Article pubs.acs.org/est
Aqueous-Phase Secondary Organic Aerosol and Organosulfate Formation in Atmospheric Aerosols: A Modeling Study V. Faye McNeill,* Joseph L. Woo, Derek D. Kim, Allison N. Schwier, Neal J. Wannell, Andrew J. Sumner, and Joseph M. Barakat Department of Chemical Engineering, Columbia University, New York, New York 10027, United States S Supporting Information *
ABSTRACT: We have examined aqueous-phase secondary organic aerosol (SOA) and organosulfate (OS) formation in atmospheric aerosols using a photochemical box model with coupled gas-phase chemistry and detailed aqueous aerosol chemistry. SOA formation in deliquesced ammonium sulfate aerosol is highest under low-NOx conditions, with acidic aerosol (pH = 1) and low ambient relative humidity (40%). Under these conditions, with an initial sulfate loading of 4.0 μg m−3, 0.9 μg m−3 SOA is predicted after 12 h. Low-NOx aqueous-aerosol SOA (aaSOA) and OS formation is dominated by isoprene-derived epoxydiol (IEPOX) pathways; 69% or more of aaSOA is composed of IEPOX, 2methyltetrol, and 2-methyltetrol sulfate ester. 2-Methyltetrol sulfate ester comprises >99% of OS mass (66 ng m−3 at 40% RH and pH 1). In urban (high-NOx) environments, aaSOA is primarily formed via reversible glyoxal uptake, with 0.12 μg m−3 formed after 12 h at 80% RH, with 20 μg m−3 initial sulfate. OS formation under all conditions studied is maximum at low pH and lower relative humidities (99% of OS mass (66 ng m−3 at 40% RH and pH 1). Figure 3 considers the relative
■
DISCUSSION In contrast to what has been predicted for SOA generated in cloudwater,1,3 organic acids were predicted to contribute less than 2% of aaSOA mass under low-NOx conditions and under 25% under high-NOx conditions. Total SOA mass generated in the aqueous aerosol phase is predicted to be highest under lowNOx conditions where aaSOA formation from IEPOX is favorable (i.e., low pH and low RH). For low-NOx conditions with aerosol pH = 1, aaSOA O:C ratio was 1.08 (±0.04) after 12 h of simulation, with a mild dependence on relative humidity, while H:C = 2.39 (±0.008). For high-NOx conditions, O:C varied from 2 to 1.8 and H:C varied from 3 to 2.75 over an RH range of 40−80%. In all cases, the O:C values are higher at a given H:C ratio for aaSOA than what is typical for oxidized primary organic aerosol or “traditional” SOA derived from the condensation of semivolatile oxidized VOCs reported by Heald et al.63 This is consistent with the notion that aqueous-phase aerosol chemistry may be a source of highly oxidized, low volatility organic aerosol (LV-OOA).64,65 Although the emissions and initial gas-phase conditions used in the model were not tailored to any specific location, comparing GAMMA model output to measurements made at representative locations in the Southeast U.S. can provide some insight into the potential significance of aqueous aerosol SOA formation pathways. Tanner et al. reported organic carbon (OC) values of 3.15−3.53 μg m−3 from the SEARCH rural site at Yorkville, GA during summer months in 2002−2004.59 On the basis of the results in Figure 1 and the predicted O:C ratio, our aaSOA mass predictions comprise within 0.2−11.25% of this value after 12 h of simulation, depending on aerosol RH and pH. This is consistent with the observations of Lin et al., who reported that IEPOX tracers comprised 7.9% of OC in Yorkville in Summer 2010.66 They reported the average mass concentration of 2-methyltetrol to be 330 ng m−3 during that time period. GAMMA reproduces this observed mass concentration after 12 h of simulation at 40−45% RH and pH = 1. Lin et al. also observed 72 ng m−3 organosulfates derived from 2-methyltetrol, suggesting a branching ratio of 21.8%. This is lower than that suggested by the bulk phase laboratory studies of Eddingsaas et al. (2010) and used in GAMMA. This difference suggests that this chemistry may proceed differently in the concentrated aerosol environment as compared to a bulk laboratory solution. Our predictions of aaSOA mass generated under high-NOx conditions after 12 h represent approximately 1% of the WSOC measured by Hennigan et al. in Atlanta (3−4 μgC m−3).67 In a control simulation, we found that changing the urban isoprene initial concentrations and emissions rates to match the remote case (roughly 10× increase in both emissions and initial concentration) resulted in a 10% increase in overall aaSOA mass (9% increase in mass due to CVOC and 17% increase in organic acid SOA mass). Atmospheric Implications. The results of these simulations using our newly developed model, GAMMA, shed light on aaSOA and OS formation in aerosol water. The maximum aaSOA was formed under conditions where the IEPOX SOA formation pathways are active (low NOx, low pH, and low RH).
Figure 3. In-particle concentrations of organosulfate species formed via radical mechanisms under low-NOx conditions (65% RH, pH = 1) as a function of time after sunrise.
contributions of each of the photochemically derived organosulfate species under low-NOx conditions. Note that concentration is shown on a log scale, and the absolute concentration of each of these species is very small. Glyoxal sulfate is predicted to be the most abundant photochemically derived OS species. Note that, due to their very small Henry’s Law constants, isoprene, methylvinylketone, and methacrolein concentrations are too low in the particle phase for significant direct OS formation.31 High-NOx. Glyoxal uptake dominates aaSOA formation under high-NOx conditions (i.e., when IEPOX is not expected to be formed from isoprene oxidation45,46) (Figure 2, panels (B) and (D)). As a result, in contrast to the low-NOx scenario, aaSOA formation is insensitive to pH and increases with increasing relative humidity (Figure 2, panels (A) and (C)). Mass of aaSOA predicted after 12 h ranges from 58 to 120 ng m−3 at 40% and 80% RH, respectively (Figure 2, panel (A)). OS does not contribute significantly to aaSOA mass under high-NOx conditions (maximum 0.8 ng m−3 at 40% RH and 8078
dx.doi.org/10.1021/es3002986 | Environ. Sci. Technol. 2012, 46, 8075−8081
Environmental Science & Technology
Article
considerably higher than for water or bulk electrolyte solutions (which are less concentrated than aerosol at
