Environ. Sei. Technol. 1984, 18, 460-468
Photooxidation of Methyl Sulfide, Ethyl Sulfide, and Methanethiol Daniel Grosjean* Environmental Research & Technology, Inc., Westlake Village, California 9 1361
Products of sunlight-irradiated mixtures of oxides of nitrogen and alkyl sulfides (RSR, R = CH,, C,H,) and methanethiol (CH3SH) in air include formaldehyde (R = CH,), acetaldehyde and PAN (R = C2HS),sulfur dioxide, and alkyl nitrates (RONO,) as well as particulate alkanesulfonic acids (RS0,OH) and inorganic sulfate. The nature and yields of gaseous and particulate products are discussed in terms of OH-initiated reaction pathways, including C-S bond scission, and subsequent reactions of alkylthiyl radicals (RS), including those leading to photolabile RSNO and stable RSNO, products for which indirect evidence is presented. SOzyields are found to vary according to the relative importance of the competing pathways RS + 0, (a) and RS + NO, (b), for which a ratio k b / k , 2 X lo6 is derived from data for irradiated RSRNO,, RSH-Cl,, and RSH-C1,-N02 mixtures.
-
Introduction The atmospheric photochemistry of sulfur-containing organic compounds such as thiols (RSH, R = alkyl groups), sulfides (RSR, RSR’), and disulfides (RSSR, RSSR’) is relevant to a number of important environmental issues. As anthropogenic pollutants, organosulfur compounds are relevant to both existing (e.g., petrochemical) and emerging (e.g., synthetic fuels) energy production technologies (1-3). As trace atmospheric constituents of biogenic origin, organosulfur compounds have a significant role in the atmospheric sulfur cycle and global sulfur budget (4-8). Furthermore, the recently documented formation of the strong acids sulfuric acid and methanesulfonic acid upon photooxidation of methyl sulfide (9,10) is relevant to the currently much-investigaed issue of precipitation chemistry, including precipitation acidity in “remote” areas (11). Expanding upon initial studies conducted with methyl sulfide (10, 12),the present work includes a more comprehensive investigation of the atmospheric chemistry of the organosulfur compounds methyl sulfide, ethyl sulfide, and methanethiol, with emphasis on gaseous and particulate reaction products and their modes of formation. The corresponding studies of physical processes involved in aerosol formation are reported elsewhere (12-14). This work has been carried out as part of a larger study of the photochemical and physical aspects of organic aerosol formation from a number of precursor hydrocarbons (12-1 4). Experimental Section Sunlight irradiations of organosulfur compound-NO, mixtures were carried out in outdoors chambers. The outdoor environmentalchamber facility has been described in detail elsewhere (15),and only a summary of the protocol is given here. Large (80 m3, initial surface to volume ratio -2.5 m-l) and smaller (4 m3) chambers were constructed from FEP 200A Teflon film. With the chamber covered with opaque plastic film, the reactants NO, NO,, and organosulfur compound were introduced in that order
* Address corespondence to this author at Daniel Grosjean and Associates, Inc., Suite 645, 350 N. Lantana Street, Camarillo, CA 93010. 460
Environ. Sci. Technol., Vol. 18, No. 6, 1984
into the matrix air provided by an Aadco Model 737-14 air purification system. After mixing, the opaque cover was removed, and the mixture was expossed to sunlight for several hours. Parameters measured on-line included oxides of nitrogen, ozone, organosulfur compound, sulfur dioxide, peroxyacetyl nitrate (PAN), temperature, humidity, and solar ultraviolet radiation intensity. The corresponding instruments and calibration methods are listed in Table I. Off-line samples collected from the chamber included impinger samples for gaseous carbonyls and alkanesulfonic acids and filter samples for particulate carbon, sulfate, nitrate, and alkanesulfonates. The corresponding analytical methods are also listted in Table I. Control experiments aimed at characterizing the chamber behavior with respect to trace atmospheric pollutants were performed at regular intervals. These control experiments included irradiations of pure air, studies of the stability of reactants and products in purified air, and “memory effects” runs involving irradiation of pure air following several NO,-organosulfur compound irradiations. A summary of typical results from these control experiments is given in Table 11. Studies of aerosol losses on the chamber walls as a function of particle size have also been performed (14, 23). Nitric oxide (lecture bottle; Scott Environmental Technology), methyl sulfide (298% purity; Aldrich Chemical Co.), ethyl sulfide (198% ;Aldrich), methanethiol (54.6 ppm in N,; Scott Environmental Technology), methanesulfonic acid (Aldrich; purity 98%), and ethanesulfonic acid (Aldrich) were employed without further purification. The purity of methyl sulfide, as determined by gas chromatographic analysis on a Porapak QS column and using a sulfur-selective Hall detector, was >99%. In irradiations involving chlorine (lecture bottle; Scott Environmental Technology), a gas mask was worn during handling and injection steps. Measurements of sulfur compounds involved the use of a pulsed fluorescent analyzer whose response is specific to SO, and of a total sulfur flame photometric analyzer whose response was calibrated with NBS-traceable SO2 (45.7 ppm in air in a passivated aluminum cylinder), methyl sulfide (two independent sources, a GC Industries permeation device and a certified gas cylinder, 100 ppm in N,) and methanethiol (certified gas cylinder, 54.6 f 2.7 ppm in N,; Scott Environmental Technology). The flame photometric total sulfur analyzer response was different for each sulfur compound, i.e., 0.91 for SO,, 0.82 for methyl sulfide, and 0.77 for methanethiol (all relative to the pulsed fluorescent instrument’s response to SO2 = 1.00). The pulsed fluorescent analyzer was calibrated with NBStraceable SOz according to US.Environmental Protection Agency recommended procedures (24). We verified experimentally that, as expected (24), organosulfur compounds do not interfere with SO, measurements by pulsed fluorescence at the concentrations studied. By use of the proper calibration curves, the organosulfur compound concentration was thus taken as the difference between total sulfur and SO, concentrations. We also verified by comparing instrument readings obtained directly at the chamber sampling port to those obtained downstream of the sampling manifold that no measurable losses of SO2
0013-936X/84/0918-0460$01.50/0
0 1984 American Chemical Society
Table I. Summary of Parameters Measured and Analytical Methods
on-line NO, NO2 ozone
so2
total sulfur PAN, CH30N02 temperature dew point sunlight intensity aerosol size distribution off-line formaldehyde and other carbonyls particulate sulfate, nitrate, and alkanesulfonic acids nitric acid particulate carbon aerosol molecular composition
comment/ ref
instrument and method
parameter
ThermoElectron 14 BE chemiluminescent analyzer Dasibi 1003-AH ultraviolet photometer Monitor Labs 8850 pulsed fluorescent analyzer Meloy SA-160 flame photometric sulfur analyzer electron capture gas chromatography YSI 741A-10 sensor EG&G 88041 hygrometer Eppley ultraviolet radiometer TSI 3030 electrical aerosol analyzer/dedicated microcomputer; Climet 208 optical particle counter with pulse height analyzer DNPH impinger/high pressure liquid chromatography (Altex 332 HPLC with ultraviolet detector) Teflon filter/ion chromatography (Dionex Model 10 ion chromatograph) nylon filter/ion chromatography quartz filter/combustion (ERT-modified Dohrmann DC-50 carbon analyzer) filter extract/mass spectrometry (Kratos MS-25 with electron impact and chemical ionization sources)
15-18, e
f 20
21, 22
"In the NO2mode, the analyzer responds nearly quantitatively to nitric acid and PAN. In the NO mode, organosulfur compounds introduce a small positive interference of -0.1 %. bSpecificto SO2, negligible interference from organosulfur compounds. Measures sum of SO2and organosulfur compound, with a different response to each sulfur compound. dCalibratedwith PAN synthesized from irradiated mixtures of chlorine, acetaldehyde, and NOz. CH30N02observed but not quantitated. eConfiimationof the carbonyl structure was obtained by mass spectrometry analysis of their 2,4-dinitrophenylhydrazones(19). f See details under Experimental Section. Table 11. Summary of Outdoor Chamber Control and Reference Experiments Relevant to This Study (1)Pollutant Stability Studies
pollutant
range of concn, ppb
loss rate, % h-l
ozone, dark ozone, sunlight NO,, dark NO,, sunlight Nitric acid, dark SO2, dark PAN, darkb PAN, sunlightb
20-1500 10-1100 10-700 120-570 125 355 50-310 60-380
0.8-1.4"
3.3-4.7" 0.15 1.5 9 0.8 1.2
3.8
(2) Purified Air Irradiations (3-8 h in Sunlight)
net O3 initial NO,, maximum Os, formation rate, PPb PPb PPb 6-24 2-2 1 0.3-7.0 range mean (19 runs) 13 7 1.6 "Lower value is for large chambers. Ozone loss rates were independent of matrix air humidity in the range 040% RH. the presence of NO2 and with NO2 >> NO. (80-900 ppb) and methyl sulfide (80 ppb) occurred in the sampling lines (25 f t X in. diameter Teflon TFE) and glass sampling manifold. In our earlier work (10, 12) methanesulfonic acid was identified as a product of methyl sulfide photooxidation by using solvent extraction of filter samples, esterification of the extract, and mass spectrometry by comparison with electron impact and chemical ionization spectra of authentic samples. For this study, we developed a simpler procedure (25) involving filter sample extraction by sonication for 30 min with a buffer solution (17 X M NaZCO3and 22 X M NaHCO,) followed by ion chro-
matographic analysis of the carbonate buffer extract. Under these conditions, base-line separation was achieved between methanesulfonate (or ethanesulfonate) and other ions thay may interfere as products in organosulfur-NO, irradiations (see Discussion), namely, formate, acetate, nitrate, chloride, and inorganic sulfate. Calibration curves were prepared from standard solutions of methane- and ethanesulfonic acids in the carbonate buffer. The analytical detection limit was 0.1 pg/mL. For a 30-min filter sample collected at 40 L min-l and extracted with 10 mL of buffer, the corresponding detection limit was 0.8 pg m-,. In an experiment conducted with, for example, 300 ppb of methyl sulfide, the above detection limit was sufficient to detect a -0.1% yield of methanesulfonic acid as a reaction product. Results and Discussion Methyl Sulfide. Ten sunlight irradiation experiments were carried out with mixtures of methyl sulfide (CH3SCH3,initial concentrations 0.2-2.5 ppm) and oxides of nitrogen (86-580 ppb) in particle-free purified air. Reactants and products concentration are given in Table 111. Typical concentration-time profiles are shown in Figures 1 and 2. Features common to all runs included high photochemical reactivity (e.g., rapid conversion of NO to NOz and formation of substantial amounts of ozone), significant consumption of NOz, formation of sulfur dioxide, nitric acid, formaldehyde, and methyl nitrate as gas-phase products, and formation of aerosols including inorganic sulfate (S042-)and methanesulfonic acid (CH3SOZ0H). Two apparent NO2maxima were observed in several runs, the second maximum including in fact a substantial contribution from nitric acid, to which the chemiluminescent NO, analyzer responds nearly quantitatively in the NOz mode of the instrument (26,27). As is shown in Figures 1 and 2 and discussed in more detail later, the products listed above did not account for dl of the reacted CH3SCH3 Environ. Sci. Technol., Vol. 18,No. 6, 1984
481
Table 111. S u m m a r y o f Methyl Sulfide-NO,
Sunlight I r r a d i a t i o n Experiments run
initial concn, ppb CHBSCH, NO NO2 irradiation time, h NO-NOZ, crossover, min from to maxima, ppb ozone SO2 NO2, 1 s t NOz, 2ndd products concn at end of runf formaldehyde, ppb SO2-, pg m-3g CH3S03-, wg m-3g
-
65”
66‘
67
68
69
590 580 0 7 47
2500 180 0 4 16
2500 110 0 4 16
720 200 50 3 15
700 220 0 3 25
240 180 160
205 524 107 127
165 475 75 70
500 310 200 150
460 285 165 125
52‘ 17.5 58
1570 27 33
1350 19 12
760 72 250
680 41 11
63 8 4
53 143
74c 230 90 5 2 54
75
76
77
113
135 110 20 5 100
125 140 5 5 115
310 360 20 6 97
680 240
133 54 90
177 56 75
480 165 230
350 195 84
20 3
10 21
175 139
290 54 94
0 6
129
“All runs in dry matrix air (dew point -20 “C) except runs 39,47, and 74, 4-m3 chambers. 30-40%. bRuns 39,47,66,68,76, and 77 in large Teflon chamber; other runs in smaller 4-m3 chambers. COvercast,run “slower” than others due to lower ultraviolet radiation intensity. dNitric acid accounts for a large fraction of the apparent NO2 in the later phase of the runs, e.g., 113 ppb = 75% at second NOz maximum in run 68; see text and ref 10. eStill increasing a t end of run. f T h e absence of acetaldehyde (52 ppb) and PAN (51 ppb), which may form from hydrocarbon impurities in the matrix air, was verified experimentally in most runs. BSamples collected on quartz filters in runs 68, 76, 77, and 113 and on Teflon filters in all other runs.
AN NOx
...
.....
NO;! --
AS
zoo NO
-
-
I
0
oL/ 0
Z
2
3
4
I
I
3
TIME,
4
I
5
6
f
6
hr
Flgure 1. Concentration-time profiles for methyl sulfkle-NO, irradiation, run 77. Top: NO, NO2, ozone, total NO, (as measured by chemiiuminescence), and AN = nitrogen not accounted for as NO,. Bottom: Total sulfur (TS), SO2, and methyl sulfide.
or reacted NO,. Typically, about 60% of the reacted CH3SCH3and 40% of the reacted NO, could be accounted for as gaseous and particulate products. Under the conditions of our study, reaction with the hydroxyl radical is the major removal process for methyl sulfide. Photolysis (28,29),reaction with ozone (28,29), and reaction with atomic oxygen (30) are essentially negligible. At ambient temperature, the OH-methyl sulfide reaction rate constant is (-0.5-1.0) X cm3 molecule-l s-l (28, 31, 32); i.e., methyl sulfide reacts with OH -2-3 times faster than most paraffins but 3-4 times slower than most aromatics and olefins (32). The observed photo462
S
TS/TSo
CH3SCH3
/
I
1
Environ. Sci. Technol., Vol. 18, No. 8 , 1984
DMs/DMso
TIME, HOURS
Flgure 2. Concentration-time profiles for methyl sulfide-NO, irradiation, run 113. Top: NO, NOz, ozone, total sulfur (TS), SO2, TS-SO2, and AS = sulfur not accounted for as SO2. Bottom: log plots of TSITS, and (CH3SCH3)/(CH3SCH3),concentration ratios.
chemical reactivity of the CH3SCH3-NO, mixture is higher than expected from OH reactivity considerations and is consistent with free radical production from formaldehyde, a major reaction product.
0
0 CH SCH3, this work CH~SCH,, reference IO 0 CZH5SC~Hs.run X 116 A CH CH:SH-CIZ SH-CIz-NOz
11
CH3SCH3 CH SCH3 /
3
CH3S0
Ho 2
.t
CH3
CH3 + CH3SOH ,CH3S03HO2 A
