An SOA Model for Toluene Oxidation in the Presence of Inorganic

Dec 17, 2009 - National Institute for Occupational Safety and Health,. 1095 Willowdale Rd., M/S 4020, Morgantown,. West Virginia 26505, and Department...
0 downloads 0 Views 430KB Size
Environ. Sci. Technol. 2010, 44, 727–733

An SOA Model for Toluene Oxidation in the Presence of Inorganic Aerosols G A N G C A O † A N D M Y O S E O N J A N G * ,‡ National Institute for Occupational Safety and Health, 1095 Willowdale Rd., M/S 4020, Morgantown, West Virginia 26505, and Department of Environmental Engineering Sciences, University of Florida, P.O. Box 116450, Gainesville, Florida 32611

Received June 8, 2009. Revised manuscript received November 6, 2009. Accepted November 24, 2009.

A predictive model for secondary organic aerosol (SOA) formation including both partitioning and heterogeneous reactions is explored for the SOA produced from the oxidation of toluene in the presence of inorganic seed aerosols. The predictive SOA model comprises the explicit gas-phase chemistry of toluene, gas-particle partitioning, and heterogeneous chemistry. The resulting products from the explicit gas phase chemistry are lumped into several classes of chemical species based on their vapor pressure and reactivity for heterogeneous reactions. Both the gas-particle partitioning coefficient and the heterogeneous reaction rate constant of each lumped gas-phase product are theoretically determined using group contribution and molecular structure-reactivity. In the SOA model, the predictive SOAmassisdecoupledintopartitioning(OMP)andheterogeneous aerosolproduction(OMH).OMP isestimatedfromtheSOApartitioning model developed by Schell et al. (J. Geophys. Res. 2001, 106, 28275-28293) that has been used in a regional air quality model (CMAQ 4.7). OMH is predicted from the heterogeneous SOA model developed by Jang et al. (Environ. Sci. Technol. 2006, 40, 3013-3022). The SOA model is evaluated using a number of the experimental SOA data that are generated in a 2 m3 indoor Teflon film chamber under various experimental conditions (e.g., humidity, inorganic seed compositions, NOx concentrations). The SOA model reasonably predicts not only the gas-phase chemistry, such as the ozone formation, the conversion of NO to NO2, and the toluene decay, but also the SOA production. The model predicted that the OMH fraction of the total toluene SOA mass increases as NOx concentrations decrease: 0.73-0.83 at low NOx levels and 0.17-0.47 at middle and high NOx levels for SOA experiments with high initial toluene concentrations. Our study also finds a significant increase in the OMH mass fractionintheSOAgeneratedwithlowinitialtolueneconcentrations, compared to those with high initial toluene concentrations. On average, more than a 1-fold increase in OMH fraction is observed when the comparison is made between SOA experiments with 40 ppb toluene to those with 630 ppb toluene. Such an observation implies that heterogeneous reactions of the second-generation products of toluene oxidation can contribute considerably to the total SOA mass under atmospheric relevant conditions. * Corresponding author phone: (352) 846-1744; fax: (352) 3923076; e-mail: [email protected]. † National Institute for Occupational Safety and Health. ‡ University of Florida. 10.1021/es901682r

 2010 American Chemical Society

Published on Web 12/17/2009

Introduction Secondary organic aerosol (SOA) formed from gas-phase products of volatile organic compounds (VOC) oxidation constitutes a significant organic fraction of fine particulate matter in ambient air. The potential impacts of fine particulate matter on climate change, visibility, and human health have stimulated research interest in SOA formation from either biogenic or anthropogenic precursors. The SOA formation mechanisms and the SOA models (1) for a single organic precursor have often been studied using laboratory chamber experiments due to the complexity in real atmospheric aerosols. Recent laboratory studies (2-5) have evinced that heterogeneous reactions are important pathways for SOA generation, in addition to the gas-particle partitioning. Moreover, many laboratory SOA studies (3, 4, 6-10) show that heterogeneous reactions in aerosols can be accelerated by the presence of inorganic acidic species, leading to greater SOA production. Field data for the aerosol compositions characterized using aerosol mass spectroscopy (AMS) often found that organic and inorganic components in the accumulation mode are internally mixed (11, 12). However, the current SOA model that is incorporated in the regional air quality model does not consider either the newly identified contribution of heterogeneous chemistry of organics or the chemical interaction between inorganic and organic components in aerosols, although efforts have been made for the latest version of the SOA model in the air quality model (CMAQ). The SOA submodel used for air quality tends to underestimate the observed SOA formation in the atmosphere (13, 14). In the recent study, Donahue et al. (15) also suggested that the current SOA model frames used in the air quality model oversimplified by using either very low saturation concentrations for SOA components (16, 17) or the two product partitioning-base model (18). Such oversimplification potentially would result in deviation of model prediction from observation. Inclusion of heterogeneous reactions of atmospheric organic compounds in the presence of inorganic species into the SOA model will likely overcome the shortcomings of the current SOA model and reduce the deviation between field data and the predicted SOA mass. Jang et al. (19) in their recent study for the first time incorporated heterogeneous reactions of organic compounds into the SOA model. In their model, a mathematical relationship between rate constants and molecular structures of various carbonyl compounds has been semiempirically derived for heterogeneous reactions and has been later applied to develop a predictive SOA model for the SOA produced from R-pinene ozonolysis in the presence of inorganic seed (2). The SOA model that includes both the gas-particle partitioning and heterogeneous reactions greatly improves predictability of SOA mass compared to the one based on partitioning solely. On the global scale the total aromatic VOCs account for 15% of the annual anthropogenic nonmethane hydrocarbon budget (20, 21). Toluene is the most abundant aromatic VOC in the urban atmosphere, comprising about 20-40% of the total aromatic VOCs (22), and produces high SOA yields compared to other aromatic VOCs (23). In recent laboratory chamber studies, researchers (6, 7, 24) have reported that SOA mass produced from photo-oxidation of toluene is much higher than that observed by earlier studies (25, 26). Hence, studies on toluene SOA production once again receive considerable attention. In this study, the toluene SOA model in the presence of inorganic seed has been developed using the newly derived VOL. 44, NO. 2, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

727

TABLE 1. Experimental Conditions and Resulting SOA Data from Photooxidation of Toluenea,b date

exp no.c

%RH

[H+] (µg/m3)

initial NO (ppb)

initial NO2 (ppb)

∆Tol (ppm)

Vseed (nL/m3)

Vmix (nL/m3)

OM (µg/m3)

mass ratios of org to inorg

6/14/06 6/9/06 6/14/06 6/9/06 1/25/08 1/26/08 1/24/08 2/17/08 2/15/08 2/7/08 2/5/08 1/28/08

LNOx-NA-LRH LNOx-NA-MRHe LNOx-A-LRHe LNOx-A-MRH MNOx-NA-LRHe MNOx-NA-MRH MNOx-A-LRHe MNOx-A-MRH HNOx-NA-LRH HNOx-NA-MRHe HNOx-A-LRHe HNOx-A-MRH

18.4 48.2 17.0 48.2 15.8 41.0 21.1 38.5 14.6 42.7 19.9 40.9

0.01 0.01 0.80 0.69 0.01 0.02 1.05 0.87 0.01 0.02 1.19 0.94