Secondary Organic Aerosol from Photooxidation of Polycyclic Aromatic Hydrocarbons KABINDRA M. SHAKYA AND ROBERT J. GRIFFIN* Department of Civil and Environmental Engineering, Rice University, 6100 Main St., Houston, Texas 77005, United States Climate Change Research Center, University of New Hampshire, Durham, New Hampshire 03824, United States Received June 8, 2010. Revised manuscript received September 10, 2010. Accepted September 21, 2010. Secondary organic aerosol (SOA) formation from the photooxidation of five polycyclic aromatic hydrocarbons (PAHs, naphthalene, 1- and 2-methylnaphthalene, acenaphthylene, and acenaphthene) was investigated in a 9-m 3 chamber in the presence of nitrogen oxides and the absence of seed aerosols. Aerosol size distributions and PAH decay were monitored by a scanning mobility particle sizer and a gas chromatograph with a flame ionization detector. Over a wide range of conditions, the aerosol yields for the investigated PAHs were observed to be in the range of 2-22%. The observed evolution of aerosol and PAH decay indicate that light and oxidant sources influence the time required to form aerosol and the required threshold reacted concentration of the PAHs. The SOA yields also were related to this induction period and the hydroxyl radical concentrations, particularly for smaller aerosol loadings ( <∼6 µgm -3 ). Estimation of SOA production from oxidation of PAHs emitted from mobile sources in Houston shows that PAHs could account for more than 10% of the SOA formed from emissions from mobile sources in this region. Introduction Secondary organic aerosol (SOA) constitutes a significant fraction of the fine aerosol in the atmosphere (1). However, there is large uncertainty in predicted SOA formation, with global modeling indicating a range of 12-70 Tg SOA yr -1 (2). Reactive organic gases (ROGs) emitted from natural and anthropogenic sources are oxidized by hydroxyl (OH) and nitrate (NO 3 ) radicals, ozone (O 3 ), and/or halogen atoms. The subset of oxidation products that are non- and/or semivolatile partitions into either a new or pre-existing aerosol phase in one SOA formation pathway (3). There still may exist a number of unidentified compounds that con- tribute to SOA formation in the atmosphere via this route (4). Identifying these compounds, quantifying their SOA yields, and studying their effects on climate and health are critical. The main objective of this study is to quantify further the conversion of gas-phase polycyclic aromatic hydrocar- bons (PAHs) with two aromatic rings into particle-phase products. These PAHs are products of incomplete combustion, are ubiquitous in the environment, and are emitted from both natural and anthropogenic sources. The main anthropogenic emission sources are coal, oil, gas, wood, tobacco, and refuse combustion and domestic heating (5, 6). Naphthalene, 1- and 2-methylnaphthalene, acenaphthylene, and acenaph- thene are often the dominant PAHs found in an urban environment (7, 8), and their gas-phase oxidation forms products, some toxic, that partition into the aerosol phase (9–15). A large fraction of photooxidation products from PAHs are reported to form SOA in chamber experiments (10, 11). PAHs could contribute an important fraction of urban SOA based on using phthalic acid and 4-nitro-1-naphthol as naphthalene SOA tracers (13). Conventionally, SOA formation has been described by a dimensionless fractional aerosol yield, Y (16), defined as the ratio of the mass concentration of SOA formed (ΔM) to the mass concentration of ROG consumed (ΔROG): Using a semiempirical partitioning model for semivolatile products, SOA yield can be expressed as a function of the organic aerosol mass concentration (17): where R i is the mass-based stoichiometric coefficient for product i, K om,i (m 3 µg -1 ) is its equilibrium absorptive partitioning coefficient between the gas phase and an absorbing condensed organic medium based on Raoult’s law (18), and ΔM o (µgm -3 ) is the total organic aerosol mass concentration. Generally, two hypothetical products are used to parametrize yields. Combination of eqs 1 and 2 yields a required threshold ΔROG (THC or threshold reacted hy- drocarbon concentration) to initiate SOA formation within chambers (19, 20): To further understand the volatility distribution of the products, a volatility basis set (VBS) approach can be used (21–23): where  is the overall aerosol mass fraction, c j * is the effective saturation concentration of product j (analogous to the inverse of the partitioning coefficient), and c OA is the organic aerosol mass concentration. The VBS approach typically uses nine bins with set c j * that step by an order of magnitude to fit observed aerosol yields () and masses (c OA ) to determine R j . Experimental Section A9m 3 Teflon laboratory chamber (surface-to-volume ratio 2.97 m -1 ) (Welch Fluorocarbon) was used for all experiments. A set of Sylvania (30 W, G30T8; 254-nm ultraviolet (UV)) germicidal lamps (10 tubes) was used for the experiments in which hydrogen peroxide (H 2 O 2 ) was used as the OH source; a set of Sylvania (40 W, 350 BL; 365-nm UV) blacklights (20 tubes) was used for the experiments in which nitrous acid (HONO) was used as the OH source. Illumination of the appropriate set of lights initiated the experiments by pho- * Corresponding author phone: 713-348-2093; fax: 713-348-5268; e-mail: rob.griffin@rice.edu. Y ) ΔM ΔROG (1) Y ) ΔM o ∑ i R i K om,i 1 + ΔM o K om,i (2) THC ) ( ∑ i R i K om,i ) -1 (3)  ) Δc OA ΔROG ) ∑ j R j 1 + (c j * /c OA ) (4) Environ. Sci. Technol. 2010, 44, 8134–8139 8134 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 21, 2010 10.1021/es1019417  2010 American Chemical Society Published on Web 10/04/2010