Environ. Sci. Technol. 1996, 30, 1048-1052
Isomerization of β-Hydroxyalkoxy Radicals Formed from the OH Radical-Initiated Reactions of C4-C8 1-Alkenes
SCHEME 1
ERIC S. C. KWOK,† R O G E R A T K I N S O N , * ,‡ A N D J A N E T A R E Y * ,‡ Statewide Air Pollution Research Center, University of California, Riverside, California 92521
The dihydroxycarbonyl products of the isomerization reactions of the β-hydroxyalkoxy radicals formed after OH radical addition to the 1-alkenes 1-butene through 1-octene have been observed by direct air sampling atmospheric pressure ionization mass spectrometry (API-MS). The experimental data obtained provide direct evidence for β-hydroxyalkoxy radical isomerization via a six-member transition state. Our data show that the importance of β-hydroxyalkoxy radical isomerization in the OH radical-initiated reactions of the 1-alkenes increases with carbon number, being unimportant for 1-butene and dominating for 1-heptene and 1-octene.
Introduction The alkenes present in ambient air (1-3) originate from both anthropogenic and biogenic sources. In the troposphere, alkenes react with OH radicals, NO3 radicals, and O3 (4), with the daytime OH radical reaction often dominating as their removal process (5, 6). While the kinetics of the OH radical reactions are reasonably well understood for a large number of alkenes (4, 7), only for a few alkenes have product studies been conducted at room temperature and 1 atm of air (4, 8-17). The gas-phase reactions of the OH radical with alkenes proceed mainly by initial addition to the >CdC< bond(s) to form a β-hydroxyalkyl radical (4, 7), which then rapidly adds O2 to form the corresponding β-hydroxyalkyl peroxy radical (4). For example, for 1-pentene: OH + CH3CH2CH2CH CH2 . . CH3CH2CH2CHCH2OH and CH3CH2CH2CH(OH)CH2 O2
. CH3CH2CH2CH(OO)CH2OH
(1)
O2
. CH3CH2CH2CH(OH)CH2OO
* Authors to whom correspondence should be addressed; fax: (909) 787-5004. † Present address: Environmental Chemistry Division, Brookhaven National Laboratory, Upton, NY 11973. ‡ Also at the Department of Soil and Environmental Sciences, University of California, Riverside, CA 92521.
1048
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 3, 1996
In the presence of NO, the β-hydroxyalkyl peroxy radicals react with NO to form either the β-hydroxynitrate or the β-hydroxyalkoxy radical plus NO2 (4, 18, 19): . CH3CH2CH2CH(OH)CH2OO + NO M
CH3CH2CH2CH(OH)CH2ONO2 . CH3CH2CH2CH(OH)CH2O + NO2
(2a) (2b)
The formation of β-hydroxynitrates from reaction 2a has been observed from the OH radical-initiated reactions of propene (18) and cis-2-butene (19) in the presence of NO, and rate constant ratios k2a/k2b for the β-hydroxyalkyl peroxy radicals formed from propene and cis-2-butene of ∼0.0150.017 and 0.037 ( 0.009, respectively, were measured (18, 19). Reactions of the β-hydroxyalkoxy radicals with NO and NO2 are of minor importance in the troposphere (4), and the β-hydroxyalkoxy radicals react with O2, decompose, or isomerize by a 1,5-H shift through a six-member transition state (4). The reactions predicted (4) for the CH3CH2CH2CH(OH)CH2O• radical formed after internal OH radical addition to 1-pentene, leading to “first-generation” products, are shown in Scheme 1. Previous studies have shown at room temperature and 1 atm of air that the HOCH2CH2O• radical formed after OH radical addition to ethene both reacts with O2 and decomposes (9), while the β-hydroxyalkoxy radicals formed from propene, 1-butene, and trans-2-butene predominantly decompose (8, 10). However, the formation yields of the carbonyls RCHO + HCHO from the OH radical reactions with the 1-alkenes (RCHdCH2) 1-pentene through 1-octene decrease markedly with increasing carbon number of the alkene (17), suggesting that isomerization of the intermediate β-hydroxyalkoxy radicals occurs. To date, the products predicted to be formed from these β-hydroxyalkoxy radical isomerization reactions have not been observed. In this study, we have used a direct air sampling atmospheric pressure ionization mass spectrometer (APIMS) to investigate the occurrence of isomerization of β-hydroxyalkoxy radicals formed in the OH radical-initiated reactions of 1-butene through 1-octene in the presence of NO at 296 ( 2 K and atmospheric pressure.
0013-936X/96/0930-1048$12.00/0
1996 American Chemical Society
Experimental Section The experimental methods used were similar to those described previously (20, 21). Reactions were carried out at 296 ( 2 K and 740 Torr total pressure of purified air or O2-N2 mixtures at ∼5% relative humidity in a ∼6500-L all-Teflon chamber equipped with blacklamps for irradiation. Hydroxyl radicals were generated by the photolysis of methyl nitrite (CH3ONO) in air at wavelengths >300 nm (22), and NO was added to the reactant mixtures to suppress the formation of O3 and hence of NO3 radicals (22). The initial CH3ONO, NO, and 1-alkene concentrations were maintained equal in all experiments and were (2.4-4.8) × 1013 molecule cm-3 each. Irradiations of CH3ONO-NO1-alkene-air (or N2 + O2) mixtures were carried out at 20% of the maximum light intensity for 2-5 min, resulting in 23 ( 6% reaction of the initially present 1-alkene. The reaction mixtures were analyzed with a PE SCIEX API III MS/MS direct air sampling atmospheric pressure ionization tandem mass spectrometer interfaced to the collapsible Teflon chamber via a 25 mm diameter × 75 cm length Pyrex tube. All experiments were conducted in the positive ion mode and with a sample flow rate of ∼22 L min-1. Under positive API conditions, a corona discharge in the sample stream from the chamber containing water vapor at ∼5% relative humidity generates hydronium ionwater clusters, H3O+(H2O)n, which are responsible for the protonation of analytes (23):
H3O+(H2O)n + M f MH+(H2O)m + (n - m + 1)H2O (3) where M is the neutral analyte of interest. Ions are drawn by an electric potential from the ion source through the sampling orifice into the mass analyzing first quadrupole (Q1) or third quadrupole. Neutral molecules and particles are prevented from entering the orifice by a counterflow of ultra pure nitrogen (“curtain”) gas. The [M + H]+ ions mass analyzed are the result of the declustering action of the curtain gas, with the sensitivity of the API-MS to the analyte M being determined by the stability of MH+(H2O)m (23). It is not possible to eliminate fragmentation prior to Q1, but the system can be optimized to favor protonated molecular ions [M + H]+ and/or protonated dimer ions [M2 + H]+. The MS (scanning) mode was used to provide mass spectra of the reactant and reacted mixtures. Daughter ion and parent ion spectra of a given ion peak observed in the MS scanning mode were obtained in the MS/MS mode [with collision activated dissociation (CAD)] to provide structural information and information concerning the origin of an ion peak (i.e., whether it is a molecular ion peak or a fragment ion peak) (21). For additional product confirmation in certain experiments, the API-MS instrument was tuned to optimize dimer formation (20, 21). Because the API-MS instrument was relatively insensitive to the reactant alkenes in these experiments, the concentrations of the alkenes were measured during the experiments by gas chromatography with flame ionization detection (GCFID) (17). The chemicals used and their stated purities were as follows: 1-pentene (99%), 1-hexene (99%), 1-heptene (99%), 1-octene (98%), propanal (99+%), butanal [n-butyraldehyde] (99%), pentanal (99%), hexanal (99%), and heptanal (95%), Aldrich Chemical Company; 1-butene (>99%) and NO (g99.0%), Matheson Gas Products. Methyl nitrite was
FIGURE 1. API-MS spectra of irradiated CH3ONO-NO-1-alkeneair mixtures. The [M + H]+ ion peaks of the Cn-1 aldehyde products formed from the Cn 1-alkenes are noted and shown by an arrow, and the [M - H]+ and [M + H - H2O]+ ion peaks of the Cndihydroxycarbonyls are noted and shown by the connected vertical lines (the [M + H]+ ion peaks of the dihydroxycarbonyls are too weak to be readily observed in this figure).
prepared as described previously (22) and was stored at 77 K under vacuum.
Results and Discussion Figure 1 shows the API mass spectra of irradiated CH3ONO-NO-1-alkene-air mixtures for 1-butene through 1-octene. Spectra taken prior to irradiation showed no significant ion peaks from the alkene reactants. Previous studies using GC-FID and in situ Fourier transform infrared (FTIR) absorption spectroscopy identified Cn-1 aldehydes and formaldehyde as products of the OH radical reactions with these Cn 1-alkenes (10, 17). Carbonyl compounds give strong protonated molecular ions ([M + H]+) under our instrumental conditions (20, 21) and, therefore, the ion peaks observed at 59, 73, 87, 101, and 115 u (unified atomic mass unit) from the 1-butene, 1-pentene, 1-hexene, 1-heptene, and 1-octene reactions, respectively, were attributed to propanal, butanal, pentanal, hexanal, and heptanal, respectively. API-MS/MS CAD spectra of the 59, 73, 87, 101, and 115 u ion peaks matched CAD spectra of the [M + H]+ ion peaks of authentic standards of the corresponding aldehydes, confirming these assignments. Furthermore, the assumption that the ion peaks at the masses listed above are protonated molecular ion peaks and not fragment ions was confirmed by operating the API-MS instrument under
VOL. 30, NO. 3, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1049
FIGURE 2. API-MS/MS CAD spectrum of the 119 u [M + H]+ ion of the C5-dihydroxycarbonyl formed from the OH radical-initiated reaction of 1-pentene.
conditions favoring dimer formation and observing ion peaks corresponding to the aldehyde dimer in the OH radical reactions with 1-pentene through 1-octene. The absence of an [M2 + H]+ ion of propanal from the 1-butene reaction is consistent with the lack of dimer formation from authentic samples of propanal in the chamber analyzed under these instrumental conditions. Although standards of β-hydroxycarbonyls and of dihydroxycarbonyls are generally not available, we have observed strong [M + H - H2O]+ fragment ions from all the compounds containing an alcohol functional group that we have examined by API-MS. These include authentic standards of several C4-C8 alcohols, the δ-hydroxycarbonyl 5-hydroxy-2-pentanone, and 4-hydroxy-4-methyl-2-pentanone, as well as the hydroxycarbonyls formed from the OH radical reactions with the C4-C8 n-alkanes (20, 21). In addition to the strong [M + H - H2O]+ ion peaks in our API-MS spectra of alcohols and hydroxycarbonyls, weak [M + H]+ and [M - H]+ ion peaks were also observed (20, 21). In addition to the ion peaks corresponding to the formation of the aldehyde RCHO from the alkene RCHdCH2, API-MS spectra of irradiated CH3ONO-NO1-alkene-air mixtures showed the presence of prominent higher mass ion peaks at 87, 101, 115, 129, and 143 u from 1-butene, 1-pentene, 1-hexene, 1-heptene, and 1-octene, respectively (Figure 1). The molecular weights of the expected dihydroxycarbonyls formed after isomerization of the β-hydroxyalkoxy radicals formed from 1-butene through 1-octene are 104 (C4), 118 (C5), 132 (C6), 146 (C7), and 160 u (C8), suggesting that the observed prominant peaks are the [M + H - H2O]+ ions of these dihydroxycarbonyls. API-MS/MS daughter ion spectra of the weak ion peaks observed at 119, 133, 147, and 161 u from the 1-pentene through 1-octene reactions (Figure 1) as well as API-MS/MS parent ion spectra of the [M + H - H2O]+ ions from these reactions were consistent with this interpretation. The ion peaks at 117, 131, 145, and 159 u for the 1-pentene through 1-octene reactions, respectively (see Figure 1), are attributed to the [M - H]+ ions of the dihydroxycarbonyls. As shown in Figure 2 for the 1-pentene reaction, the API-MS/MS CAD spectra of the 119, 133, 147, and 161 u [M + H]+ ion peaks from the 1-pentene through 1-octene reactions can be rationalized by the expected fragmentation pattern of the dihydroxycarbonyls, with large peaks corresponding to [M + H - H2O]+ and [M + H-2H2O]+ ions being consistent with fragmentation of the diol under these
1050
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 3, 1996
FIGURE 3. API-MS spectra of irradiated CH3ONO-NO-1-octeneN2-O2 mixtures at 4% (upper), 21% (middle), and 80% (lower) O2 levels. The [M + H]+ ion peak of heptanal is at 115, and the [M + H]+, [M - H]+, and [M + H - H2O]+ ion peaks of the C8dihydroxycarbonyls are at 161, 159, and 143 u, respectively.
conditions. Ion peaks corresponding to [M + H-CO]+ and [M + H - H2O-CO]+ fragments were also observed in these CAD spectra of the dihydroxycarbonyls (Figure 2). Figure 3 shows the API mass spectra obtained from the reaction of the OH radical with 1-octene at low (∼4%), normal (21%), and high (∼80%) O2 levels, with 19 ( 4% 1-octene reacted. The ion signals are interpreted as follows: the signal at 115 u ([M + H]+ for heptanal) resulting from the decomposition pathway; the signals at 161, 159, and 143 u (C8-dihydroxycarbonyl) resulting from the isomerization pathway; and the signal at 127 u could be the [M + H - H2O]+ of the C8-hydroxycarbonyl (expected to be the dominant peak of the hydroxycarbonyl API mass spectrum) resulting from the O2 reaction. Our data from two independent sets of experiments (one set being shown in Figure 3) showed that the integrated signal intensities of the 127 u ion peak relative to the integrated signal intensities of the 159, 143, or 115 u ion peaks varied by factors of e1.8. Furthermore, the integrated signal intensities of the 115 u ion peak relative to those of the 143 u ion peak also varied by a factor of
