Ozone loss in soot aerosols

The fractal-like structure of atmospheric soot (e.g., elemental carbon) pro vides a large surface area available for heterogeneous chemistry in the upp er troposphere and lower stratosphere [Blake and Kato, 1995]. One potential ly important reaction is ozone decomposition on soot. Although extensively studied in the laboratory, a wide range of reaction probabilities have been observed (gamma similar to 10(-3) to gamma similar to 10(-7)) which have b een attributed to differences in reactivity between fresh (i.e., nonoxidize d) versus aged (i.e., oxidized) soot [Schurath and Naumann, 1998]. The impo rtance in understanding soot-ozone chemistry is particularly important in l ight of recent nighttime field measurements [Berkowitz et al., 2000] made o ver Portland, Oregon. The data revealed episodes of an anticorrelation betw een ozone mixing ratio and aerosol surface area density. During these episo des a single scattering albedo in the range 0.8-0.9 was measured, indicatin g an increased absorptive component of the aerosol, perhaps due to elementa l carbon. In addition, an increase in the concentration of aerosols contain ed in the small size range of the fine mode (<0.1-0.15 mu m) was observed, suggestive of new aerosol formation. In this article we attempt to explain these field observations. One explanation of the field observations is ozon e loss occurring on atmospheric soot aerosol. Here we present laboratory re sults obtained using a static aerosol reactor that indicate that direct ozo ne loss on soot aerosol is unlikely under ambient conditions in the troposp here. An alternative and more likely explanation of the field data is based on ozone-mediated organic aerosol production. This could occur by either n ighttime nitrate radical oxidation or direct ozone oxidation of hydrocarbon s as suggested previously [Starn et al., 1998; Griffin ed al., 1999; Kamens et al., 1999; Yu et al., 1999; De Gouw and Lovejoy, 1998].