Methane Sulfonic Acid in the Marine Atmosphere

Methane sulfonic acid (MSA) is an oxidation product of the reaction of OH radical with dimethyl sulfide and, hence, should be an important constituent of marine air. MSA concentrations in marine aerosols ranged from 0.009 to 0.075 #g/m 3 in samples from the Pacific and Indian oceans and Miami, Florida. In the samples from remote areas (Pacific and Indian oceans), MSA levels averaged 6.7% (S = 1.9) of the non-sea-salt (nss) SO,(cid:127) = values. In the Miami area, ratios were occasionally lower because of the impact of local sulfur emissions (probably pollutant SO2). MSA concentrations in seven rainwater samples collected at Miami, Florida, ranged from 0.001 to 0.034 ppm. Cascade impactor samples from Miami, Florida, and the Gulf of Mexico indicate that MSA occurs primarily in the smaller particles, as does nss SO½ =. In the two samples analyzed, MSA is concentrated on slightly coarser particles than nss SO½ =. The size distributions of both MSA and nss SO½ = differ markedly from that of sea salt Na +, suggesting that at least the first step in the oxidation of organosulfur compounds is a gas phase reaction. The observed concentrations of MSA in marine air are the result of formation from gas phase reaction of dimethyl sulfide (and other volatile organosulfur compounds) with OH radical and destruction by aqueous phase reaction with OH radical in aerosols. Preliminary experiments indicate that MSA oxidation probably yields SO½ = under atmospheric conditions. Therefore, MSA destruction is a potentially important pathway for the formation of nss SO½ = in the marine atmosphere. The magnitude of the fluxes involved in the organosulfur cycle cannot be calculated from the MSA data because of the uncertainty in the free radical chemistry of aerosols.


INTRODUCTION
The presence of non-sea-salt (nss) sulfate in aerosols throughout the marine troposphere [-Georqii, 1978; M•z•iros, 1978J suggests that there is a significant flux of reduced sulfur compounds from the sea surface to the atmosphere. Recent measurements of reduced sulfur compounds in the oceans strongly support these speculations and indicate that the estimated flux of dimethyl sulfide (DMS), the most abundant of these compounds, is sufficient to account for most, if not all, of the background nss sulfate in marine aerosols [Barnard et al., 1982;Andreae and Raemdonck, 1983].
Kinetic studies indicate that reaction with OH radical may be a major pathway in the oxidation of DMS. Assuming an OH concentration of 106 molec/cm s, the reported rate constants 8.3 x 10 -•2 [Kurylo, 1978] and 1.5 x 10 -• [Cox, 1975] cm3/molec s yield lifetimes of DMS against destruction by OH of 33.5 and 18.6 hours. Recent experimental studies, Hatakeyama et al., 1982;Hatakeyama and Akimoto, 1983;Grosjean and Lewis, 1982;Niki et al., 1983] showed that the reaction of OH and DMS resulted in the formation of substantial amounts of both methane sulfonic acid (MSA) and SO2. These studies suggest that MSA should be a major component of the sulfur in the marine troposphere. Methane sulfonic acid has been identified in continental aerosols [Panter and Penzhorn, 1980] but has never been measured, to our knowledge, in maritime air masses.
In this report we present measurements of the concentration of methane sulfonic acid in marine aerosols and rain. From this data we assess the importance of this species both in the depositional flux of sulfur to the oceans and as an intermediate oxidation product of DMS in the formation of aerosol sulfate. Prospero, 1982a]. Assuming a particle density of 1.2 g/cm s (sea salt at 80% relative humidity), the 50½/0 cutoff diameters for stages 1-6 were 16.7, 8.4, 4.2, 2.1, 1.0, and 0.5 microns, respectively. Rainwater samples were collected on an event basis by using an automatic wet-dry collector; the wet-side collection bucket was exposed only when it was raining. For analysis, quarter sections of the Whatman 41 filters were extracted with 20 ml of Milli-Q water (16 Mohm cm). The entire polycarbonate impaction sheets were extracted ultrasonically in 25 ml of Milli-Q water. All the extracts were filtered through pre-washed 0.45 ttm filters and, along with the rainwater samples, were maintained at 5øC until analyzed. In addition to MSA, aerosol extracts and rainwaters were analyzed for sodium, nitrate, and sulfate' sodium to ___2% by atomic absorption spectrometry and nitrate and sulfate to q-5% by ion chromatography [Savoie and Prospero, 1982a].

SAMPLING AND ANALYSIS
Nss SO,• = values were calculated from Na + and SO,• = concentrations by assuming that Na + is derived exclusively from seawater, which has a SO,• =/Na +of0.2517.

Previous measurements of MSA have involved esterification
to the methyl ester (CHsSOsCHs), which was then detected and quantified by using a gas chromatograph with a flame photometric detector and/or a mass spectrometer [Penzhorn and Filby, 1976]. Because MSA is a relatively strong acid and, thus, conductive in aqueous solutions, ion chromatography appeared to be a simple and viable alternative; this proved to be the case. MSA was determined by using a Dionex model 10 Ion Chromatograph with a 0.8 ml sample volume, 3 ml/min flow rate, and 0.0003 M NailCOs eluent. Calibration curves were prepared as concentration versus the peak areas as mea- sured by a Hewlett-Packard 3390A integrator. In this configuration, the detection limit was about 1 ppb with standard errors of _+ 5%. With these parameters, good separation was achieved between MSA and possible interferents such as low molecular weight organic acids (primarily formate and acetate) and fluoride (see Figure 1). No MSA was observed in filter blanks.
The presence of MSA in environmental samples was independently verified by esterification with diazomethane and detection of the resultant methyl sulfonate by GC-MS.

STABILITY OF MSA SAMPLES
A simple experiment was performed to test the stability of MSA on the filters during sampling. Air was drawn through two pump and filter systems which were run simultaneously. In one system the filter was changed at intervals of 1-3 days. In the other, the filter was left in place for 8 days. If MSA was produced or destroyed on the filters by some component in the air, the effect should be most pronounced on the continuous sample, as it was exposed to a considerably greater volume of air than the other samples. The results, Table 1    ser particles. This difference has been noted by numerous authors previously and is explained by a sea-salt origin for Na + versus a "gas to particle conversion" origin for nss SO,• =. Size distribution data have been used to argue against a heterogeneous origin for nss SO,• =, as this would lead to a correlation between nss SO,•= and particle surface area, which is not observed.

MSA, like nss SO,• =, is concentrated near the lower end of the size spectrum. However, in detail there appear to be significant differences between their size distributions. In the Key
Biscayne sample, the MSA mass distribution is accurately (within analytical error) defined by a single peak with a lognormal distribution that has mass median diameter of 0.7 gm and a standard geometric deviation of 4.2. In contrast, the MSA distribution in the Gulf of Mexico sample appears to be bimodal. While a log-normal distribution similar to that found in Miami holds for the larger particle size portion, there is, in addition, a definite peak in the smallest size fractions.
If both MSA and nss SO,• = are formed by homogeneous gas phase reactions, the observed differences in their size distributions is somewhat surprising. The physical processes of particle formation and coagulation must be similar for the two species because they have very similar physical properties. Chemically, however, there is a significant difference; SO,• = is stable in the aerosol, while MSA is not. In a subsequent section of this paper we discuss evidence for the atmospheric destruction of MSA via reaction with OH radical. We expect the rate of MSA oxidation to vary with particle size primarily because of the concentration of CO3 = and C1-in the coarser sea-salt particles. These species react rapidly with OH, lowering the OH concentration and therefore slowing the rate of MSA oxidation in the larger particles. This stabilization process would tend to shift the MSA peak to the coarser end of the size spectrum. In terms of this process, the Miami sample is clearly more evolved than that of the Gulf of Mexico sample. In the Gulf of Mexico sample the higher MSA concentration and the presence of an MSA peak in the smaller particles suggest either that the sample was taken close to the dimethyl sulfide source or that the OH concentrations were low. The extent to which aerosols exhibit the MSA shift to larger particles is probably a function of those chemical and meteorological factors that control the oxidation rate of MSA and its precursors and the residence time of aerosols in a given air mass.
Because only two sample sets have been analyzed, the above discussion is clearly speculative. Further field measurements and laboratory experiments are needed to confirm the observed distributions and to test the hypothesis.

Rainwater
Concentrations of Na+, CI-, NO 3 -, total SO½ = nss SO½ =, and MSA in seven Miami rainwater samples are listed in Table 3. MSA is present in all samples at concentrations ranging from 1 to 34 ppb. The MSA/nss SO½ = ratio ranged from 0.4 to 2.1%, values similar to those in the Miami aerosol data previously discussed. Therefore, the origin of both the MSA and nss SO½ = in these samples may be the scavenging of aerosols within and below the clouds. If droplet uptake of gaseous species is important, we would expect to see lower MSA/nss SO½ = ratios in rainwater than in aerosols, due to the more rapid dissolution and oxidation of SO2 than of DMS. However, the variability of the nss SO,• = in Miami aerosols is large and may obscure the signal from such reactions.

Atmospheric Destruction of Methane Sulfonic Acid
The results of this study show that MSA is present in remote marine aerosols at levels approximately 5% that of nss  We have performed some preliminary experiments that suggest that reaction with OH radical in aerosols and cloud droplets is probably the major atmospheric pathway for the destruction of atmospheric MSA. No detectable reaction occurred between MSA and H202 or 03 in aqueous solution. In concentrated 03 solutions a small amount of SO,• = was produced from MSA; we attribute the production of SO,• = in this case to the reaction of MSA with OH radicals generated by O3 decomposition. Lind and Eriksen [1975]  The preceding calculations are presented to demonstrate that the reactivity of MSA may be an important factor controlling ambient MSA levels and that SO,•--formation via MSA may be a significant process in the marine atmosphere.
A direct calculation of the fluxes involved will require independent assessments (model or measured) of aerosol OH concentrations and models that incorporate gas phase, cloud water, and aerosol-free radical chemistry.
Implications for the Ori•lin of Maritime S02 and nss SO4 = Do the observed MSA concentrations over the oceans support the idea that DMS is the major source of background nss sulfur compounds in marine air9. One way to test this hypothesis is to compare the ambient ratio of MSA to SO2 to the ratio obtained by experimental oxidation of DMS by OH radical.
We estimate the marine background molar ratio of MSA to SO2 to be roughly 1'3.7 assuming a mean MSA concentration of 0.04 #g/m 3 and a mean SO2 concentration of 0. i #g/m 3 [Bonsan•l, et al., 1980]. This ratio may be increased slightly if we consider that at least some of the SO2 is derived from the background CS2 and COS. The lifetimes of these compounds are sufficiently long that they are uniformly distributed in the troposphere, and their source may be both biological (terrestrial and marine) and anthropogenic. Bandy and Maroulis