FT-IR product study of the photo-oxidation of dimethyl sulfide: Temperature and O2 partial pressure dependence

Article abstract of DOI:10.1039/a907211j

The products of the OH-radical initiated oxidation of dimethyl sulfur (DMS) were investigated as a function of temperature (284, 295 and 306 ± 2 K) and O₂ partial pressure (20, 200 and 500 mbar) and under NO_x free conditions using the photolysis of H₂O₅ as the OH radical source and FTIR spectroscopy to monitor reactants and products at 1000 mbar total pressure (N₂ + O₂). Dimethyl sulfoxide (DMSO), SO₂ and methane sulfonic acid (MSA:CH₃SO₃H) were the major sulfur-containing products and evidence was found for the formation of methane sulfinic acid (MSIA: CH₃S(O)OH) in considerable yield. High overall molar formation yields of SO₂ were measured suggesting that further oxidation of the products in the addition and abstraction channels result mainly in SO₂ production under NO_x free conditions used in the experiments. The implications of the results for the atmospheric oxidation of DMS under conditions of low NO_x were considered such as those occurring in the remote marine boundary layer.

Full text of DOI:10.1039/a907211j

View Article Online / Journal Homepage / Table of Contents for this issue
FT-IR product study of the photo-oxidation of dimethyl sulÐde:
Temperature and O partial pressure dependence
2
Cecilia Arsene, Ian Barnes* and Karl H. Becker
Bergische Universitat-Gesamthochschule W uppertal, Physikalische Chemie/FB 9,
Gauss Strasse 20, D-42097 W uppertal, Germany. E-mail: barnes=physchem.uni-wuppertal.de;
Fax: ]49 202 439 2505; T el: ]49 202 439 2510
Received 7th September 1999, Accepted 5th November 1999
The products of the OH-radical initiated oxidation of dimethyl sulfur (DMS) have been investigated under
NO free conditions using the photolysis of H O as the OH radical source and FT-IR spectroscopy to
x
2 2
monitor reactants and products at 1000 mbar total pressure (N ] O ). The products were investigated as a
2
2
function of the temperature (284, 295 and 306 ^ 2 K) and the O partial pressure (20, 200 and 500 mbar).
Dimethyl sulfoxide (DMSO), SO and methane sulfonic acid (MSA: CH SO H) were the major
2
2
3
3
sulfur-containing products and evidence has been found for the formation of methane sulÐnic acid (MSIA:
CH S(O)OH) in considerable yield. The variation of the formation of DMSO and SO with temperature and
3
2
O partial pressure is consistent with a mechanism involving both addition and abstraction channels. The
2
molar formation yields of DMSO support that a major fraction of the OH-DMS adduct formed in the
addition channel reacts with O to form DMSO. Due to the uncertainty in the literature rate coefficients used
2
for the corrections in the formation yield of DMSO the possibility of other channels cannot be completely
excluded. High overall molar formation yields of SO have been measured, suggesting that further oxidation of
2
the products in both the addition and abstraction channels result mainly in SO production under the NO
2
x
free conditions employed in the experiments. Consideration is made of the implications of the results for the
atmospheric oxidation of DMS under conditions of low NO such as occur in the remote marine boundary
x
layer.
sea-salt contributes little to the CCN population. However,
1. Introduction
recent evidence suggests that the contribution from sea-salt to
CCN can often be quite substantial and under conditions of
high wind speed may even dominate and also that scavenging
of SO and H SO by sea-salt aerosol is signiÐcant.9h11
Dimethyl sulÐde (DMS) is a product of biological processes
involving marine phytoplankton and is estimated to account
for approximately 60% of the total natural sulfur gases re-
leased to the atmosphere.1h4 It has been postulated that the
emission of DMS from the oceans and the subsequent forma-
tion of sulfate aerosols may have a signiÐcant inÑuence on the
EarthÏs radiation budget and possibly in climate regulation,
the so called CLAW hypothesis.5 The major question con-
cerning DMS in the atmosphere is the extent it plays in con-
trolling the levels of aerosol in the marine boundary layer,
(MBL). There are three main sources of aerosol in the remote
MBL: sea salt, non-sea-salt (nss) sources and entrainment of
free tropospheric aerosol. The principal component of nss-
aerosol is sulfate derived from the oxidation of gaseous
dimethyl sulÐde (DMS: CH SCH ). The relative contributions
2
2
4
In spite of intensive e†orts determination of the quantitative
contribution of DMS-derived non-sea-salt sulfate to CCN in
the marine boundary layer is still highly uncertain. Further
progress in the elucidation of the DMS oxidation mechanism
requires both further advances in Ðeld studies as well as
detailed kinetic/product studies combined with modelling of
the Ðeld data.
In the atmosphere DMS is oxidised mainly by OH and
NO radicals and currently the potential importance of reac-
3
tions with halogens and halogen oxides, in particular BrO, is
under scrutiny. Under pristine background marine conditions
with very low NO , reactions with NO are relatively unim-
3
3
x
3
of the three marine aerosol sources is dependent upon many
factors such as wind speed, frequency of occurrence of clouds
and precipitation, sea-surface DMS emission rates, oxidation
portant. Several reviews give detailed accounts of the present
status of laboratory investigations on these reactions.4,12h16
The reaction of OH with DMS is now known to proceed by
two independent channels: abstraction and addition.12,17,18
mechanism of DMS to SO and MSA, rate of entrainment of
2
free tropospheric aerosol, etc. The SO produced from DMS
Reaction of DMS with NO appears to proceed exclusively by
hydrogen abstraction. In laboratory photo-reactor studies at
2
oxidation is the predominant source of in situ nss-sulfate. SO
2
3
is oxidised to H SO , in the gas phase by reaction with OH
room temperature SO , methane sulfonic acid (CH SO OH:
2
4
2 3 2
MSA), dimethyl sulfoxide (CH SOCH : DMSO) and
radicals and in sea-salt aerosols6 and cloud droplets.7 H SO
2
4
3 3
dimethyl sulfone (CH SO CH : DMSO ) have been observed
contributes either to aerosol growth or to the formation of
new sulfate particles.8 Any e†ect that DMS will have on the
climate is critically dependent on the production of gas phase
sulfuric acid (H SO ) and new particles. The conventional
3
2
3
2
as products of the OH-radical initiated oxidation of
DMS.19,20 Formation of methane sulÐnic acid (CH S(O)OH:
3
MSIA) in low yield has been reported.20 DMSO has been
2
4
wisdom has generally been that nss-sulfate is the primary
source of CCN in the remote marine boundary layer and that
reported to be a major product of the addition channel,20,21
however, the yields are still highly uncertain. DMSO , SO
2
2
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470
This journal is ( The Owner Societies 1999
5463

View Article Online
and MSA have been observed as products of the further oxi-
dation of DMSO in smog chamber studies,20,22 however,
there are discrepancies concerning the yields.
size 7 lm, 100 mm length ] 4.6 mm id) operated at ambient
temperature was used for the anion separation. To analyse for
MSIA, MSA and sulfate a mobile phase consisting of a
mixture of bicarbonate (0.24 mM) and carbonate (0.3 mM)
solutions was chosen as eluent. The Ñow of the eluent was
maintained at a rate of 0.5 mL min~1. Calibrations of the
CH SO ~, CH SO ~ and SO 2~ anions were performed
A recent study suggests that the OH ] DMSO reaction
produces MSIA and CH in high yield,23 a Ðnding which is in
3
contradiction with smog chamber observations. In both the
reactions of OH with DMS and DMSO detailed information
on the distribution of the reaction products with temperature
is not available and there are considerable uncertainties con-
cerning the oxidation mechanism of DMSO. In DMS oxida-
3
2
3
3
4
using standard solutions of the sodium salts of methane sul-
Ðnic and methane sulfonic acids and sulfuric acid.
MSIA, MSA and sulfate were sampled from the reactor
using a system composed of a short TeÑon tube attached to
an empty glass cryogenic U-tube trap immersed in an
ethanolÈliquid nitrogen slush bath ([112 ¡C), a Ñow control-
ler and pump. Because of the low temperature any MSIA,
MSA or H SO passing through the trap is condensed on the
tion systems free from NO the formation of carbonyl sulÐde
x
(OCS)21,24 and methyl thiolformate (MTF: CH SCHO)19,25
3
has been observed.
Results of several Ðeld studies support that DMS oxidation
chemistry, in particular, the ratio of MSA and nss-sulfate
depends on temperature.26
As part of ongoing research projects in this laboratory
aimed at quantifying the contribution of DMS-derived nss-
sulfate to CCN in the marine boundary layer the present
paper deals with laboratory investigations to quantify the gas
phase DMS oxidation products for various atmospheric con-
ditions. We report here the Ðrst detailed product study of the
OH-radical initiated oxidation of DMS as a function of tem-
2
4
walls of the tube. The pump was adjusted to pass reactor air
at a rate of 1 L min~1 through the cold trap and between
10È40 L were sampled. Due to the high reactivity of MSIA
sampling was continuous during the irradiation period rather
than after termination of the irradiation. After collection,
reagent grade water was added to the U-tube to dissolve the
sampled material which was then transferred with washing to
a small storage Ñask. A total of 5 ml reagent grade water was
used for this procedure. The Ñasks were then stored at dry ice
temperature ([78 ¡C) to reduce oxidation of MSIA to MSA.
Generally the samples were analysed as soon as possible after
collection by ion chromatography. The samples were analysed
for MSIA, MSA and sulfate by injecting a 200 lL aliquot of
the collected sample via an injection loop into the IC system.
With the exception of methyl thiolformate all the chemicals
employed were commercial products and were used without
further puriÐcation. Methyl thiolformate was synthesised as
described in Patroescu et al.25 and had a GC purity of [98%.
perature in the absence of NO , i.e. under conditions applic-
x
able to the remote marine atmosphere.
2. Experimental
The product investigations on the OH-radical photo-
oxidation of DMS were performed in an 1080 L quartz glass
reaction chamber equipped with 32 super actinic Ñuorescent
lamps (Philips TL05/40 W: 320 \ j \ 480 nm, j \ 360
max
nm) and 32 low-pressure mercury lamps (Philips TUV 40 W:
j
\ 254 nm). The experimental set-up has been described in
max
detail elsewhere.27 The experiments were performed at a total
pressure of 1000 mbar (O ] N ) with O partial pressures of
3. Results and discussion
The sulfur-containing products observed in the irradiation of
2
2
2
20, 200 and 500 mbar in the absence of NO . Although the
x
chamber has been used previously for experiments with NO ,
x
liberation of NO from the chamberÏs walls on irradiation is
x
DMSÈH O È(N ] O ) reaction mixtures included SO ,
2
2
2
2
2
DMSO, DMSO , MSA, OCS, MTF, MSIA and SO 2~. Fig.
very minor. For example, NO at ppb levels would be enough
2
4
1aÈf shows plots of all the measured yieldÈtime proÐles of (a)
to completely suppress the formation of methyl thiolformate25
SO , (b) DMSO, (c) DMSO , (d) MSA (e) OCS and (f) MTF
and the presence of NO would be indicated by the appear-
2
2
2
as a function of the di†erent temperatures and O partial
ance of the peroxynitrate CH SO O NO .19 None of these
2
3
2
2
2
pressures investigated. In order to better highlight the
e†ects was observed, supporting that formation of NO in the
system is negligible. The reactor can be temperature regulated
x
observed trends with temperature and O partial pressure the
2
yieldÈtime proÐles of (a) SO , (b) DMSO, (c) DMSO , (d)
within the range 283 to 308 K with a precision of ^ 2 K.
Experiments were performed at temperatures of 284, 295 and
2
2
MSA (e) OCS and (f) MTF as a function of temperature for a
constant O partial pressure of 200 mbar are shown separa-
306 K for each of the O partial pressures.
2
2
tely in Fig. 2aÈf. In the Ðgures the yields have not been cor-
The UV photolysis of H O at 254 nm with the mercury
2
2
rected for possible wall loss or further reaction with OH
radicals/other oxidants in the system. Corrected formation
lamps was used to generate the OH radicals:
H O ] hl(j \ 254 nm) ] OH ] OH
yields for SO , DMSO and MTF are discussed later in the
2
2
2
text.
Typical initial concentrations of DMS and H O , were 6.5È
2
2
7.5 and approximately 32 ppm, respectively, where
1
ppm \ 2.46 ] 1013 molecules cm~3 at 298 K.
3.1. General trends in products with temperature and O
2
The concentrationÈtime behaviours of the DMS and pro-
ducts were monitored in situ using long path FT-IR spectros-
copy (Bruker IFS-88, 492 m path length, HgCdTe detector).
Generally infrared spectra were recorded at 1 cm~1 resolution
and 10È15 spectra were collected over irradiation periods of
20È30 min. Concentrations of reactant and products were
mainly determined by computer-aided spectral abstraction
using calibrated reference spectra generated in this laboratory.
The only exceptions were MSIA and inorganic sulfate
which were determined along with MSA using an Odyssey
Ion Chromatography (IC) System from Alltech consisting of a
526 HPLC pump, an ERIS 1000 HP autosuppressor, a 550
conductivity detector and a 7725/7725i rheodyne valve. An
Allsep hydrophilic anion exchange column based on methyl-
acrylate with quaternary amine functional groups (particle
partial pressure
Sulfur dioxide (SO ) and dimethyl sulfoxide (DMSO). From
2
Figs. 1 and 2 it is evident that the observed yield of SO , (i)
2
increases with increasing temperature at constant O partial
2
pressure and (ii) increases with decreasing O partial pressure
2
at constant temperature. Although not quite so evident
because of the fast further oxidation with OH radicals the
converse is true for DMSO. The yield of DMSO, (i) decreases
with increasing temperature at constant O partial pressure
2
and (ii) increases with increasing O partial pressure at con-
2
stant temperature. The shapes of the curves for DMSO are
determined by competition between production of DMSO
from DMS oxidation, loss to the reactor wall and further oxi-
dation of DMSO.
5464
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470

View Article Online
Fig. 1 Plots of the yields of (a) SO , (b) DMSO, (c) DMSO , (d) MSA, (e) OCS and (f) MTF as a function of O partial pressure (20, 200 and
2
2
2
500 mbar) and temperature (284, 295 and 306 K) against reaction time.
The observed behaviour of the yields of SO and DMSO
DMSO . Formation of DMSO is observed in the reaction
2
with variation in temperature and O partial pressure are
2
2
2
system, the yields are highly variable, not generally above 8%
S and tend to be higher at the higher temperatures. The obser-
vations are difficult to interpret. A large fraction of the
DMSO formation can certainly be ascribed to H O ÈDMS
broadly in line with what would be expected from current
mechanistic models, the initial steps of which can be rep-
resented by the following simpliÐed mechanism:
2
2 2
interactions probably heterogeneously at the reactor walls.
~OH ] CH SCH ] CH SCH ~ ] H O
Production of DMSO is observed in DMSÈH O Èair mix-
3
3
3
2
2
2
2 2
tures left in the dark and its formation has also been reported
from the reaction of OH with DMSO.19h22 There is, however,
evidence from work in this laboratory28 and also from the
(abstraction pathway)
~OH ] CH SCH ] CH S(OH)CH
3
3
3
3
laboratory of SÔrensen et al.20 that formation of DMSO in
chamber studies on the reaction of OH with DMSO is mainly
(addition pathway)
2
CH SCH ~ ] O ] products
due to secondary oxidation processes and is dependent on the
initial DMSO concentration. Because of high variability in the
3
2
2
(including SO , MTF, MSA, etc.)
heterogeneous production of DMSO and also uncertainty in
2
2
its formation pathway from the oxidation of DMSO, it is not
CH S(OH)CH ] M ] OH ] CH SCH ] M
3
3
3
3
possible to establish the extent to which DMSO is being
2
CH S(OH)CH ] O ] products
formed from the OH ] DMS reaction. However, if account is
3
3
2
taken of the relatively large heterogeneous contribution and
the relatively low intermediate DMSO concentrations in the
(including DMSO, other products?)
Increasing the temperature and lowering the O partial pres-
present experimental system, the yield of DMSO from
2
sure will favour the abstraction pathway and hence SO for-
2
2
OH ] DMS is not considered to be particularly signiÐcant.
mation, whereas reversing these conditions will favour the
Gas phase methane sulfonic acid (MSA). The levels of gas
phase MSA are not very high and vary between approx-
imately 1 and 4% S. Trends in the MSA data are very difficult
to interpret. In general, the highest values were obtained at
addition channel and DMSO formation. Although the
observed trends are generally in line with the simpliÐed
mechanism, in its present state the model ignores possible for-
mation of SO from further oxidation of DMSO, which is
known to occur.20.22 This aspect is discussed in the section on
formation yields.
the highest temperature and highest O partial pressure and
2
2
the lowest at the lowest temperature and lowest O partial
2
pressure. The di†erences between the results at 295 and 284 K
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470
5465

View Article Online
abstraction channel but also from the addition route, i.e. from
the further gas phase oxidation of DMSO.28
Ion chromatographic analysis (IC) of methane sulÐnic acid
(MSIA), methane sulfonic acid (MSA) and sulfate ions. In con-
trast to MSA the infrared spectrum of methane sulÐnic acid is
not known, therefore, it is not presently possible to establish
the formation of this compound using FT-IR spectroscopy.
Similarly H SO can only be determined by IC. Ion chroma-
2
4
tography was therefore the analytical tool used to establish
the presence of MSIA and inorganic sulfate in the DMS oxi-
dation system. This analytical technique also gives informa-
tion on the formation of MSA (gas ] aerosol) in the oxidation
system.
Presently results are only available for 284 and 295 K. Fig.
3 shows a chromatogram demonstrating the detection and
resolution of the CH SO ~, CH SO ~ and SO 2~ anions
3
2
3
3
4
from MSIA, MSA and H SO , respectively, sampled from a
2
4
DMSÈH O oxidation system at 295 K in 1000 mbar of syn-
thetic air. The yields for the CH SO ~, CH SO ~ and SO 2~
2
2
3
2
3
3
4
anions for measurements performed at 284 and 295 K are col-
lected in Table 1. Since the reaction system was probed con-
tinuously these values are integral yield values for the time
period of the experiment and represent composite values of
production and loss processes. They do not represent the for-
mation yields of the species, however, in the case of MSIA
they represent a lower limit of the formation yield. For 295 K
the integral yield, incorporating all production and loss pro-
cesses of MSIA, is between 4 and 10% S whereas at 284 K it is
much higher at approximately 20% S. The measured yields of
MSIA at 295 K are much higher than the yield of 1.5 ^ 1.1%
S yield reported by SÔrensen et al.20 from studies on the oxi-
dation of DMS at 295 ^ 3 K. This can probably be explained
by the method of sampling used by SÔrensen et al. They col-
lected samples at the end of the irradiation period and not
during the irradiation as in this study. Since MSIA will be
subject to fast further oxidation with OH16 and wall loss pro-
Fig. 2 Plots of the yields of (a) SO , (b) DMSO, (c) DMSO , (d)
2
2
MSA, (e) OCS and (f) MTF as a function of temperature (284, 295 and
306 K) against reaction time for a constant O partial pressure of 200
2
mbar, i.e. atmospheric conditions.
are, however, very marginal and may just reÑect the higher
vapour pressure of MSA at the higher temperature. It is not
possible to extract any meaningful mechanistic information on
the origin of MSA from the results, except to say that the gas
phase production of MSA appears to be generally low at all of
the temperatures. Di†erences in the MSA formation yields
may be masked by several e†ects. Wall loss rates of MSA, for
example, will probably be temperature dependent and may
a†ect the observed MSA yield. Loss of MSA to the aerosol
phase may also be di†erent at the various temperatures.
However, we expect none of these e†ects to be particularly
pronounced over the small temperature range investigated.
This is supported by the large observed formation yields of
SO (see below), any large variation in the gas phase forma-
2
Fig. 3 Example of a chromatogram demonstrating the detection and
resolution of ion chromatographic measurements of (1) MSIA, (2)
MSA and (3) sulfate from samples collected from the irradiation of
DMSÈH O È(O ] N ) mixtures in the photo-reactor. The example
tion of MSA would have led to large variations in the SO
yield.
2
Interpretation has been made even more complicated by
recent work from this laboratory which has shown, contrary
to previous wisdom, that MSA is not only formed from the
2
2
2
2
shown is for a measurement performed at 295 K in synthetic air, i.e.
200 mbar partial pressure of O .
2
Table 1 Ion chromatographic analysis of methane sulÐnate, methane sulfonate and sulfate ions from samples collected from the oxidation of
DMS in 1000 mbar synthetic air at 295 K
CH S(O)O~
CH S(O) O~
SO 2~
3
3
2
4
Expt.
no.
Temperature
284 K
Volume/L
ppm
%S
ppm
%S
ppm
%S
1
2
3
11.8
10.9
11.9
0.37
0.39
0.32
19.24
17.35
18.62
0.33
0.32
0.39
17.41
14.3
23.22
0.81
0.46
0.57
42.06
20.83
33.24
295 K
4
5
6
7
8
9
28.7
41.0
42.4
13.1
11.3
29.4
0.184
0.135
0.218
0.182
0.146
0.423
4.26
2.53
4.62
6.07
4.88
10.26
0.45
0.33
0.24
0.24
0.25
0.24
10.44
6.13
4.98
7.88
8.16
5.89
0.62
0.81
0.81
a
a
a
14.42
15.19
17.32
a
a
a
a Not calibrated at the time of the experiment.
5466
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470

View Article Online
cesses a low yield of MSIA will be obtained if the measure-
ments are conducted at the end of the irradiation. In addition
we have used much lower collection and storage temperatures
which probably further assists in inhibiting the further oxida-
tion of MSIA to MSA by any oxidising material co-collected
in the sampling medium.
Methyl thiolformate (MTF). The observation of the forma-
tion of MTF has been reported several times in the photo-
oxidation of DMS under conditions of low NO .21,24,25 This
x
NO free study conÐrms once again the formation of this
x
compound, which can only be formed from the OH ] DMS
abstraction channel. Possible reactions channels leading to
formation of MTF under the present experimental conditions
include:25
The yields of MSA observed using ion chromatography are
higher than the quantities measured in the gas phase, espe-
cially at low temperature. This is due to, (i) the method of
sampling, i.e. continuous collection of the liquid samples
during irradiation and (ii) oxidation of MSIA to MSA either
in the sampling line or in the liquid sample itself. The mea-
sured values of MSA using the ion chromatography method
can, therefore, not be compared directly with the gas phase
formation yield of MSA. The measured values are composite
values of the gas phase MSA yield plus contributions from the
oxidation of MSIA. Because of the high reactivity of MSIA,
further oxidation to MSA will be very difficult to avoid in
analytical probing systems. The higher values compared to the
gas phase values are a further indication of the importance of
MSIA production in the reaction system. This is particularly
evident from the much higher values at 284 K where oxida-
tion of MSIA to MSA in the sampling line probably becomes
a complicating factor.
OH ] CH SCH ] CH SCH ~ ] H O
3
3
3
2
2
(abstraction pathway)
CH SCH ~ ] O ] CH SCH O ~
3
2
2
3
2 2
CH SCH O ~ ] HO ~ ] CH SCHO ] O ] H O
3
2
2
2
3
2
2
CH SCH O ~ ] CH SCH O ~ ] CH SCH O~
3
2
2
3
2
2
3
2
] CH SCH O~ ] O
3
2
2
CH SCH O ~ ] CH SCH O ~ ] CH SCHO
3
2
2
3
2
2
3
] CH SCH OH ] O
3
2
2
CH SCH O~ ] O ] CH SCHO ] HO ~
3
2
2
3
2
Wall loss of MTF is negligible in the reactor, however, MTF
is photolysed by the mercury lamps and also reacts with OH
radicals. Production of MTF, therefore, competes with pho-
tolysis and reaction with OH and this accounts for the shape
of the MTF yieldÈtime proÐles in Fig. 1 and 2. The yield of
MTF was observed to (i) increase with decreasing temperature
Sulfate (SO 2~) formation was observed with values of
4
approximately 14È17% S at 295 K and 20È40% S at 284 K.
The yields at 284 K were very variable. The fraction of SO
2
oxidised by OH radicals to gas phase H SO is estimated to
2
4
be of the order of 15% at 295 K and 13% at 284 K. From
Table 1 it is evident that, whereas the SO 2~ formation yields
at a constant O partial pressure and (ii) increase with
4
2
at 295 K are of the order expected from OH-radical induced
decreasing O partial pressure at a constant temperature. The
2
oxidation of gas phase SO , those at 284 K are considerably
formation behaviour of MTF with temperature is opposite to
2
higher. As for MSA it is suspected that part of the variability
that observed for SO whereas the direction of the e†ect on
2
in the H SO yields at low temperature is a sampling
changing the O partial pressure parallels that of SO . It is
2
4
2
2
problem. At 284 K more condensation of H O in the sam-
difficult to predict how changes in temperature and O partial
pressure will a†ect the direct formation of MTF, either from
2
2
2
pling line than at 295 K is likely, this could result in more
e†ective oxidation of SO and thus erroneously high sulfate
yields. Because of the possible sampling problems the values
self-reactions of CH SCH O ~ or reaction with HO ~, since
2
3
2
2
2
the chemistry of these reactions is not well established. Under
the experimental conditions of the study self reaction of
CH SCH O ~ is expected to be the dominant fate of this
of the SO 2~ yield should be viewed with caution, particularly
4
at low temperature. Further work is continuing in this area to
3
2 2
better quantify the SO 2~ formation yields in the reaction
radical, which can form CH SCH O~ radicals, MTF and
4
3 2
system, possibly either by heating the sampling line or by use
CH SCH OH. The observed O and temperature e†ects
3
2
2
of a dark, H O and NO free, OH radical source.
argue against direct formation of MTF from self reaction
CH SCH O ~ being the only source of MTF in the reaction
2
2
x
Tests have shown that mixtures of DMS or DMSO and
H O in the dark do not produce MSIA, MSA or H SO ,
3
2 2
system. The observed temperature e†ect is in line with a
simple competition between thermal decomposition of the
CH SCH O~ radical to CH S~ and HCHO and reaction with
2
2
2
4
thus the only route to MSIA in the present experiments
appears to be gas phase OH-radical initiated oxidation of
DMS. The experiments show that there is a large increase in
the measured yield of MSIA on decreasing the temperature
from 295 to 284 K. The decrease in temperature will increase
the importance of the addition pathway and also the forma-
tion of DMSO as evidenced by this study. Therefore, the
results support that the further oxidation of DMSO with OH
will be the major pathway leading to production of MSIA in
the present system:
3
2
3
O forming MTF also being of importance. The further oxida-
2
tion of the CH S~ radical will produce SO . At low tem-
peratures the reverse will be true, reaction with O will be
favoured over the thermal decomposition. The increase in the
3
2
2
yield of MTF with decreasing O partial pressure would at
2
Ðrst seem to be in contradiction with the above reaction
mechanism since the opposite e†ect would be expected, i.e. on
decreasing the O the formation of MTF would be expected
2
to decrease due to the increased competitiveness of the
CH SOCH ] OH ] CH SO(OH)CH
3
3
3
3
thermal decomposition channel. However, it must be borne in
(thermal decomposition)
mind that when the O partial pressure is decreased the frac-
2
tion of the reaction proceeding via the abstraction channel is
] CH S(O)OH ] CH .
considerably enhanced over the addition channel. The magni-
tude of this enhancement is such that any expected reduction
3
3
The present results do not presently allow an exact determi-
nation of the formation yield of MSIA, however, from a
simple consideration of the magnitude of the measured MSIA
yield in combination with the collection efficiency and the loss
processes involved (wall loss, gas phase oxidation, oxidation
to MSA in the liquid sample) the fraction of the
OH ] DMSO reaction leading to MSIA formation must be
fairly substantial. This would be in line with the recent absol-
ute study of Urbanski et al.23 which supports near unit yield
in the MTF yield with reduction in the O partial pressure is
2
probably over compensated by the increase in the fraction of
the reaction proceeding via the abstraction channel. Therefore,
the observed changes in the MTF yield with variation in tem-
perature and O partial pressure is fully in line with its pro-
2
duction, at least in part, via the abstraction channel and
reaction of the CH SCH O~ radical with O . The evidence
3
2
2
although circumstantial is the Ðrst to show the possible exis-
tence of this pathway, the evidence from previous studies has
for the process OH ] DMSO ] MSIA ] CH .
3
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470
5467

View Article Online
always been inconclusive.21,24,25 In the presence of NO, oxi-
dation of CH SCHO ~ to CH SCH O~ followed by thermal
3
2
3
2
decomposition appears to dominate.25 The size of the branch-
ing ratio between thermal decomposition of CH SCH O~ and
its further reaction with O is discussed below.
3
2
2
Carbonyl sulÐde (OCS). The formation yield of OCS was
found to (i) increase with increasing temperature at a constant
O partial pressure and (ii) increase with increasing O partial
pressure at a constant temperature. There was generally very
little di†erence between the results with 200 and 500 mbar of
O , the results with 20 mbar of O are signiÐcantly lower
2
2
2
2
than those with either 200 or 500 mbar O . The yield of OCS
at 295 K in synthetic air is similar to that previously reported
2
from studies in this laboratory on the oxidation of
DMS.19,21,24 The mechanism producing OCS is presently
unknown. The photolysis of MTF is known to produce
OCS.25 Both the temperature and O partial pressure trends
2
of MTF and OCS production are anti-correlated, therefore,
although photolysis of MTF will contribute to the formation
of OCS it is deÐnitely not the only source. Other potential
sources could include production from OH radical attack at
the methyl group of compounds such as CH SCHO and
3
CH SCH OOH.
Fig. 4 Plots of the DMSO yields (corrected for wall loss and reac-
tion with OH radicals; see text) against the consumption of DMS.
3
2
3.2. Formation yields of SO , DMSO and MTF
2
In order to obtain the formation yields for SO and DMSO
their measured concentrations have to be corrected for reac-
tion with OH radicals and loss to the reactor walls. Correc-
tions for atmospheric conditions, i.e. 200 mbar partial
shown in Fig. 4 and 5, respectively, for an O partial pressure
of 200 mbar (atmospheric conditions) and temperatures of
284, 295 and 306 K. The slopes of the lines give the formation
yield for the respective compound and temperature. The cor-
2
2
pressure of O , and temperatures of 284, 295 and 306 K have
rected formation yields for DMSO and SO obtained by least-
2
2
been performed using the mathematical procedure described
squares analyses of these plots are collected in Tables 2 and 3,
by Tuazon et al.29 which is based on the reaction sequence:
respectively. The errors quoted in the tables are the 2p
random scatter in the data plots (Fig. 4 and 5) and reÑect the
k1
OH ] DMS ÈÈÈ Y (SO or DMSO)
2
k2
OH ] (SO or DMSO) ÈÈÈ products
2
k3
(SO or DMSO) ÈÈÈ wall
2
In this sequence Y is the formation yield of either SO or
DMSO from the oxidation of DMS and k , k and k are the
rate coefficients for the reactions indicated above. For the
2
1
2
3
reaction of OH with DMS values of k (overall reaction
1
abstraction ] addition) have been calculated for the di†erent
temperatures using the recommended literature values.18,30
For the reaction of OH with SO values for k have been
2
2
calculated from the Arrhenius expression 5.07 ] 10~13
exp(231/T ) cm3 molecule~1 s~1 used in modelling
studies.16,31 Wall loss for SO was very minor, dark experi-
ments showed that its loss over the time period of the mea-
2
surements was within the estimation detection error limits.
This allowed a value of \1 ] 10~6 s~1 to be put on the wall
loss rate k for SO .
3
2
The rate coefficient for the reaction of OH with DMSO is
not very well established and is associated with a high degree
of uncertainty.23,32 For the purposes of the yield correction
for OH ] DMSO the Arrhenius expression 10~11.2B0.7
exp(800 ^ 540/T ) cm3 molecule~1 s~1 from Hynes and
Wine32 has been used to calculate k (OH ] DMSO) for the
2
individual temperatures. Wall loss of DMSO is also quite sig-
niÐcant, Ðrst order loss rates of k of 7 ] 10~5 s~1 for 284 K
3
and 1 ] 10~4 s~1 for 295 and 306 K, determined from wall
loss experiments in the dark, have been used for the yield cor-
rection calculations. Due to extensive pumping and purging of
the system between experiments, variation of the rate between
the di†erent experiments was minor, indicating that condi-
tioning of the chamber walls, which could a†ect the loss rates,
was not a problem.
Plots of the DMSO and SO concentrations, corrected for
reaction with OH radicals and wall loss, against *[DMS] are
Fig. 5 Plots of the SO yields (corrected for wall loss and reaction
2
2
with OH radicals; see text) against the consumption of DMS.
5468
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470

View Article Online
Table 2 Corrected yields for the formation of DMSO in the reaction
of ~OH with DMS in 1000 mbar of synthetic air as a function of tem-
perature (see text for details)
However, both the results from this study and that of
Turnipseed et al.33 support that the branching ratio of the
path leading to DMSO is quite substantial.
The corrected formation yields for SO in synthetic air as
function of temperature are compared in Table 3 with the
Contribution
2
Corrected DMSO
yieldsa
(% molar yield, ^2p)
of addition
pathway
(%)
fractions of the OH ] DMS reaction proceeding via the addi-
tion and abstraction pathways taken from the data of Hynes
et al.18 The formation yields are all high, the yield increases
progressively from 84.3 ^ 6.5 at 284 K, to 95 ^ 3.8 at 295 K
Temperature/K
284
295
306
46.3 ^ 5.0
34.8 ^ 7.6
24.4 ^ 2.8
52
33
17
to 99.0 ^ 6.5 at 306 K. The accuracy of the SO formation
2
yields is much superior to those for DMSO since the rate co-
a The yields are compared to the fraction of the reaction occurring via
the addition pathway calculated using the results of Hynes et al.18 for
the speciÐc reaction conditions. See text for details of the correction
calculation.
efficient for k (OH ] SO ) is much better established than k
2
2
2
(OH ] DMSO).
From
a
comparison of the relative fractions of the
abstraction and addition channels in Table 3 it is obvious
from the high measured formation yields of SO that it is
2
being formed via both channels. The high formation yields
extent of the measurement errors in SO , DMSO and
*[DMS] using FT-IR spectroscopy. Uncertainties in the rate
coefficients used for the correction are not included.
2
also infer that under the NO free conditions of the experi-
ments the principal fate of DMSO, a major addition pathway
product, is oxidation to SO . Since the present results also
support that MSIA is a major DMSO oxidation product this
further implies that the predominant fate of MSIA in the reac-
tion system is oxidation via reaction with OH radicals to form
x
In Table 2 the formation yields of DMSO are compared
with the fraction of the OH ] DMS reaction occurring via the
addition pathway calculated from the data of Hynes et al.18
The DMSO yield values for 284 and 295 K are approximately
the same as the fraction of the reaction occurring via the addi-
tion channel while the yield value for 306 K is somewhat
higher. The numbers, as they stand, suggest that virtually all
of the DMSÈOH adduct formed in the OH ] DMS addition
2
SO , presumably via the reaction sequence:
2
CH S(O)OH ] ~OH ] CH S(O)O~ ] H O
3
3
2
CH S(O)O~ ] CH ~ ] SO
3
3
2
pathway is reacting with O to form DMSO. Turnipseed et
Any MTF formed in the reaction system will be photolysed
and also further oxidised by reaction with OH radicals; wall
loss is negligible.
2
al.,33 in a pulsed laser photolysis/pulsed laser-induced Ñuores-
cence study on the reaction of OH with DMS, have reported a
branching ratio of U \ 0.5 ^ 0.15 at 234 and 258 K for HO
2
2
production from the CH S(OH)CH ] O reaction, the HO
k1
3
3
2
~OH ] DMS ÈÈÈ Y (MTF)
being formed via abstraction of the hydroxy H atom. The
co-product in the study of Turnipseed et al. was assumed to
be DMSO in equivalent yield. The yield of 50% from the
work of Turnipseed et al. is considerably lower than the yield
of near unity which the results from the present study would
suggest. The work of Turnipseed et al. was performed at low
temperature and pressure, however, their work suggested that
k2
~OH ] MTF ÈÈÈ products
k3
MTF ] hl ÈÈÈ products
Unfortunately the rate coefficient for the reaction of MTF
with OH, k , is known only for room temperature, therefore a
formation yield correction was only performed on the results
for 295 K in 1000 mbar of synthetic air. A value of k
2
the HO yield was temperature independent and there is no
2
reason to expect a signiÐcant pressure dependence for the
2
DMSÈOH] O reaction. The formation yields for DMSO
(OH ] MTF) \ 1.11 ] 10~11 cm3 molecule~1 s~1 taken from
2
which we calculate are very sensitive to the value used for k
Patroescu et al.25 was used. The photolysis frequency for
2
(OH ] DMSO). As stated above the absolute value of k
2
MTF under the conditions of the experiment was k \
3
(OH ] DMSO) is associated with a high degree of uncertainty
5 ] 10~3 s~1, this value was determined in earlier work on
which has not been incorporated into the formation yield cal-
culation. Interestingly, if the results of Turnipseed et al. are
correct this implies a lower value for the rate coefficient of the
reaction of OH with DMSO and suggests previous determi-
nations of the rate, both absolute and relative, have been
MTF in this laboratory. A plot of the MTF concentration
corrected for reaction with OH radicals and photolysis
against *[DMS] for an O partial pressure of 200 mbar and
295 K (not shown) was linear. A formation yield of 7.2 ^ 1.4%
2
has been obtained for 295 K.
subject to artefacts. More precise values of k (OH ] DMSO)
As outlined above there are several potential pathways
which can be responsible for the formation of MTF all stem-
ming from further reactions of the CH SCH O ~ radical. All
of these processes may be occurring in the present system and
it is not presently possible to di†erentiate between the di†er-
ent channels. However, the observed sensitivity of the MTF to
2
as a function temperature would greatly help to improve the
accuracy of the formation yield data for DMSO obtained in
this work for the OH ] DMS reaction. Until such values
become available the exact yield of DMSO from reaction of
3
2 2
O
with the DMSÈOH adduct is still open to question.
2
Table 3 Corrected yields for the formation of SO in the reaction of ~OH with DMS in 1000 mbar of synthetic air as a function of temperature
2
(see text for details)
Contribution
of addition
pathway
(%)
Contribution
of abstraction
pathway
Corrected SO
yieldsa
(% molar yield, ^ 2p)
2
Temperature/K
(%)
284
295
306
84.3 ^ 6.5
95.0 ^ 3.8
99.0 ^ 6.5
52
33
17
48
67
83
a The yields are compared to the fractions of the reaction occurring via the addition pathway and abstraction pathways calculated using the
results of Hynes et al.18 for the speciÐc reaction conditions. See text for details of the correction calculation.
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470
5469

View Article Online
3
4
P. A. Spiro, D. J. Jacob and J. A. Logan, J. Geophys. Res., 1992,
97, 6023.
H. Berresheim, P. H. Wine and D. D. Davis, in Composition,
Chemistry and Climate of the Atmosphere, ed. H. B. Singh, Van
Nostrand Reinhold, New York, ISBN 0-662-01264-0, 1995, pp.
251È302.
R. J. Charlson, J. E. Lovelock, M. O. Andreae and S. G. Warren,
Nature, 1987, 326, 655.
W. C. Keene, R. Sander, A. A. P. Pszenny, R. Vogt, P. J. Crutzen
and J. N. Galloway, J. Aerosol Sci., 1998, 29, 339.
Atmospheric Chemistry and Physics, ed. J. H. Seinfeld and S. N.
Pandis, Wiley, New York, 1998.
D. Katoshevski, A. Nenes and J. H. Seinfeld, J. Aerosol Sci., 1999,
30, 503.
D. M. Murphy, J. R. Anderson, P. K. Quinn, L. M. McLnnes,
F. J. Brechtel, S. M. Kreidenweis, A. M. Middlebrook, M. Pofai,
D. S. Thomson and R. P. Buseck, Nature, 1998, 392, 62.
changes in temperature and O partial pressure suggest that
2
competition between thermal decomposition of CH SCH O~
3
2
radicals and reaction with O may be a major source of MTF
2
formation in the present reaction system. Since the yield of
MTF is fairly low this implies that under conditions in the
atmosphere where the chemistry of CH SCH O ~ is not con-
5
6
7
8
9
3
2 2
trolled by reaction with NO, i.e. the remote MBL, the pro-
duction of MTF will also be low. It would be interesting to
calculate the results for lower temperatures since the impor-
tance of the MTF product pathway from CH SCH O~ ] O ,
3
2
2
if operative, should increase with decreasing temperature.
Unfortunately, this can only be resolved when temperature
kinetic data for OH ] MTF become available. The above
considerations are presently still very speculative. In order to
properly gauge the importance of the various MTF producing
pathways from CH SCH O ~ reactions it will be necessary to
10 C. OÏDowd, M. H. Smith, I. E. Consterdine and J. A. Lowe,
Atmos. Environ., 1997, 31, 73.
3
2 2
investigate the various channels individually.
11 A. Pszenny, W. Keene, C. OÏDowd, M. Smith and P Quinn,
IGACtivities News L ett., 1998, 11, 6.
12 F. Campolongo, A. Saltelli, N. R. Jensen, J. Wilson and J. Hjorth,
J. Atmos. Chem., 1999, 32, 327.
4. Summary
13 A. R. Ravishankara, Y. Rudich, R. Talukdar and S. B. Barone,
Philos. T rans. R. Soc. L ondon, Ser. A, 1997, 332, 171.
14 S. B. Barone, A. A. Andrew and A. R. Ravishankara, Faraday
Discuss., 1995, 100, 39.
The OH radical initiated oxidation of DMS has been investi-
gated as a function temperature and O partial pressure. For-
2
mation of SO , DMSO, DMSO , MSIA, MSA, OCS, MTF
2
2
and SO 2~ was observed. The observed formation behaviour
of SO and DMSO with variation in temperature and O
partial pressure is in line with current mechanistic models
involving addition and abstraction channels. The data support
that DMSO, formed in the reaction of a DMSÈOH adduct
15 G. S. Tyndall and A. R. Ravishankara, Int. J. Chem. Kinet., 1991,
23, 483.
4
2
2
16 F. Yin, D. Grosjean and J. H. Seinfeld, J. Atmos. Chem., 1990, 11,
309.
17 A. J. Hynes, R. B. Stoker, A. J. Pounds, T. McKay, J. D. Brad-
shaw, J. M. Nicovich and P. H. Wine, J. Phys. Chem., 1995, 99,
16,967.
with O , is the major product of the OH ] DMS addition
2
channel. The observation of high yields of methane sulÐnic
18 A. J. Hynes, P. H.Wine and D. H. Semmes, J. Phys. Chem., 1986,
acid also support that further OH-radical induced oxida-
tion of DMSO results in formation of the acid in substantial
yield. Preliminary ion chromatographic analyses of
DMSOÈH O Èair photolysis systems show quite deÐnitely
90, 4148.
19 I. V. Patroescu, I. Barnes, K. H. Becker and N. Mihalopoulos,
Atmos. Environ., 1999, 33, 25.
20 S. SÔrensen, H. Falbe-Hansen, M. Mangoni, J. Hjorth and N. R.
Jensen, J. Atmos. Chem., 1996, 24, 299.
2
2
that methane sulÐnic acid is formed in the reaction of OH
with DMSO, more experiments need to be performed before
information on the yield can be given. However, the experi-
ments here on DMS and the preliminary work on DMSO
show that in assessing the contribution of DMS to CCN for-
mation in the remote MBL consideration of the fate of MSIA
21 I. Barnes, K. H. Becker and I. Patroescu, Atmos. Environ., 1996,
30, 1805.
22 I. Barnes, V. Bastian, K. H. Becker and D. Martin, in Biogenic
Sulfur in the Environment, ed. E. S. Saltzman and W. J. Cooper,
ACS Symp. Ser. 393, American Chemical Society, Washington,
DC, 1989, p. 476.
23 S. P. Urbanski, R. E. Stickel and P. H. Wine, J. Phys. Chem.,
(oxidation to form SO or MSA, formation of new particles or
1998, 102, 10,522.
2
uptake by existing particles) is an important aspect in the
24 I. Barnes, K. H. Becker and I. Patroescu, Geophys. Res. L ett.,
1994, 21, 2389.
chemistry of DMS which needs to be taken into account.
MSIA chemistry is presently not adequately represented in
atmospheric chemical models. The results of the present study
suggest that the role of the chemistry of MSIA in controlling
the contribution of DMS to CCN formation in the remote
MBL will be become increasingly important with decreasing
temperature since the importance of the DMSO producing
channel increases with decreasing temperature.
25 I. Patroescu, I. Barnes and K. H. Becker, J. Phys. Chem., 1996,
100, 17,207.
26 T. S. Bates, J. A. Calhoun and P. K. Quinn, J. Geophys. Res.,
1992, 97, 9859.
27 I. Barnes, K. H. Becker and N. Mihalopoulos, J. Atmos. Chem.,
1994, 18, 267.
28 H. Beyer, Diploma Thesis, Bergische Universitat Wuppertal, June
1998.
29 E. C. Tuazon, H. MacLeod, R. Atkinson and W. P. L. Carter,
Environ. Sci. T echnol., 1986, 20, 383.
30 R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, Jr., J. A.
Kerr, M. J. Rossi and J. Troe, J. Phys. Chem. Ref. Data, 1997, 25,
1329.
Acknowledgements
Financial support for this work by the European Commission
and the BTMF within the AFS project is gratefully
acknowledged.
31 O. Hertel, J. Christensen and ^. Hov, Atmos. Environ., 1994, 28,
2431.
32 A. J. Hynes and P. H. Wine, J. Atmos. Chem., 1996, 24, 23.
33 A. A. Turnipseed, S. B. Barone and A. R. Ravishankara, J. Phys.
Chem., 1996, 100, 14,703.
References
1
2
M. O. Andreae, Marine Chem., 1990, 30, 1.
T. S. Bates, B. K. Lamb, A. Guenther, J. Dignon and R. E.
Stoiber, J. Atmos. Chem., 1992, 14, 315.
Paper 9/07211J
5470
Phys. Chem. Chem. Phys., 1999, 1, 5463È5470