Environ. Sci. Techno/. 1995, 29, 1674-1680
Prducts of the b s - f h s e
ROGER ATKINSON,*,' ERNEST0 C. TUAZON, AND S A R A M . ASCHMANN Statewide Air Pollution Research Center, University of California, Riverside, California 92521
The products of the gas-phase reactions of the OH radical with 1-pentene, 1-hexene, 1-heptene, 1-octene, 2,3-dimethyl-l-butene, and l-methylcyclohexene have been investigated in the presence of NO a t room temperature and 740 Torr total pressure of air. Products were identified and quantified by gas chromatography and in situ Fourier transform infrared absorption spectroscopy. The carbonyl products identified and their yields were as follows: from 1-pentene, butanal (0.73 f0.09) and HCHO (0.88 f0.1 1); from 1-hexene, pentanal (0.46 f0.07) and HCHO (0.57 Ifr: 0.08); from 1-heptene, hexanal (0.30 f 0.04) and HCHO (0.49 f 0.06); from 1-octene, heptanal (0.21 f 0.03) and HCHO (0.39 f 0.06); from 2,3-dimethyl-lbutene, acetone (0.27 f 0.041, 3-methyl-2-butanone, (0.45 f 0.06), and HCHO (0.50 f 0.04); and from 1-methylcyclohexene, 5-acetylpentanal (0.31 f 0.08). These product yield data suggest that the intermediate P-hydroxyal koxy radicals undergo isomerization and/ or reaction with 02 in competition with decomposition, and the decrease in the carbonyl and HCHO yields with increasing carbon number in the 1-alkenes 1-pentene through 1-octene suggests that isomerization of the intermediate P-hydroxyalkoxy radical is occurring.
Introduction Alkenes are important constituents of gasoline fuels ( 1 -41, automobile exhaust (1-3, and ambient air in urban areas (6-8). In the atmosphere, alkenes react with OH radicals, NO3 radicals, and O3 (9), with the daytime OH radical reaction often dominating as the alkene removal process (10,11). Thekinetics and products ofthe gas-phase reaction of OH radicals with alkenes have been studied previously (9, 12). However, while the kinetics of these OH radical reactions are now reasonably well understood for a large number of alkenes (9, 12), only for a few simple acyclic alkenes have product studies been conducted at room temperature and atmospheric pressure of air (9, 12-19). The gas-phase reactions of the OH radical with alkenes proceed mainly by initial addition of the OH radical to the >C=C< bond(s) to form a 8-hydroxyalkyl radical (9, 12), which then adds 0 2 to form the B-hydroxyalkyl peroxy radical (9). For example, for 1-butene: OH
M
+ CH3CHzCH=CHp
---)
CH3CH26HCH20H
and
CH~CH$X(OH)~HZ
1
I
CH3CH2CH(O6)CH>OH
CH3CH2CH(OH)CH206
O2
(1)
O2
In the presence of NO, the 8-hydroxyalkyl peroxy radical reacts with NO to form either the B-hydroxynitrate or the B-hydroxyalkoxy radical plus NOz (9):
-r
CH3CH2CH(OH)CH20N02
CH3CH2CH(OH)CH206 + NO
(2a)
CH3CH2CH(OH)CH26
+
NO2
(2b)
The B-hydroxyalkoxy radical can then react with 02, decompose, or, for zC4 side chains, isomerize via a sixmember transition state (91, as shown, for example, in Scheme 1for the CH~CHZCH(OH)CHZO radical formed after internal OH radical addition to 1-butene [notethat the CH3CH*CH(O)CH*OH radical formed after terminal OH radical addition to 1-butene cannot undergo isomerization via a six-member transition state] with the a-hydroxyalkyl radicals reacting rapidly with 0 2 to form the carbonyl (9): CH,CH,CHOH
+ 0, - CH,CH,CHO + HO,
(3)
The alkoxy radical (OCH2CH2CH(OH)CH20H)formed subsequent to the isomerization reaction is predicted to undergo a further isomerization to form the polyfunctional product HOCH2CH2CH(OH)CH0(9). Previous studies have shown that at room temperature and atmospheric pressure of air the HOCH2CH2O radical formed from OH radical addition to ethene reacts with 0 2 and decomposes (141, while the B-hydroxyalkoxy radicals formed from propene, 1-butene, and trans-2-butene predominantly decompose (13, 15). However, the carbonyl products observed and quantified from the OH radical * Author to whom correspondence should be addressed. Also with the Department of Soil and Environmental Sciences.
+
1674 1 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 29, NO. 6,1995
0013-936)(/95/0929-1674$09.00/0 @ 1995 American Chemical Society
SCHEME 1 02
CH3CH,CH(OH)CH,6
/ \
decomposition
CH,CH,CHOH
+ HCHO
-. CH,CH,CH(OH)CHO + HO, isomerization
CH,CH,CH(OH)CH,OH
products
reactions with isoprene (16, 17, 191, 1-octene (181, and a series of C10H16 monoterpenes (20, 21) indicate that decomposition of the intermediate ,5-hydroxyalkoxyradicals is not the sole process and that other reaction pathways, possibly including isomerization (91, must be occurring. To investigate the expectation that isomerization of the /3-hydroxyalkoxy radical becomes more important as the alkyl substituent chain-length increases,we have measured the carbonyl formationyields from the gas-phasereactions of the OH radical with the 1-alkenes 1-pentene through 1-octene and with 2,3-dimethyl-l-butene in the presence of NO. In addition, the OH radical reaction with l-methylcyclohexene (a model compound for the cyclic monoterpenes a-pinene, 3-carene, and limonene) has been studied.
Experimental Section Experiments were carried out at 298 f 2 K and 740 Torr total pressure of dry synthetic air (80%NZ 20% 0 2 ) in a 5800-L evacuable, Teflon-coated chamber with an in situ multiple reflection optical system interfaced to a Nicolet 7199 Fourier transform infrared (FT-IR)absorption spectrometer, with radiation provided by a 24-kW xenon arc filtered through a 0.25-in.v e x pane to remove wavelengths 1300 nm, and in a 7900-L all-Teflon chamber equipped with two parallel banks of blacklamps at 296 f2 K and 740 Torr total pressure of purified air (-5% relative humidity). Both chambers are fittedwith Teflon-coatedfans to ensure rapid mixing of reactants. Hydroxyl radicals were generated by the photolysis of methyl nitrite (CH30NO)or ethyl nitrite (CZHSONO) in air at wavelengths '300 nm:
+
RCH,ONO
+ hv - RCH,O + NO
+ 0, - RCHO + HO, HO, + NO - OH + NO,
RCH,O
(4)
(5) (6)
where R = H or CH3. NO was added to the reactant mixtures to avoid the formation of 0 3 , and, hence, of NO3 radicals (22). Experiments were carried out in the 7900-Lall-Teflon chamber for the measurement of the formation yields of carbonyls other than HCHO and in the 5800-L evacuable, Teflon-coated chamber for FT-IR absorptionspectroscopic determination of the HCHO formation yields. Because HCHO is the primary photolytic product of methyl nitrite (see above), the photolysis of ethyl nitrite in air was used as the OH radical source for the determination of HCHO formation yields in the 5800-L evacuable chamber. In the 7900-L all-Teflon chamber, OH radicals were generated by the photolysis of CH30N0 in air. For the experimentscarried out in the 7900-L all-Teflon chamber, the initial reactant concentrations (in molecule
~ m units) - ~ were as follows: CH30N0, (1.9-2.2) x 1014; NO, (2.0-2.2) x 1014; and alkene, (2.28-3.03) x 1013. Irradiations were carried out at 20% of the maximum light intensity for 1.5-7.5 min. The alkenes were introduced into the chamber by introducinga measured volume of the liquid alkene into a 1-L Pyrex bulb and then flushing the contents of the bulb into the chamber with a stream of Nz. The concentrations of the alkenes and the carbonyl (apart from HCHO) products were measured during the experiments by gas chromatography with flame ionization detection (GC-FID). For the analyses of 1-pentene, l-hexene, 2,3-dimethyl-l-butene, and 1-methylcyclohexene,gas samples were collected from the chamber into 100 cm3 all-glass, gas-tight syringes, and the gas samples were transferred via a 1-cm3stainless steel loop and gas sampling valve onto a 30-m DB-5 megabore column in a Hewlett Packard (HP) 5890 GC, initially held at -25 "C and then temperature programmed to 200 "C at 8 "C min-'. For the analyses of 1-heptene and 1-octene and the carbonyl products acetone, butanal, pentanal, hexanal, heptanal, 3-methyl-2-butanone,and 5-acetylpentanal, 100 cm3volume gas samples were collected from the chamber onto Tenax-TA solid adsorbent with subsequent thermal desorption at -225 "C onto a 30-m DB-5.625 megabore column in a HP 5710 GC, initially held at 0 "C and then temperature programmed to 200 "C at 8 "C min-'. The data from the HP 5710 GC were computer processed using an HP 3395 integrator and HP Peak96 software. In addition, gas sampleswere collected from the chamber onto TenaxTA solid adsorbent for subsequent thermal desorption and analysis by combined gas chromatographylmass spectrometry (GUMS) using a 60-m HP-5 fused silica capillary column in a HP 5890 GC interfaced to a HP 5970 mass selective detector operated in the scanning mode. Two sets of experiments were carried out for the 1-alkenes in the 5800-L evacuable chamber, with differing initial reactant concentrations and different analysis methods for the alkenes. In the first set of experiments, the initial reactant concentrations (inmolecule ~ munits) - ~ were as follows: CZH~ONO, 2.4 x 1014;NO, 1.2 x 1014;and alkene, 4.8 x 1014. The reactant mixtures were irradiated continuously with FT-IR absorption spectroscopic analyses of the alkene and HCHO concentrations. In the second set of experiments for 1-pentene through 1-octene, the initial reactant concentrations (in molecule ~ m units) - ~ were as follows: C,H~ONO, 2.4 x 1014; NO, 1.8 x 1014 (with an additional 1.2 x 1014molecule ~ m being - ~ added after the second irradiation period); and alkene, 4.8 x Twominute intermittent irradiations were carried out with FTIR spectroscopic and GC-FID analyses during the dark periods. The alkenes were introduced into the chamber by flushing measured partial pressures in a 5.25-L Pyrex bulb into the chamber with a stream of N2 gas. The concentrations of HCHO and (for the first series of experiments) the alkenes were measured by FT-IR absorption spectroscopy,with 64 co-addedinterferograms (scans) per spectrum (1.8 min measurement time) recorded with a fullwidth at half maximum (fwhm)resolution of 0.7 cm-l and a path length of 57.7 m. The GC-FID response factors for the reactants and products were obtained by introducing measured amounts of the chemical into the 7900-L chamber and conducting several replicate GC-FID analyses. The chamber volume was determinedby introducing a measured volume of truns2-butene and determining the trans-2-butene concentraVOL. 29, NO. 6 , 1995 / ENVIRONMENTAL SCIENCE &TECHNOLOGY
1675
tion with a precalibrated HP 5890 GC and looplgas sampling valve injection system. For 2,3-dimethyl-l-butene, 1-heptene, 1-hexene, 3-methyl-2-butanone, 1-methylcyclohexene, 1-octene, and 1-pentene, measured volumes of the liquid were introduced into a 1-L Pyrex bulb, and the contents of the bulb were flushed into the chamber by a stream of NZgas. For acetone, butanal, heptanal, hexanal, and pentanal, partial pressures of the gaseous compound were measured in a 2.03-L Pyrex bulb using an MKS Baratron 0-100 Torr sensor, and the contents of the bulb were flushed into the chamber with a stream of Nz gas. Replicate calibrations of butanal, pentanal, hexanal, heptanal, and 3-methyl-2-butanoneagreed to within 9%. Furthermore, on a relative basis the measured GC-FID response factors for 1-heptene, 1-octene,butanal, pentanal, hexanal, heptanal, acetone, and 3-methyl-2-butanoneon the HP 5710 thermal desorption system agreed with the equivalent carbon numbers (ECNs)as calculated bykanlon and Willis (23) to within &lo%, indicating that the calibration factors were accurate and that the sample collection and thermal desorption procedures were quantitative. FT-IR absorptions of the alkenes were obtained by measuring known partial pressures of the gaseous compoundinto a5.25-LPyrexbulbwithanMKSBaratronO-100 Torr sensor and flushing the contents of the bulb into the 5800-L chamber with FT-IR detection of the authentic compound. IR reference spectra of HCHO were available from previous IR calibrations in this laboratory. Chemicals. The sourcesof the chemicals used and their stated purities were as follows: acetone (>99.6%),Fisher Scientific; butanal [n-butyraldehyde](99%),2,3-dimethyl1-butene(97%),heptanal(95%),1-heptene(99+%),hexanal (99%),1-hexene (99%),3-methyi-2-butanone(99%),1-methylcyclohexene (97%),1-octene (98%),pentanal (99%),and 1-pentene(99%),Aldrich Chemical Co.; and NO (r99.0%), Matheson Gas Products. Methyl nitrite was prepared as described by Taylor etal. (24)while ethyl nitrite was distilled from a commercial 15%solution by weight of CZH~ONO in ethanol (Aldrich Chemical Co.), and both nitrites were stored at 77 K under vacuum.
Results and Discussion Aseries ofmethylnitrite-NO-alkene-air irradiations were carried out in the 7900-L &-Teflon chamber to determine the yields of carbonyl products other than HCHO. Based on GC retention time and mass spectrum matching with authentic standards, butanalwas identified from 1-pentene, pentanal from 1-hexene,hexanal from 1-heptene,heptanal from 1-octene,and acetone and 3-methyl-2-butanonefrom 2,3-dimethyl-l-butene. These products were quantified using the GC-FID response factors determined from the calibrations with authentic standards. In addition, a product observed from 1-methylcyclohexene was tentatively identified as 5-acetylpentanal on the basis of its mass spectrum,which showed a weak (< 1%)molecular ion [MI+ at mlz 128, a base peak at mlz 43, and characteristic odd electron fragment ions at mlz 110 [M - COl+ and rnlz 58, and its infrared (IR) spectrum. The IR spectrum was obtained using a HP 5890 GC interfaced to a HP 5965B FTIR detector and showed absorption bands at 2950,2810, 2713, 1738 (strong), 1362, and 1158 cm-l, very similar to the IR spectra of the corresponding keto aldehydes identSed by Hakola et al.(21)from 3-carene and a-pinene. 5-Acetylpentanal was quantified using the GC-FID response factor calculated from its ECN (23). A corresponding series of 1676 ENVIRONMENTAL SCIENCE &TECHNOLOGY / VOL. 29, NO. 6,1995
ethyl nitrite-NO-alkene-air irradiations were carried out in the 5800-L evacuable chamber to determine the formation yields of HCHO from the l-alkenes by ET-IR absorption spectroscopy. Since the HCHO and other carbonyl products react with the OH radical (9, 121,the secondary reactions of the OH radical with the products were taken into account to determine the product formationyields. These secondary reactions were taken into account as described previously by Atkinson et al.(23,using the recommended or estimated OH radical reaction rate constants at room temperature (inunitsof 10-l2cm3molecule-1 s-1) of (9,12,26')1-pentene, 31.4; 1-hexene, 37.5; 1-heptene, 40.5; 1-octene, 41.9; 2,3dimethyl-1-butene,53.7; 1-methylcyclohexene,94.4; HCHO, 9.37; butanal, 23.5; pentanal, 28.5; hexanal, 29.0; heptanal, 30.4; acetone, 0.219; 3-methyl-2-butanone, 1.50; and 5-acetylpentanal, 31.9. The concentration of the product at time t corrected to take into account its secondary reactions with the OH radical, [product],,,, is given by
where F is the multiplicative correction factor (25) and [prOdUCt]measured is the measured concentration of the product at time t. The multiplicative correction factor F increases with the extent of reaction and with the ratio k(OH product)lk(OH alkene). These factors Ffor the various products were as follows: < 1.30for butanal,
