Environ. Sci. Technol. 2005, 39, 4049-4059
Contributions of Organic Peroxides to Secondary Aerosol Formed from Reactions of Monoterpenes with O3 KENNETH S. DOCHERTY,† WILBUR WU,‡ YONG BIN LIM,§ AND P A U L J . Z I E M A N N * ,| Air Pollution Research Center, University of California, Riverside, California 92521
The role of organic peroxides in secondary organic aerosol (SOA) formation from reactions of monoterpenes with O3 was investigated in a series of environmental chamber experiments. Reactions were performed with endocyclic (R-pinene and ∆3-carene) and exocyclic (β-pinene and sabinene) alkenes in dry and humid air and in the presence of the OH radical scavengers: cyclohexane, 1-propanol, and formaldehyde. A thermal desorption particle beam mass spectrometer was used to probe the identity and volatility of SOA components, and an iodometric-spectrophotometric method was used to quantify organic peroxides. Thermal desorption profiles and mass spectra showed that the most volatile SOA components had vapor pressures similar to pinic acid and that much of the SOA consisted of less volatile species that were probably oligomeric compounds. Peroxide analyses indicated that the SOA was predominantly organic peroxides, providing evidence that the oligomers were mostly peroxyhemiacetals formed by heterogeneous reactions of hydroperoxides and aldehydes. For example, it was estimated that organic peroxides contributed ∼47 and ∼85% of the SOA mass formed in the R- and β-pinene reactions, respectively. Reactions performed with different OH radical scavengers indicated that most of the hydroperoxides were formed through the hydroperoxide channel rather than by reactions of stabilized Criegee intermediates. The effect of the OH radical scavenger on the SOA yield was also investigated, and the results were consistent with results of recent experiments and model simulations that support a mechanism based on changes in the [HO2]/[RO2] ratios. These are the first measurements of organic peroxides in monoterpene SOA, and the results have important implications for understanding the mechanisms of SOA formation and the potential effects of atmospheric aerosol particles on the environment and human health.
Introduction Monoterpenes are an important class of atmospheric hydrocarbons consisting of cyclic and acylic alkenes with the molecular formula C10H16. They are emitted into the tropo* Corresponding author phone: (951) 827-5127; fax: (951) 827-5004; e-mail:
[email protected]. † Also in the Environmental Toxicology Graduate Program. ‡ Also in the Department of Biology. § Also in the Department of Chemistry. | Also in the Department of Environmental Sciences, Department of Chemistry, and Environmental Toxicology Graduate Program. 10.1021/es050228s CCC: $30.25 Published on Web 04/27/2005
2005 American Chemical Society
sphere in large amounts by vegetation, with recent global estimates indicating that monoterpene emissions constitute ∼10% of total biogenic hydrocarbon emissions and are approximately equal to emissions from anthropogenic sources (1, 2). In the atmosphere, monoterpenes are oxidized by OH and NO3 radicals and O3 (3, 4). Whereas reactions with OH radicals are often the major degradation pathway, reactions with O3 form many low-volatility products that can partition into particles and contribute significantly to secondary organic aerosol (SOA) formation (5). SOA accumulates in fine particles (diameter cyclohexane/humid > 1-propanol > formaldehyde, while sabinine followed the same trend with yields of 38 and 10% for the cyclohexane/dry and 1-propanol reactions, respectively. Organic peroxide yields, YOP/MT, are shown in Figure 6B. The range of values was ∼3-12%. The trends were similar to those observed for the SOA yields. As a result, when organic peroxide yields for a monoterpene were calculated relative to SOA mass, YOP/SOA, they were nearly independent of the conditions. The YOP/SOA values (equal to YOP/MT/YSOA/MT) had a range of ∼25-103%, as shown in Figure 6C. For R- and β-pinene, the values measured for the four conditions were not statistically different, resulting in average values of 47 ( 12 and 85 ( 16%, respectively. Values for ∆3-carene and sabinene were 34 ( 13 and 98 ( 7%, respectively. The trends observed in YOP/MT and YSOA/MT values for the exocyclic alkenes are the same as we observed previously for SOA yields from β-pinene (23). We suggested this occurred because reactions of OH radicals with different scavengers caused the [HO2]/[RO2] ratio to increase in the following order: formaldehyde > 1-propanol > cyclohexane. This increased the rate of HO2-RO2 reactions at the expense of RO2-RO2 reactions and led to more volatile products and less SOA. In light of the present results, it appears that those
FIGURE 6. Measured yields of (A) SOA and (B and C) SOA peroxides from the reactions of r-pinene, ∆3-carene, β-pinene, and sabinene with O3 in dry and humid air in the presence of cyclohexane, 1-propanol, and formaldehyde OH radical scavengers. Error bars are one standard deviation for replicate experiments. products are primarily hydroperoxides that are too volatile to partition to the particle phase, either permanently or long enough to react with an aldehyde to form a nonvolatile peroxyhemiacetal. As a result, both organic peroxide and SOA yields decrease. Examples of these reactions are shown in Figure 1. Formation of the lowest volatility hydroperoxide, peroxypinic acid, requires two RO2-RO2 reactions and an HO2 reaction, formation of the more volatile peroxypinalic acid involves one less RO2-RO2 reaction, and formation of the most volatile hydroperoxides (I) and (II) requires no RO2-RO2 reaction. Recently, Jenkin (28) employed a chemical model containing ∼200 compounds to investigate the effect of OH radical scavengers on SOA yields for R- and β-pinene. The results for β-pinene supported the [HO2]/[RO2] hypothesis and provided a more detailed explanation of the effect. The model predicted no significant effect of the OH radical scavenger on SOA yields from R-pinene, which is consistent with our results. Keywood et al. (52) employed a simpler model using organic acids as a surrogate for SOA to explain VOL. 39, NO. 11, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4055
TABLE 1. Measured SOA and Organic Peroxide Yields from Reactions of Monoterpenes with O3 monoterpene R-pinene
∆-3carene β-pinene
sabinene
OH scavenger
RH (%)
