AAAR 31st Annual Conference
October 8-12, 2012
Hyatt Regency Minneapolis
Minneapolis, Minnesota, USA
Abstract View
The Role of the Precursor’s Volatility and Structure on Secondary Organic Aerosol Formation: From Experiments to Models
SHANTANU JATHAR, Marissa Miracolo, Daniel S. Tkacik, Peter Adams, Allen Robinson, Carnegie Mellon University
Abstract Number: 304 Working Group: Aerosol Chemistry
Abstract Photo-oxidation experiments were conducted using a smog chamber on unburned evaporated fuel to investigate the influence of the precursor’s volatility and structure on secondary organic aerosol (SOA) formation. We perform experiments on gasoline, two Fischer-Tropschs (Sasol and Shell), two JP-8s and six diesels; they span a modest range of volatility and structure. We find that for a unit amount of fuel reacted, diesel forms the most SOA followed by JP-8, FT-Shell, gasoline and FT-Sasol. Both JP-8s exhibit the same SOA potential while all diesels exhibit very similar SOA potential. Chemically, the SOA from our experiments is lightly oxygenated (O:C~0.2-0.4) and looks similar to semi-volatile oxygenated organic aerosol (SV-OOA). A traditional SOA model (SAPRC07 lumping, Murphy and Pandis (2010) yields) results in a reasonable model-measurement comparison (fractional error = 76%, fractional bias = 26%) for the JP-8, FT-Shell and diesel experiments. The model over-predicts SOA in the FT-Sasol experiments because it is not configured to account for branched alkanes which have lower yields compared to straight/cyclic alkanes and which mostly constitute FT-Sasol (88%). The model over-predicts SOA in the gasoline experiments probably because the Murphy and Pandis (2010) yields for single-ring aromatics are biased too high. When we add a branched alkane and multi-ring aromatic model species to SAPRC07 and adjust yields for the traditional SOA model, we improve the model-measurement comparison significantly (fractional error = 54%, fractional bias=-2%). On using a volatility-based model that does not account for differences in the precursor’s structure, we find that the model (when fit) is marginally better than the traditional SOA model (fractional error = 73%, fractional bias=13%). This implies that the SOA formation across these precursor fuels can reasonably be explained by differences in their volatility alone. Since the volatility-based model runs on only 4 parameters compared to SAPRC’s 30 parameters, it would be much more efficient to use in computationally expensive chemical transport models.