Oxidation Of Sulfur Compounds Other Than So

As discussed in Chapter 2 and in more detail in Chapter If, a variety of organic sulfur compounds in addition to inorganics such as H2S and COS are emitted by biological sources. In the troposphere, they may ultimately be oxidized to S02 and H2S04. However, the chemistry of these compounds tends to be complex, and a variety of partially oxidized sulfur compounds is formed first.

Table 8.17 summarizes the rate constants and estimated tropospheric lifetimes of some of these sulfur compounds with respect to reaction with OH or N03. The assumed concentrations of these oxidants chosen for the calculations are those characteristic of more remote regions, which are major sources of reduced sulfur compounds such as dimethyl sulfide (DMS). It is seen that OH is expected to be the most important sink for these compounds and that N03 may also be important, for example, for DMS oxidation (see also Chapter 6.J).

As discussed in Chapters 6 and 7, NaCl and other chloride salts in airborne sea salt particles are now believed to react to generate photochemically active compounds such as Cl2 and ClNOz that photolyze to form chlorine atoms. Peak concentrations of atomic chlorine as high as 104-105 cm"3 in the early morning hours have been predicted (Keene et al., 1993, 1996; Pszenny et al., 1993; Wingenter et al., 1996; Singh et al., 1996; Tuckermann et al., 1997; Spicer et al., 1998). Chlorine atoms also react rapidly with reduced sulfur compounds and hence are potential oxidants for these organics in the early morning hours in coastal regions.

Finally, there has been a great deal of interest in the halogen oxides IO, BrO, and CIO as potential oxidants for organic sulfur compounds such as dimethyl sulfide. We therefore also discuss the current status of these reactions.

Dimethyl sulfide (DMS) is of particular interest since it is the major reduced sulfur compound produced in oceanic areas. In addition, as discussed in Chapter 14, it is believed to play a significant role in global climate issues; thus, its oxidation to S02 is expected to lead to the subsequent formation of sulfuric acid and sulfate particles by the chemistry that is the focus of this chapter. In addition, as we shall see, another oxidation product of DMS is methanesulfonic acid (CH3S03H), which is highly soluble and has a sufficiently low vapor pressure that, like sulfuric acid, it exists primarily in the condensed phase (e.g., Ayers et al., 1980; Clegg and Brimblecombe, 1985; Kreidenweis and Seinfeld, 1988a, 1988b). As a result, oxidation of DMS can generate condensation nuclei on which water condenses to form clouds. These clouds can scatter radiation back to space, lowering the surface temperature and hence the production of DMS, forming a feedback loop (Charlson et al., 1987). Because of this potential feedback, there have been many recent studies examining the relationship between DMS, its oxidation products, and condensation nuclei (e.g., see Kreidenweis et al., 1991; Berresheim et al., 1993; Pandis et al., 1994; Ferek et al., 1995; Clarke et al., 1996; and Bandy et al., 1992, 1996). However, it is clear that understanding such feedbacks requires understanding the chemistry of DMS and, in particular, its role as a source of low-volatility products that can act as condensation nuclei.

As a result, we focus here on what is known about the tropospheric chemistry of DMS. As we shall see, the chemistry of even this relatively simple compound is complex, and much remains to be learned about its reaction mechanisms. For larger reduced sulfur com-

TABLE 8.17 Rate Constants and Lifetimes at Room Temperature for the Reactions of OH and NO 5 with Some Reduced Sulfur Compounds Emitted Biogenically

Compound k (cm3 molecule 's ')" tOHh k 'cm3 molecule 's ')" tno,'

CH3SCH3 6.5 X 10"12 2 days 1.1 X10"12 10 h

CH3SSCH3 2.3x10" 111 1.2 h 0.7x10"12 16 h

CH3SH 3.3 x 10"11 8 h 0.92X10"12 12 h

H2S 4.8X10"12 2.4 days <lxl0"15 >lyr

CS2 4.7X10"12 2.5 days <1x10"15 >lyr

COS 2.0X10"15 16 yr <lxl0"16 >13yr

" Rate constants at 1 atm in air at 298 K; from Atkinson et al. (1997a).

h Ton = I/AfOH], where [OH] is taken to be 1 X 10'' radicals cm"3.

' tno, = 1 A[N03], where [N03] is taken to be 1 ppt = 2.5 X 107 radicals cm"3.

pounds, including disulfides, even less is known. However, there are a number of organic sulfur intermediates formed in the DMS reactions that are common to the oxidations of other sulfur compounds as well. Hence some "educated guesses" as to the chemistry of other sulfur compounds can be made based on what is known about DMS chemistry as discussed in the following sections. For a review of dimethyl sulfide and other sulfur compounds in the atmosphere, see Berresheim et al. (f 995).

1. Reactions of Dimethyl Sulfide (CH3SCH3) a. Reaction with OH

The reaction of OH with DMS is believed to proceed by two channels, abstraction of a hydrogen from a methyl group and addition of OH to the sulfur atom:

OH + CH3SCH3 -> CH3SCH2 + H20, (45a) ki5.d = 4.8 X 10"12 cm3 molecule"1 s"1 at 298 K

(Atkinson et al., 1997a) OH + CH3SCH3 + M <-> CH3S(OH)CH3 + M,

k45h = 1.7 X 10"12 cm3 molecule"1 s"1 at 298 K, 1 atm (Atkinson etal., 1997a).

Thus, at 1 atm in air and 298 K, abstraction predominates. The addition channel (45b) would be expected to have a pressure dependence and a negative temperature dependence (see Chapter 5.A.2). Thus is consistent with the observation that the effective overall bimolecular rate constant in 1 atm of air decreases as the temperature increases from 250 to 310 K and that the fraction of the reaction that proceeds via (45a) increases from 0.24 to 0.87 over the same temperature range (e.g., Hynes et al., 1986).

Experimental studies of the reaction of OH with fully deuterated DMS (Hynes et al., 1995; Barone et al., 1996) give a bond strength for the adduct at 258 K of ~ 10-13 kcal mol"1. In air, reaction (46) of the adduct with 02 can occur in competition with its decomposition back to reactants [reaction ( — 45b)]:

The rate constant (k46) is independent of pressure from 100 to 700 Torr and temperature from 2f7 to 300 K and is reported to be (8-10) X 10"13 cm3 molecule~ 1 s~' (Hynes et al., 1995; Barone et al., 1996). At 1 atm in air, the first-order removal rate of the adduct by reaction with 02 is thus 4 X 106 s"1, similar to the unimolecular rate of decomposition of the adduct of 3.5 X 106 s" 1 at 261 K (Hynes et al., 1986).

At least part of reaction (46) forms dimethyl sulfoxide (DMSO):

Figure 8.24a, for example, shows the FTIR spectrum before the photolysis of mixtures of DMS in air with H202 as the OH source and the residual spectrum after 5 min of photolysis (Barnes et al., 1996). The reactants, as well as the product S02 have been subtracted out in Fig. 8.24b. Dimethyl sulfoxide (DMSO) as well as dimethyl sulfone, CH3S02CH3 (DMS02), and small amounts of COS are observed as products. DMSO is so reactive that it is rapidly converted into DMS02 in this system and hence both are observed in Fig. 8.24b. However, Barnes and co-workers calculate that the DMSO yield corrected for secondary oxidation is about the same as the fraction of the OH-DMS reaction that proceeds by addition under these conditions, i.e., that the major fate of the adduct is reaction (47). Turnipseed et al. (1996) measured the yield of H02 from reaction (47) to be 0.50 ± 0.15 at both 234 and 258 K, suggesting that there are other reaction paths than (47) as well. The mechanism of formation of COS is not clear but may involve the oxidation of thioformaldehyde (H2C=S). The implications for the global budget of COS are discussed by Barnes et al. (1994b, 1996).

The radical formed by abstraction, (45a), is, in essence, an alkyl radical and is therefore expected to add 02:

Wavenumber (cm-1)

FIGURE 8.24 Infrared spectra of (a) DMS and H202 at 1 atm air before photolysis and (b) the product spectrum after 5-min irradiation and subtraction of peaks due to unreacted DMS and H202 and product S02 (adapted from Barnes et al., 1996).

Wavenumber (cm-1)

FIGURE 8.24 Infrared spectra of (a) DMS and H202 at 1 atm air before photolysis and (b) the product spectrum after 5-min irradiation and subtraction of peaks due to unreacted DMS and H202 and product S02 (adapted from Barnes et al., 1996).

The rate constant for reaction (48) has been reported as (5.7 ± 0.4) X 10"12 cm3 molecule 1 s"1 at 1 atm (Wallington et al., 1993) and somewhat smaller at lower pressures (Butkovskaya and Le Bras, 1994), consistent with similar reactions of alkyl radicals (see Chapter 6.D.1).

In the presence of NO, the alkylperoxy radical oxidizes NO to NOz, forming an alkoxy radical, which can decompose to formaldehyde and the CH3S radical (Wallington et al., 1993; Butkovskaya and Le Bras, 1994):

CH3SCH200 + NO -> CH3SCH20 + NOz, (49) CH3SCH20 -> CH3S + HCHO. (50)

The rate constant for reaction (49) has been estimated to be ~1 X 10"" cm3 molecule"1 s"1 at 298 K, similar to that of other R02 + NO reactions (Turnip-seed et al., 1996).

An alternate potential fate for the CH3SCH20 radical is the well-known abstraction reaction of alkoxy radicals with 02, forming methyl thioformate (Crutzen, 1983; Butkovskaya and Le Bras, 1994):

Infrared absorption bands attributable to methyl thioformate have been observed in the oxidation of DMS by OH in the absence of NOx but not when NOx was present (Barnes et al., 1996; Patroescu et al., 1999). Reaction (51) appears to be quite slow (k < 1 X 10"15 cm3 molecule"1 s"1), so that the dominant fate of CH3SCH20 is decomposition to HCHO + CH3S; Turnipseed et al. (1996) measured the production of CH3S in the reaction of OH with DMS and also suggest, based on its relatively high yield, that the thermal decomposition (50) predominates over reaction (51) with 02 to form methyl thioformate. The methyl thioformate observed in laboratory systems in the absence of NO is thus likely due to cross reactions of CH3SCH200 with itself or other R02 (Barnes et al., 1994a), and the abstraction channel in the OH + DMS

reaction leads primarily to the formation of CH3S. Methyl thioformate itself reacts rapidly with OH (k = 1.1 X 10"" cm3 molecule"1 s"1 at room temperature), with evidence from FTIR studies for both abstraction of the aldehydic hydrogen and addition to the sulfur; photolysis of methyl thioformate also occurs but is relatively slow, with a lifetime of > 5.4 days (Patroescu et al., 1996).

Thus, in low-NOx environments, the CH3SCH200 radical will react with H02 or other R02 radicals:

The fate of the CH3S radical in the atmosphere is not clear but may include reaction with 02, 03, or NOz. Table 8.18 summarizes the rate constants for these reactions at 298 K and the corresponding lifetimes under typical tropospheric conditions. Although only an upper limit can be placed on the rate constant for the 02 reaction, it may still be the predominant reaction of CH3S because of the large oxygen concentration in the atmosphere. However, this process forms a weakly bound adduct with a CH3S-00 bond energy of about 11 kcal mol"1, and decomposition back to reactants also occurs:

Evidence for formation of this adduct has been obtained in laboratory studies between 216 and 258 K, where CH3S is observed to come to equilibrium in the presence of 02 (Turnipseed et al., 1992). A contribution from the back reaction is difficult to avoid in experimental systems, making measurements of the true forward rate constant somewhat uncertain. Extrapolation of the measured kinetics to 298 K suggests that approximately 30-75% of the CH3S would be in the form of the adduct at typical tropospheric temperatures of 298-275 K and 1 atm pressure (Turnipseed et al., 1992).

TABLE 8.18 Removal Rates of CH3 S in the Troposphere at 298 K "

A (298 K) [Reactantl

Reaction of CH,S (cm3 molecule-1 s"1) (molecules cm 3) t(s)

CH3S + 02 -> products <6.0x10"18 5.2 X 1018 >0.03

CH3S + 03 -» products 5.4x10"12 9.8 X 1011 (40 ppb) 0.19

CH3S + N02 -> CH3SO + NO 5.8 x 10" " 2.5 x 1010 (1 ppb) 0.69

" Rate constants from Tyndall and Ravishankara (1991), Dominé et al. (1992), and Turnipseed et al. (1993).

The fate of the CH3SOO adduct is not known but by analogy to other peroxy radical reactions is expected to include reactions with NO and N02 (Turnipseed et al., 1993):

CH3SOO + NO -> CH3SO + N02, (54) fc54(227-256 K) = 1.1 X 10"" cm3 molecule"1 s"1 CH^SOO + NO2

(227-246 K) = 2.2 X 10"" cm3 molecule"1 s"1.

Another alternative is isomerization to CH3S02 (Turnipseed and Ravishankara, 1993).

The data in Table 8.18 show that reactions of CH3S with N02 and 03 may also be important. This also appears to be the case with larger thio radicals such as C2H5S (Black et al., 1988). The reaction with N02 produces primarily CH3SO + NO (Barnes et al., 1987; Hatakeyama, 1989; Tyndall and Ravishankara, 1989; Dominé et al., 1990); a minor addition channel produces CH3SN02, which has been observed in laboratory systems using FTfR (Barnes et al., 1987).

CH3SO has been observed as a product of the reaction of CH3S with 03 (Dominé et al., 1992), suggesting that one channel is an oxygen atom transfer:

The yield of CH3SO at low (Torr) pressures is only 15%. However, the yield has not been determined at 1 atm. Since (56a) is highly exothermic (~ 59 kcal mol"1 ), the CH3SO may contain sufficient energy to decompose at low pressures to CH3 + SO; quenching of an excited CH3SO at 1 atm may lead to much greater yields of this radical. Other potential channels that are exothermic are the following:

Although Dominé et al. (1992) place an upper limit of 4% on the contribution of (56e), Barnes et al. (1996) point out that even a small production of CH2 S followed by its oxidation could be responsible for the small yields of COS they observe by FTIR.

The CH3SO radical will be further oxidized under tropospheric conditions:

k57(298 K) = 6.0 X 10"13 cm3 molecule"1 s" 1 (Dominé et al., 1992) CH3SO + NOz -» CH3S02 (or CH3 + S02) + NO,

k5H(298 K) = f .2 X 10"11 cm3 molecule"1 s"1 (Mellouki etal., 1988; Hatakeyama, 1989; Dominé et al., 1990).

There is also evidence for an addition reaction of CH3SO with 02 (e.g., Hatakeyama et al., f989; Barone et al., f 995):

For example, methylsulfinyl peroxynitrate, CH3S(0)00N02, has been observed by FTIR (e.g., see Barnes et al., 1987; Hatakeyama, 1989; and Jensen et al., f992), presumably from reaction (59) followed by reaction of the peroxy radical formed with N02.

By analogy to other reactions, there are several other fates of the CH3S(0)00 radical formed in (59) that might be expected to be important. These include decomposition or reaction with NO or other peroxy radicals (Barnes et al., 1987):

The S03 formed in (60) will react with HzO to form H2S04 as described in detail earlier in this chapter, whereas CH3S(0)0 formed in (61) may decompose or react further.

The CH3S02 radical has a number of potential fates, including decomposition to CH3 + S02 and reactions with N02, 02, and 03:

CH3S02 + M -> CH3 + S02 + M, (62) /c62(298 K) = 510 s"1 at 1 Torr pressure (Ray et al., 1996) CH3S02 + N02 -» CH3S03 + NO, (63) ¿h3(298 K) = 2.2 X fO"12 cm3 molecule"1 s"1 (Ray et al., 1996).

Only upper limits for rate constants for the reaction of CH3S02 with 02 and 03 of < 6 X 10"18 and <8 X 10"13 cm3 molecule"1 s"1 have been measured (Turnipseed and Ravishankara, 1993). If the rate constant for decomposition measured at f Torr is the same at 1 atm, the decomposition is expected to predominate at 298 K. As discussed by Ray et al. (1996), as the temperature decreases, the rate of the thermal decomposition will also decrease and some of the other reactions may become competitive.

However, Patroescu et al. (1999) observed in FTIR experiments in 1 atm air that the yield of methane-sulfonic acid increased as the NOx concentration increased. They attributed this observation to the addition of 02 to CH3S02, with secondary reactions involving NO leading to the formation of methanesul-fonic acid, CH3S020H. Thus, at 1 atm in air, the reaction of CH3S02 appeared to be much faster than its thermal decomposition. Reaction of the CH3S02(00) adduct with N02 was postulated to give CH3S0200N02, methanesulfonyl peroxynitrate, which was observed by FTIR (Patroescu et al., 1999).

The reaction of OH with dimethyl sulfoxide also appears to occur primarily by addition. Urbanski et al. (1998) have reported yields of CH3 in this reaction of unity and suggest that the reaction is

methanesulfinic acid

The rate constant at 298 K, based on the production of CH3, was measured to be kM = (8.7 ± 1.6) X 10"" cm3 molecule" 1 s" 1 (Urbanski et al., 1998). The formation of methanesulfinic acid in the OH-dimethyl sulfide system has been reported by S0rensen et al. (1996), who proposed that it resulted from secondary reactions of the OH-DMS adduct rather than reaction (64). Further addition of OH to methanesulfinic acid followed by reaction with 02 may then form methanesul-fonic acid, CH3S(0)(0)0H, which in the atmosphere is rapidly scavenged into particles (e.g., Clegg and Brim-blecombe, 1985). The adduct may also react with 02 and, by analogy to the OH-CH3SCH3 adduct, is expected to give H02 + CH3S(0)(0)CH3, dimethyl sulfone.

The oxidation of dimethyl sulfide (DMS) to dimethyl sulfoxide (DMSO) and the subsequent oxidation of the latter to methanesulfonic acid (MSA) have been observed in field studies. For example, one study in Antarctica which focused on the chemistry of dimethyl sulfide (Berresheim and Eisele, 1998) measured not only DMS but also a variety of its oxidation products, including DMSO, MSA, and dimethyl sulfone (Berresheim et al., 1998). The measured concentrations of DMSO were in agreement with model results if 80-f00% of the OH + DMS reaction gave DMSO; as discussed earlier, the addition channel that leads to DMSO becomes relatively more important at the lower temperatures found in Antarctica. Furthermore, the

MSA concentrations were generally consistent with its formation from the OH-DMSO reaction if 80% of the reaction generated methanesulfinic acid and 50% of this compound was oxidized to MSA (Davis et al., f998).

Figure 8.25 summarizes the current understanding of the mechanism of DMS oxidation in the troposphere. For reviews, see Plane (f989), Turnipseed and Ravishankara (f 993), Barnes et al. (1993, 1996), Barone et al. (1995), and Berresheim et al. (1995).

b. Reaction with the Nitrate Radical (N03)

The nitrate radical is also known to react rapidly with DMS:

CH3SCH3 + N03 ^ CH3SCH2 + HN03, (65) kh5 = 1.1 X 10"12 cm3 molecule"1 s"1 (Atkinson et al., 1997a).

In contrast to the OH reaction, which occurs during the day due to the photolytic sources of OH, reaction (65) is a nighttime reaction due to the rapid photolysis of N03 at dawn. It is interesting that while the overall reaction appears to correspond to a hydrogen atom abstraction rather than addition to the sulfur atom, the mechanism is believed to involve the initial formation of an adduct, followed by its decomposition to HN03 and the alkyl radical shown in reaction (65) (Jensen et al., 1992; Daykin and Wine, f990; Butkovskaya and Le Bras, 1994). N02 is not formed in the N03-DMS reaction (Dlugokencky and Howard, 1988), ruling out an oxygen atom transfer either directly or via decomposition of the adduct.

c. Oxidation by Chlorine Atoms

Atomic chlorine reacts rapidly with DMS, with an overall rate constant of (3.3 + 0.5) X f0"'° cm3 molecule"1 s"1 at 298 K and 700 Torr total pressure (Stickel et al., 1992). As is the case for the OH reaction, the chlorine atom reaction proceeds by two reaction channels, one an abstraction and the other addition to the sulfur atom:

A potential reaction path producing CH3S + CH3C1 can be ruled out based on the observation of very small yields of CH3C1, only 0.13% at 1 atm pressure (Langer et al., 1996), and Zhao et al. (1996) show a path producing CH3 is also not important. Stickel et al. (1992) suggest that at 298 K and 1 atm pressure, the two paths, (66a) and (66b), are about equally important. The fate of the Cl-DMS adduct is not known.

d. Oxidation by Halogen Oxides: 10, BrO, and CIO

Halogen oxides are also potential reactants with DMS in the marine boundary layer (e.g., see Barnes et al., 1989). As discussed in Chapter 12, CH3I as well as CH3C1 and CH3Br have natural oceanic sources. While CH3CI and CH3Br do not absorb light significantly in the actinic region, CH3I and other alkyl halides do (see Chapter 4.V), photolyzing to form iodine atoms:

The iodine atoms react with 03, forming the IO radical, which can potentially oxidize DMS to DMSO, regenerating iodine atoms in a chain process:

The current recommendation for &ftX at 298 K is 1.2 X fO"14 cm3 molecule" 1 s" 1 (Atkinson et al., 1997a). The concentration of IO radicals is not well known, as it has only just been directly detected and measured in the troposphere at concentrations up to 6 ppt at a coastal site (Alicke et al., 1999). Using 108 cm"3, one can calculate a lifetime for DMS with respect to reaction (68) of about 10 days, too slow to compete with OH, N03, and CI atoms.

However, with the recent recognition of the potential importance of atomic chlorine and bromine under certain conditions in the Arctic at polar sunrise (e.g., see Barrie et al., 1988; and Niki and Becker, 1993), the potential for BrO and CIO chemistry has been reconsidered. As described in Chapter 6.J.4, at polar sunrise there is a rapid loss of ground-level 03 that appears to be associated with reaction with atomic bromine and at the same time, there is evidence that chlorine atoms are playing a major role in the organic removal (Jobson et al., 1994). This is consistent with reactions of sea salt particles generating atomic bromine and chlorine, although the exact nature of the reactions and halogen atom precursors remains unknown.

Was this article helpful?

0 0

Post a comment