Abstract View
The Impact of Structure on the Estimation of Atmospherically Relevant Physicochemical Parameters
GABRIEL ISAACMAN-VANWERTZ, Bernard Aumont, Virginia Tech
Abstract Number: 408
Working Group: Instrumentation and Methods
Abstract
Recent advances in field-deployable mass spectrometry of gas- and particle-phase organic compounds have provided unprecedented characterization of atmospheric mixtures. However, while organic carbon across the entire range of atmospheric properties has become measurable by current state-of-the-art tools, many of these instruments identify analytes only by elemental formula with little or no structural information. This lack of structural information stymies the estimation of many physicochemical properties, which have strong structural dependencies. Consequently, a substantial body of work has generated and utilized empirical parameterizations of molecular properties (e.g. volatility and reactivity) based on elemental formulas, and little work has quantified the extent to which ignoring molecular structure degrades estimates of these parameters. In this presentation, we compare the estimated vapor pressures, Henry’s Law Constants, and OH reactivity of isomers of the same molecular formulas. Differences between isomers are compared to uncertainties between different structure-based estimation methods, and to errors in formula-based estimation methods. This analysis is performed using a set of ~35,000 structures (of ~1,200 formulas) predicted by the GECKO-A explicit chemical mechanism generator as atmospheric oxidation products of α-pinene, decane, and toluene. We find that differences between isomers are greater than differences between structure-based methods, indicating structural information improves estimates. Furthermore, particle-phase components suffer higher ranges and uncertainties in their estimated properties. However, formula-based estimation is possible for all three parameters with little bias and an approximately normally distributed error. Consequently, formula-based estimation is reasonable when necessary, but creates uncertainty commensurate with the lack of structural information.