American Association for Aerosol Research - Abstract Submission

AAAR 32nd Annual Conference
September 30 - October 4, 2013
Oregon Convention Center
Portland, Oregon, USA

Abstract View


Ambient Primary PM2.5 from Petroleum Refinery Operations

LI DU, Jay Turner, Washington University in St. Louis

     Abstract Number: 485
     Working Group: Urban Aerosols

Abstract
The Roxana Air Quality Study (RAQS), being conducted in Roxana (IL) at the fenceline of a petroleum refinery, will include three years of 24-hour integrated 1-in-6 day sampling for PM2.5 mass and speciation, volatile organic compounds (VOC), and gaseous carbonyl compounds. Routine operations at petroleum refineries are generally not considered to be a major source of primary fine particulate matter emissions. However, episodic releases from unit operations such as fluidized-bed catalytic cracking (FCC) can have considerable impacts. The Chellum group at the University of Houston has demonstrated a framework to track PM emissions from FCC operations using rare earth elements (REEs) as elemental tracers. We have analyzed RAQS PM2.5 samples from an 18-month period for REEs and interpreted the data using a two-source model assuming soil and FCC operations are the dominant sources of lanthanum (La) and cerium (Ce). On most days the La/Ce ratio is ~0.8 which is consistent with the bulk composition reported in the literature for soil. However, on some days the La concentration is much higher than predicted by the soil ratio. 20% of the 90 samples analyzed to date exhibit La/Ce > 1.6 (i.e. more than double the ratio expected for soil) and the 95th percentile of La/Ca ratio 4.6 (i.e. more than five time the ratio expected for soil). Nonparametric wind regression on the excess La, relative to soil, is generally consistent with the location of the FCC unit. This presentation summarizes the REEs patterns observed at this site including the two-source modeling results. PM2.5 major species reported for the Chemical Speciation Network (EC, OC, major ions, and elements by XRF) at the RAQS site are placed in a broader context by comparing to speciation network data at four other sites in the St. Louis metropolitan area.