 system generates CH3SH and CH3SD
molecules with the observed emission arising from the OHqCH3SH and ODqCH3SD secondary reactions. q 1999
Elsevier Science B.V. All rights reserved.
1. Introduction
Dimethyl disulfide CH3SSCH3, abbreviated as
of the reaction with rapid decomposition to CH3Sq
CH3SOH as the likely fate of the adduct.
Ž
OHqCH SSCH ™CH S OH SCH
Ž
.
3
3
3
3
.
DMDS is one of the naturally occurring sulfur
compounds that plays an important role in atmo-
spheric chemistry. The oxidation in the troposphere
is mainly initiated by reaction with OH radicals and
the final products depend on subsequent elementary
™CH3SOHqCH3S
1a
Ž
.
The final products of the OHqDMDS reaction in
w x
an airqNO atmosphere 5 also have been inter-
Ž .
preted in terms of 1a . Later, Domine and Ravis-
w
x
steps involving the radicals 1–4 produced in the
primary step. Studies of the reaction of CH3SSCH3
with OH were reviewed by Tyndall and Ravis-
w
x
hankara 6,7 using laser-induced fluorescence and
photoionization mass-spectrometry observed both
CH3S and CH3SOH as products of the OHqDMDS
reaction system. However, the measured CH3S yield
was only 28"20% and the CH3SOH yield was not
w x
hankara 2 , and the recommended value of the rate
y10
cm3
Ž
.
coefficient at 298 K is k1 s 2.0"0.3 =10
moleculey1 sy1. Wine et al. 4 found a negative
temperature dependence in the 244–367 K range and
suggested that addition to DMDS was the first step
w x
w x
quantified 6 . The rate constant of the OHq
CH3SCH3 reaction, which proceeds only by abstrac-
w x
tion for our low pressure experimental conditions 8 ,
Ž
is nearly two orders of magnitude smaller, 4.4"
y12
cm3 moleculey1 sy1 2 , than k1, which
)
.
w x
0.3 =10
Corresponding author. Fax: q1-785-532-6666; e-mail:
is strong evidence that the OHqCH3SSCH3 reac-
setserdw@ksu.edu
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 99 00862-3

(
)
38
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
tion proceeds by a mechanism other than abstraction
of a hydrogen atom. Based on the above-mentioned
OH adduct partly decomposes with formation of a
CH3SH molecule that subsequently reacts with OH
radical to give the observed H2O chemilumines-
cence.
Ž
.
results, the addition–decomposition mechanism 1a
has been accepted for modeling the DMDS oxidation
w
x
under atmospheric conditions 3,9 .
OHqCH SSCH ™CH S OH SCH
Ž
.
3
3
3
3
The reaction of OH with methanethiol, another
atmospherically important sulfur compound, has a
large rate constant, 33"4=10y12 cm3 moleculey1
sy1, and it mainly proceeds through addition of OH
to the sulfur atom followed by the elimination of
™CH3SOqCH3SH
1b
Ž
.
In addition to the similarity of the spectra, the depen-
dencies of the HOD and D2O integrated emission
intensities upon reagent concentrations for ODq
DMDS reaction were compared with the HOD emis-
sion intensity from ODqH2 S reaction to support
the proposed mechanism. The very large rate con-
w
x
H2O 2,8 :
OHqCH SH™CH S OH H™CH SqH O
Ž
.
3
3
3
2
2a
2b
Ž
Ž
.
.
Ž .
Ž .
stant for 1 plus the sizeable rate constant for 2 or
Ž .
OHqCH3SH™CH2 SHqH2O
3 enables the IR emission from water molecules to
Ž
.
be observed.
The dominant role of 2a was confirmed in our
recent studies of the reactions of methanethiol with
OH and OD radicals using a discharge flow reactor
Ž
.
with a Fourier-transfer infrared IR spectrometer
2. Experimental method
w
x
8,10 to observe the water products. The available
energy of 2a , 34.4 kcal moly1, is sufficient to
excite up to three stretching quanta in the water
molecule; the vibrational distributions of H2O and
Ž
.
The experiments were done in the KSU laboratory
employing methods described in previous papers
w
x
8,10–13 . The IR chemiluminescence from vibra-
tionally excited products was recorded by a Fourier-
Ž .
HDO from reaction 2 were characteristic for an
addition–elimination reaction and differed substan-
tially from the distribution from the OH OD q
CH3 2 S reaction, which proceeds by abstraction.
Examination of the analogous reaction with CH3SD,
where channels a and b give distinguishable prod-
ucts,
Ž
.
transform IR spectrometer BIORAD , which viewed
the emission from a fast-flow chemical reactor 40
Ž
.
Ž
Ž
.
.
mm ID Pyrex glass pipe through a NaCl window.
Each measured spectrum represents an average of
512 scans of the spectrometer operated with a spec-
tral resolution of 2 cmy1. The OH or OD radicals
were produced via the reaction of H or D with NO2
in the pre-reactor section of the flow reactor. The H
or D atoms were generated by a microwave dis-
Ž .
Ž .
ODqCH3SD™CH3SqD2O
ODqCH3SD™CH2 SDqHOD
3a
3b
Ž
Ž
.
.
Ž
.
Ž
.
revealed that 3b cannot be entirely neglected, since
charge through H2 D2 rAr mixtures. The DMDS,
CH3SH or H2 S reactants were introduced into the
center of the linear reactor through a ring injector
located 20 cm downstream of the hydrogen and
nitrogen dioxide inlets and 3.5 cm upstream of the
center of the NaCl observation window. The total
pressure in the reactor was 0.5 torr, the flow velocity
was about 130 m sy1, corresponding to a reaction
w
x
both D2O and HOD were observed 10 , but the ratio
of the channels 0.16"0.05 is roughly in propor-
tion to the rate constant per H atom for abstraction
from CH3 2 S versus the rate constant for reaction
Ž .
As part of our comprehensive study 8,10–13 of
OH reactions with various molecules using the IR
chemiluminescence experimental technique, we ob-
tained the rather unexpected result that the observed
spectra from OHqDMDS were identical to the
spectra from OHqCH3SH and the spectra from
ODqDMDS were identical to that from ODq
CH3SD. This observation implies that the DMDSP
Ž
.
Ž
.
Ž
.
2 .
w
x
Ž
.
time between OH OD and DMDS of Dt;0.25 ms.
The concentration of H or D atoms was about 2=
1013 molecule cmy3; the NO2 concentration was
about 1=1014 molecule cmy3; the reactant concen-
trations were varied from 3.1=1011 to 1.1=1013
molecule cmy3. Commercial tank grade Ar passed

(
)
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
39
through three molecular sieve traps cooled by ace-
tonerdry ice mixture and liquid N2 before being
added to the reactor. These traps reduce the H2O and
CO2 impurities. The isotopic and chemical purity of
were selected for comparison that had similar reagent
Ž
.
concentrations see the figure caption , and the rela-
tive intensities in Fig. 1 reflect the relative concentra-
tions of the isotopic water molecules generated by
Ž
.
Ž
.
CH3SD CDN Isotopes were )94% and )99.6%,
the CH3SH CH3SD and DMDS chemical reaction
systems. One should note the comparable yields of
water molecules in both systems. The observation of
D2O emission spectra from ODqDMDS is convinc-
Ž
respectively. The CH3SSCH3 purchased as a liquid
from Aldrich Chemical, 99% was degassed in
.
freeze–thaw–pump cycles. The central fraction from
a room temperature liquid sample was expanded into
a 10 l Pyrex glass bulb and a 7% mixture in Ar was
prepared. This mixture was used during the same
working day. The DMDS mixture was metered to the
reactor through a 10 mm ID glass line approximately
50 cm in length. Experiments were done with two
independently prepared DMDSrAr mixtures.
Ž
.
ing evidence for the importance of reaction 1b . The
observation of D2O emission from the ODqDMDS
reaction also eliminates the possibility that the de-
composition of DMDS in the gas handling procedure
could be the source of the water emission, e.g., the
source cannot be ODqCH3SH.
The spectra from the DMDS and CH3SH
CH3SD systems are very similar. Analysis of the
Ž
.
The computer simulation procedure used to assign
stretching and bending mode distributions from the
H2O, HOD and D2O spectra has been described
spectra by computer simulation showed only very
small differences, which can be explained from con-
tamination in the ODqCH3SD system by the
CH3SH isotopic impurity 10 . The H2O stretching
distributions obtained by simulation of the spectra
w
x
8,10–13 . In the present work, our interpretations
w
x
are mainly based on the overall appearance of the
emission spectra and on the total emission intensi-
ties, and we will not describe the simulation proce-
dures. The limited degree of vibrational relaxation of
the product water molecules for the short reaction
time and low Ar pressure has been discussed in Refs.
Ž
.
.
are P1,3 1:2:3 s 66:30:5 from DMDS and
Ž
P1,3 1:2:3 s68:26:6 from CH3SH; the D2O stretch-
Ž
.
ing distributions are P1,3 1:2:3 s64:33:2 for DMDS
Ž
.
and P1,3 1:2:3 s65:32:3 for CH3SD. The 1,3 sub-
scripts denote the coupled n1 and n3 modes. The
HOD emission in the 3200–4000 cmy1 range is
overlapped with the Õ2 qÕ3 combination bands from
D2O, and a complete analysis of these spectra will
w
x
8,10–13 . For the conditions of the experiments to
be described, the populations in the n1 and n3
modes of H2O and D2O and the populations in the
n1 and n2 modes of HOD are coupled by collisions
with Ar. The vibrational relaxation of H2O, D2O,
and HOD molecules beyond this coupling is negligi-
ble. The only assumption in the spectral simulation is
that the rotational distributions are represented by
300 K Boltzmann distributions.
w
x
be made later 10 . It is sufficient here to note that
the two spectra in Fig. 1c,d in the 3200–4000 cmy1
range are identical. The spectral pattern from the
ODqDMDS system, including the intensity ratio
between the D2O and HOD parts of the spectra, did
not change with variation of DMDS or OD concen-
tration, when the former was varied between 1 and
7=1012 molecule cmy3 and the latter was varied
between 1.5 and 2.8=1013 molecule cmy3. The
D2O and HOD emission arise from the same source
in the ODqDMDS reaction and, in fact, this source
is the reaction of OD with CH3SD.
3. Results and discussion
3.1. Chemiluminescent spectra
The raw chemiluminescent spectra from the OH
qDMDS and ODqDMDS reaction systems are
shown in Fig. 1b,d, respectively. They can be com-
pared to the H2O spectra in the 3200–4000 cmy1
3.2. Kinetic behaÕior and reaction mechanism
Ž
.
range from the OHqCH3SH reaction Fig. 1a and
Comparative kinetic measurements were carried
out for the ODqDMDS and ODqH2 S systems in
order to quantify the water yield in the DMDS
system. The H2 S system was chosen as one nearly
y1
Ž
.
Ž
D2O 2400–3000 cm
and HOD 3200–4000
cmy1 spectra from the ODqCH3SD reaction Fig.
.
Ž
.
w
x
1c measured in our previous studies 8,10 . Spectra

(
)
40
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
Ž .
Ž .
x
Fig. 1. Raw chemiluminescent spectra obtained at 0.5 torr and Dts0.25 ms; the spectra for a and c have been elevated so that 0.02 is
13
14
12
their base line.
OHqCH3SH system, OH s2.1=10
,
NO2 s1.6=10 and CH3SH s7.1=10 molecules cmy3
Ž .
a
w
x
w
x
w
Ž .
. b
13
13
w
x
w
w
x
Ž .
w
OHqDMDS system, OH s2.1=10
,
NO2 s6.1=10 and DMDSs7.0=1012 molecules cmy3
.
c
ODqCH3SD system,
OD s2.7=10 , NO2 s1.0=10 and CH3SD s1.2=10 molecules cmy3
.
d ODqDMDS system, OD s2.8=10 , NO2
13
14
13
13
w
x
w
x
x
Ž .
x
w
x
13
s1.5=1014 and DMDS s1.1=10 molecules cmy3
.
w
x
w x
Ž
.
free of secondary chemistry 8 . The spectra were
corrected for the detector response function and the
intensity closed squares and dashed line from the
H2 S reaction with rate coefficient ks4.7=10y12
total intensity was measured in the 2400–3050 cmy1
D2O and in the 3200–4000 cm
ranges. The dependence of the D2O and HOD inte-
grated emission intensities on DMDS and H2 S con-
centration are shown in Fig. 2a. The HOD emission
cm3 moleculey1 sy1 11 shows a linear dependence
on H2 S concentration, as expected. These intensities
were converted to relative concentrations open
squares using the Einstein coefficients and the vi-
brational distribution. The plots with solid triangles
w
x
y1
Ž
.
Ž
.
HOD spectral
Ž
.

(
)
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
41
ODqH2 S and ODqDMDS for the proposed
mechanism are in satisfactory agreement.
The kinetic behavior for ODqDMDS requires a
reaction scheme in which D2O and HOD are both
formed by the same secondary reaction involving
OD radicals. We consider the following possible
mechanisms:
Ž
Ž ..
1. The initial step yields methanethiol reaction 1b ,
Ž .
which leads to production of water via 3 .
2. Hydroxyl radicals react directly with the DMDSP
OD adduct either giving methanethiol, which sub-
sequently reacts with OD,
ODqCH SS OD CH ™2CH SDqO
4a
Ž
.
Ž
.
3
3
3
2
Ž
.
Ž .
or giving water via reactions 4b and 4c .
ODqCH SS OD CH
Ž
.
3
3
™CH SS O CH qD O
Ž .
4b
4c
Ž
Ž
.
.
3
3
2
ODqCH SS OD CH
Ž
.
3
3
™CH SS OD CH qHOD
Ž
.
2
3
3. The primary reaction occurs with formation of
Ž
.
CH3S and CH3SOD, as in 1a , and water
molecules are produced by subsequent reaction of
hydroxyl radicals with methanesulfenic acid.
Fig. 2. Plots of integrated product emission intensities against
Ž .
w x
a HOD and D2O versus DMDS from
reagent concentration:
ODqDMDS and HOD versus H2 S from ODqH2 S at 0.5 torr
w
x
ODqCH3SOD™CH3SOqD2O
ODqCH3SOD™CH2 SODqHOD
5a
5b
Ž
Ž
.
.
13
13
Ž
.
w
x
w
x
Dts0.25 ms with NO2 s8.2=10 and OD s2.5=10
molecule cmy3. The solid squares connected by the dotted line
are the actual integrated HOD intensities from ODqH2 S; the
open squares connected by the solid line are the HOD intensities
scaled by the Einstein coefficients to represent relative HOD
concentrations. The solid triangles and circles connected by solid
curves are the HOD and D2O concentrations from DMDS that
were obtained from the intensities after adjustment for Einstein
Ž .
Ž .
To give a spectrum similar to that from reaction 3 ,
Ž .
Ž .
alternative reaction s must fulfill two conditions: i
the absolute yield and branching ratio between D2O
and HOD producing channels must be equal to those
Ž
.
Ž
.
Ž .
for channels 3a and 3b , and ii the energetics
and, moreover, the dynamics of the hydrogen trans-
fer in both channels must be very similar, so as to
give similar vibrational distributions for both D2O
and HOD. Therefore, the enthalpy changes for the
coefficients.
b Calculated product concentrations for the condi-
Ž .
tions of d and the reaction scheme from Section 3.2 with the rate
constants quoted in the text.
and solid circles from DMDS were obtained by
adjustment of the intensities for the Einstein coeffi-
secondary reaction should be close to those for reac-
y1
Ž .
Ž .
tion 3 with Dr H08 3a sy32.0 kcal mol
and
y1
w
x
Ž
.
cients of the D2O and HOD transitions 8,10 to
obtain relative concentrations. The plot for DMDS
Dr H08 3b sy25.7 kcal mol
.
Ž
.
Reaction 4a can be removed from consideration,
Ž
.
Ž .
3
measured at 0.5 torr Dts0.25 ms displays some
curvature. Roughly, we can interpret the plot as a
linear growth below 3=1012 molecule cmy3 and a
as in this case reaction
reaction, and even if the secondary reaction 4a is as
fast as reaction 1 , the water yields would be an
would be a tertiary
Ž
.
Ž .
w
x
Ž
.
Ž .
fall off due to the OH consumption at higher
order of magnitude less than from 1b q 3 and the
emission could be observed only at substantially
higher OD and DMDS concentrations. The direct
Ž
.
DMDS concentrations. The experimental Fig. 1a
Ž
.
and calculated Fig. 2b relative concentrations from

(
)
42
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
production of water from the adduct can be ruled out
because formation of HOD via the H atom abstrac-
tion from the methyl group should have a negligible
ered as nearly thermoneutral. The enthalpy of reac-
y1
Ž
.
Ž
.
tion 1b is y28 kcal mol for Df H2988 CH3SO s
y18.5 kcal moly1, and reaction 1b is more
Ž
.
Ž
.
Ž .
rate. It is doubtful that reaction 4b is sufficiently
exothermic to give D2O with the necessary energy.
exothermic than reaction 1a .
Secondary reactions in our reactor also could
Ž .
w x
Although reaction 5 has never been observed, it
include NO2 qCH3S and CH3SO 13 , and the ter-
w x
is likely to occur 9 . We failed to find experimen-
tal thermochemical data for CH3SOH or CH3SO;
tiary OD reactions also need to be considered.
CH3SqOD™CH3SOD™CH3SOqD
CH3SqOD™CH3SOD™CH3DqSO
6a
6b
Ž
Ž
.
.
however, the equilibrium structure and Df H2988-
CH3SOH sy33.9 kcal moly1, was calculated in
Ž
.
Ž
.
an ab initio study of the CH3 2 SqOH reaction at
CH SOqOD™CH S O OD™CH DqSO
Ž .
3
3
3
2
Ž
.
w x
the G2 MP2 level of theory 14 . Turecek also
7
Ž .
y1
Ž
.
calculated Dr H2988 5a sy51.7 kcal mol
and
Dr H2988 5b fy25.8 kcal moly1, from which the
Ž .
Ž .
Ž
.
Reactions 6 and 7 have not been observed, but
y1
Ž
.
Ž
.
their rate constants should be large; Dr H2988 6a s
Df H08 CH3SO was estimated as y18.5 kcal mol
.
Ž
.
Ž .
Ž .
Ž .
y8.7, Dr H2988 6b s y58.9 and Dr H2988 7 s
Comparing reactions 5 and 3 , we see that abstrac-
y79.7 kcal moly1 and high vibrational excitation of
tion from the C–H bonds releases similar amounts of
Ž .
Ž .
7 are
Ž
.
the products is likely. Reactions
6
and
energy, but channel 5a is much more exothermic
Ž
.
mentioned because of the weak chemiluminescence
observed in the 2900–3050 cmy1 range. An en-
larged spectrum in this range is shown in Fig. 1d,
where its shape is compared with CH3D emission
than channel 3a . In other words, the C–H bond
w
x
energies are very close, 95.0 14 and 93.9 kcal
moly1 15 , but D CH3SO–H s68.8 kcal mol is
y1
w
x
Ž
.
nearly 20 kcal moly1 weaker than D CH3S–H s
Ž
.
87.4 kcal moly1
. An even smaller value,
Ž .
from the unimolecular decomposition of CH3C O D
1
D CH3SO–H s59 kcal moly1, was assigned by
Ž
.
™CH3DqCO . Although too weak to be assigned
w x
with certainty, this emission may be a progression of
red shifted Q-branches of the CH3D triply degener-
Yin et al. 9 from consideration of the O–H bond
Ž .
Ž .
energy trend in the HOS O O–H and CH3S O O–
x
x
ate vibration of the n4 mode centered at 3000 cmy1
.
H series. Undoubtedly, the weak O–H bond in
CH3SOH would lead to a shift of branching ratio, as
well as to a higher vibrational excitation of the
3.3. Comparison with preÕious studies
Ž
.
product water molecule in channel 5b compared to
w
x
Ž
.
The CH3S radical 6,7 was detected via laser-in-
3b , which is contrary to the experimental observa-
Ž .
3 seems the
Ž .
Ž
.
duced fluorescence as a product of reaction 1 and
tion. Mechanism 1b followed by
the yield was reported as 28"20%. Since the direct
most plausible scenario. As a further test, the D2O
Ž
.
Ž .
production by sequential reactions 1b and 3a are
compared with the HOD production from ODqH2 S
w
x
in Fig. 2b. The calculation corresponds to NO2 0 s
8.2=1013 and OD s2.5=10 molecule cmy3
.
13
w
x
0
The consumption of DMDS was about 70% for an
initial DMDS 0 s7.0=10 molecule cmy3. The
observed and calculated relative D2O and HOD yields
are in satisfactory agreement.
12
w
x
Ž
.
Using the calculated Df H2988 CH3SOH value
Ž
.
together with Df H2988 OH s 9.3, Df H2988-
1 Vibrational excitation of CH4 and CH3D molecules was
observed from the unimolecular decomposition of chemically
y1
Ž
Ž
.
CH3SSCH3 sy5.8 kcal mol
and Df H2988-
CH3S s 33.2 kcal moly1 11 , the enthalpy
.
Ž
.
w x
Ž
.
Ž
.
Ž .
activated CH3C O H or CH3C O D and CH3C O OH or
change for reaction 1a is y4 kcal moly1. As the
accuracy of the Df H2988 values for CH3S and
Ž
.
Ž
.
Ž
.
Ž
.
CH2 DC O OH obtained from successive CH3C O IqH or D
and ICH3C O OHqH or D reactions, respectively see Ref.
Ž
.
Ž
.
Ž
Ž
.
CH3SOH are not high, reaction 1a can be consid-
w x.
16 .

(
)
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
43
Observation of CH3SO and CH3SH is consistent
In the FTIR study of the photolysis of DMDS at
Ž
.
Ž
.
w x
with 1b . Comparison of the temporal profiles of the
CH3SOH, CH3SO and CH3S signals suggests that
all can be primary products; the CH3SH profile was
not presented in the article. It seems that the data of
760 torr in an O2 qN2 mixture, Barnes et al. 3
found that the CH3SO3H formation increased with
an increase of the concentration of OH radicals. As a
possible explanation, formation of CH3SO3H in re-
w x
Ž .
Ref. 6 are compatible with our IR measurements, if
action 10
Ž
.
Ž
.
the branching ratio of 1a to 1b is Q1r3.
CH3SOHqCH3SO3 ™CH3SOqCH3SO3H 10
was proposed. However, the same arguments are
valid for the alternative products, CH3SH and
CH3SO, owing to the possibility of reaction 11 .
Ž
.
Ž .
The products from reaction 1 are important for
modeling the oxidation mechanism in the atmo-
sphere, especially the production of sulfuric acid
versus methanesulfonic acid, CH3SO3H 17 . Ac-
cording to current views, a simplified mechanism of
the oxidation of DMDS or any sulfur containing
compound giving CH3S in the atmosphere proceeds
through the following steps to yield CH3SO3H 1 :
Ž
.
Ž
.
CH3SHqCH3SO3 ™CH3SqCH3SO3H
11
Ž
.
Ž
.
4. Conclusions
w x
CH3SqO2 ™CH3SOO
The IR emission spectra from the OH q
CH3SSCH3 and ODqCH3SSCH3 reactions were
observed from a fast flow reactor operating with 0.5
torr of Ar carrier gas. The observed spectra were
nearly identical to the spectra obtained from the
OHqCH3SH and ODqCH3SD reactions. Espe-
cially striking is the fact that the products from the
ODqCH3SSCH3 reaction are D2O and HOD in the
same proportion and with the same vibrational distri-
bution as for ODqCH3SD. Comparison of the inte-
CH3SOOqNO™CH3SOqNO2
CH SOqO ™CH S O OO
Ž .
3
2
3
CH S O OOqNO™CH SO qNO
Ž .
8
Ž .
3
3
2
2
CH SO qO ™CH S O OO
Ž .2
3
2
2
3
CH S O OOqNO™CH SO qNO
Ž .2
3
3
3
2
CH3SO3 qRH™CH3SO3HqR
Ž
.
grated emission intensities of the OH OD q
Ž
.
CH3SSCH3 reaction to those from OH OD qH2 S
and CH3SH or CH3SD shows that the observed
intensities from CH3SCH3 are consistent with the
The alternative product, H2 SO4, requires formation
of SO2 as the intermediate product. One possible
channel is collisional decomposition of CH3SO2 to
give CH3 qSO2 , which competes with reaction with
O2. Both CH3SO3H and SO2 were found by
Ž
very fast primary reaction giving CH3SH or
.
Ž
.
CH3SD qCH3SO followed by reaction of OH OD
Ž
.
with CH3SH CH3SD . Comparison of our results
with other studies in the literature suggests that the
primary reaction has a branching ratio of R3 in
favor of CH3SH q CH3SO versus CH3S q
CH3SOH. The DMDS reaction mainly proceeds via
addition of hydroxyl radical to one of the sulfur
atoms, followed by four-centered migration of a H
atom to the other sulfur atom, which is favored
because the S–H bond is stronger than the O–H
bond, followed by scission of the S–S bond.
w x
5
Hatakeyama and Akimoto
for reactions of
CH3SH, DMS and DMDS with OH under atmo-
spheric conditions. In their experiments, OH radicals
were generated by the photolysis of alkyl nitrites,
RCH2ONO. The authors invoked reaction 1a to
explain the enhanced yield of CH3SO3H from DMDS
compared to CH3SH. Though gas-phase reactions of
CH3SOH have been never observed, it was specu-
lated that the increase of the CH3SO3H yield arose
from the direct association reaction
Ž
.
CH3SOHqO2 qM™CH3SO3HqM
9
Ž .
Acknowledgements
Ž .
However, reaction 9 seems rather improbable as an
elementary reaction, and the increase of CH3SO3H
yield can be explained by the formation of CH3SO
The work at Kansas State University was sup-
ported by the National Science Foundation Grant
CHE-9505032.
Ž
.
in 1b .

(
)
44
N.I. ButkoÕskaya, D.W. SetserrChemical Physics Letters 312 1999 37–44
w x
8
Ž
.
N.I. Butkovskaya, D.W. Setser, J. Phys. Chem. A 102 1998
6395.
References
w x
9
F. Yin, D. Grosjean, J.H. Seinfeld, J. Atmos. Chem. 11
w x
1
S.B. Barone, A.A. Turnipseed, A.R. Ravishankara, Faraday
Ž
.
1990 309.
Ž
.
Discuss. Chem. Soc. 100 1995 39.
w
w
w
w
x
Ž
.
.
.
10 N.I. Butkovskaya, D.W. Setser, J. Phys. Chem. A 103 1999
w x
2
G.S. Tyndall, A.R. Ravishankara, Int. J. Chem. Kinet. 23
in press.
Ž
.
1991 483.
x
Ž
11 N.I. Butkovskaya, D.S. Setser, J. Phys. Chem. A 102 1998
w x
3
I. Barnes, K.H. Becker, N. Mihalopoulos, in: G. Restelli, G.
9715.
Ž
.
Angeletti Eds. , Dimethylsulfide: Oceans, Atmosphere and
Climate, CEC, Kluwer Academic Publishers, Netherlands,
1993, p. 197.
x
Ž
12 N.I. Butkovskaya, D.W. Setser, J. Chem. Phys. 108 1998
2434.
x
13 Chemical Kinetics and Photochemical Data for Use in Strato-
w x
4
P.H. Wine, N.M. Kreutter, C.A. Gump, A.R. Ravishankara,
spheric Modeling, NASA JPL Publ. 92-20, 1992.
Ž
.
J. Phys. Chem. 85 1981 2660.
w
w
x
Ž
.
14 F. Turecek, J. Phys. Chem. 98 1994 3701.
w x
Ž
.
5
S. Hatakeyama, H. Akimoto, J. Phys. Chem. 87 1983 2387.
F.A. Domine, A.R. Ravishankara, Int. J. Chem. Kinet. 24
x
15 J. Berkowitz, G.B. Ellison, D. Gutman, J. Phys. Chem. 98
w x
6
Ž
.
1994 2744.
Ž
.
1992 943.
w
x
16 N.I. Butovskaya, D.W. Setser, unpublished results.
w x
7
F.A. Domine, A.R. Ravishankara, C.J. Howard, J. Phys.
Ž
.
Chem. 96 1992 2171.