REACTIONS OF AROMATIC COMPOUNDS IN THE ATMOSPHERE Dale G. Hendry Physical Organic Chemistry Group SRI International Menlo Park, California 94025 This paper is a review of the tropospheric chemistry of aromatic compounds. The reactivity of aromatic compounds is discussed and rate constants for their reactions. with OH are tabulated. The reaction mechanisms are discussed in detail. Key words: Aromatics; free radicals; mechanism; reactions; tropospheric chemistry. ca. 0.038 ppmC ethane, ca. 0.043 PpmC propane and ca. 0.022 ppmC benzene. substituted benzene compose roughly 24 carbon percent of the total hydrocarbon in Los Angeles and about 37 carbon percent in Manhattan. The aromatic compounds in the atmosphere come from gasoline [3], in which they are used to enhance the octane rating. Gasoline itself is composed of 30 to 40 percent aromatic hydrocarbons and apparoximately 6 to 8 percent toluene. Hydrocarbons emitted in automobile exhaust are composed of 6 to 8 percent toluene. We have known for some time from smog chamber reactivity studies that the alkyl-substituted benzenes are reactive in promoting oxidation of NO to NO2 and formation of ozone [4]. However, only in the last few years has an effort been made to understand specifically how these compounds react. This effort has been very productive, largely because it builds on an existing background of moderately well understood smog chemistry of the alkanes and alkenes. The total conversion of the aromatics to H2O, CO, and CO2 is a complex process, of which we understand only the initial steps. 2. Initial Reactions of Alkylbenzenes Table 2 summarizes the possible reactions of toluene, a representative aromatic hydrocarbon, with the oxidizing species known to be present in the atmosphere. Best values of rate constants and approximate concentrations are included for estimating the rate of loss of toluene by the various processes. The data in table 2 show clearly that the only important reaction of toluene in the atmosphere is with OH. The contribution of the reactions with 0 atom and 03 are about 10" and 103 of that of OH reaction. The reactions of RO2° proceed extremely slowly and can account for only 10 of the total consumption of toluene. Rate constants for the reaction of OH with various alkyl benzenes are summarized in table 3. We are fortunate to have several techniques for Figures in brackets indicate literature references measuring the rate constants for reaction of OH at the end of this paper. with aromatic hydrocarbons. The agreement between In our laboratory we have been investigating the products of reaction of aromatic hydrocarbons and OH in a discharge flow system [12,13]. The products were collected in cold traps and on soli adsorbents. The product distributions were determined as a function of hydrocarbon, NO2, and O2 pressures. Table 4 summarizes some of the dat obtained as a function of NO2 pressures. The fraction of products resulting from reaction (1) is a measure of k1/(kı + K2) and remains constant over the range of conditions. For toluene we obtain 0.15 0.02, which agrees very well with the best value reported by Perry et al. [7]. The k1/(k + k2) values for various aromatic hydrocarbons obtained by these two methods are summarized in table 5. Benzene 1.20 Toluene 6.40 5.78 Table 4. Product distribution for the reaction of toluene plus Oна. We find m-nitrotoluene to be a major product however, the concentration varies with the 0/1 ratio. Thus the intermediate formed in reacti (2) appears to react by two parallel pathways. mixtures, and we believe that many nitro products observed in smog chambers may reflect heterogeneous reactions either during the actual chamber reaction or during trapping out of the products. Since phenolic compounds are especially susceptible to heterogeneous nitration, the origin of nitrophenols must be interpreted with extreme caution. 4. Reactions of Initial Products Benzaldehyde Reactions. Two processes appear to be important for the reaction of the benzaldehyde formed from toluene in the atmosphere: the reaction with OH and photolysis. Niki et al. [16] recently reported the rate constant for the reaction as k = 1.3 x 10-11 cm3 molec1s1, which is identical within the experimental uncertainties to rate constant for other aldehyde-OH reactions. Addition of OH to the ring as observed for toluene (reaction (2)) is expected to occur no faster than addition to benzene, where koH = 1.2 x 10-12 cm3 molec 1s 1. Thus the attack of OH is expected to be largely at the aldehydic position. Using our discharge flow systems, we have found that the reaction produces phenol as the only gas phase product [17]. Thus the initial reactions of the benzaldehyde with OH is In addition to the meta-nitrotoluene orthopara-isomers have also been reported in smog rber experiments [14,15]. These isomers Id potentially be formed from the meta-OH-adduct toluene in sequences similar to reaction (4). ever, since very little m-cresol is formed, route does not seem reasonable. In many es, we have observed NO to be an effective rating agent upon condensing our reaction followed by the reactions = = 1 k13-19 (21 The following estimates are applicable: k13 1.7 x 10 12 cm3 molec 3 1 s1, k20 = 5.0 x 10 12 cm molec1 s1, k19 - 1 6.0 x 10-13 cm3 molec-1 s and [02] 0.01 M in air at 1 atm. If k-19 = 107 s which is consistent with our estimate of DH°(C-02) = 10 kcal/mol, the two processes will compete equally at NO2/NO = 1. At very high ratios of NO2/NO, however, the formation of nitr phenol will predominate. If the DH°(C-02) is mu weaker than 10 kcal/mol (k_19 >> 107 s1), react (13) will predominate under most atmospheric conditions, but if it is much stronger than 10 kcal/mol (k_19 >> 107 s1), reaction (13) wil be unimportant and reactions (19) and (20) will predominate. The relative importance of reactio (13) and (19)-(20) is very critical because reaction (13) is a termination reaction whereas (19)-(20) will lead to ring degradation and further oxidation of NO by reactions of the following type. PhCHO PhH + CO Ph: + Ho Phco + H. (15) (16) (17) Thus, Although reaction (15) is energetically favorable. at all wavelengths of the visible spectrum, its measured quantum yield is significant only at wavelengths less than 300 nm [19]. Reaction (16) is energetically possible only below 300 nm. only reaction (17), which has an energy cut-off at 330 nm, appears to be important in the solar spectrum. However, the possibility of the generation of a triplet excited state that reacts with oxygen above 330 nm cannot be ruled out. This reaction sequence is speculative, although each step can be justified in most cases by analysis of competing reactions. It does sugges that a-dicarbonyl compounds should be important secondary products. These compounds absorb lig very strongly in the solar spectrum and can be significant source of radicals [20,21]. Rat Cresol Reactions. The reaction of OH plus o-cresol was studied by Perry et al. [22] over temperature range 300 to 435 K (reactions for p and m-cresols are expected to be similar). constants were reported for two processes: (1) nonreversible reaction believed to be hydrogen abstraction (k = 2.6 x 10 12 cm3 molec s1) a (2) a reversible reaction believed to be additi to the ring (k = 3.1 x 1011 cm3 molec1 s ̄1). postulate the following reaction pathways. |