AAAR 34th Annual Conference
October 12 - October 16, 2015
Hyatt Regency
Minneapolis, Minnesota, USA
Abstract View
Gas-Particle Partitioning of Primary Organic Aerosol from Vehicles Measured in a Traffic Tunnel
XIANG LI, Timothy Dallmann, Albert A. Presto, Carnegie Mellon University
Abstract Number: 307 Working Group: Carbonaceous Aerosols in the Atmosphere
Abstract Motor vehicles are a major contributor of carbonaceous particulate matter (PM), which consists of organic aerosol (OA) and elemental carbon (EC), in the urban environment. Primary organic aerosols (POA) emitted by vehicles are semivolatile. Accurate representations of POA gas-particle partitioning are required to estimate the aerosol burden in the atmosphere. Although several dynamometer studies were conducted to determine the volatility of vehicle emitted POA, little work has been done to show whether the dynamometer results can reproduce POA partitioning in the atmosphere.
In this work we studied the volatility of vehicle emitted PM in a traffic tunnel in Pittsburgh, PA. PM size from 7-400nm and nonrefectory PM composition were measured by a pair of scanning mobility particle sizers and an Aerosol Chemical Speciation Monitor, respectively. The gas-particle partitioning of vehicle emitted POA was determined using three independent approaches: 1) heating by a thermodenuder (TD) cycling between 40-100C with a residence time about 19s, 2) quartz filter samples analyzed by thermal desorption/gas chromatography/mass spectrometry (TD-GC-MS), 3) bare quartz and quartz behind Teflon filter samples analyzed by the Sunset OCEC.
Results from three independent approaches consistently show the vehicle emitted POA is semivolatile. The PM evaporation inside the TD can be adequately described by the TD model using the volatility distribution from May et al (2013); and the volatility distribution determined using TD-GC-MS in this study is similar with that of May et al (2013). The variability of volatility distributions measured in different time periods with large difference on fleet compositions is very limited, which suggests that a single volatility distribution may be adequate to describe the vehicular PM volatility in the urban environment.