AAAR 36th Annual Conference October 16 - October 20, 2017 Raleigh Convention Center Raleigh, North Carolina, USA
Abstract View
Efficient and Improved Processing of Chromatographic Data Using Peak Fitting and Deconvolution
GABRIEL ISAACMAN-VANWERTZ, Donna Sueper, Brian Lerner, Kenneth Aikin, Jessica Gilman, Joost de Gouw, Douglas Worsnop, Allen H. Goldstein, Virginia Tech
Abstract Number: 313 Working Group: Instrumentation and Methods
Abstract Decades of data on the composition of organic aerosol have been collected by using chromatography to measure the concentrations of individual components with molecular- and isomer-level specificity. However, the complexity of the ambient atmospheric organic mixture, and its tendency to contain many structurally and chemically similar compounds, challenges the separation and resolution of individual compounds. Consequently, processing chromatographic data from raw signal to integrated peak areas representing ambient concentrations requires substantial operator effort and is a “bottleneck” in analysis that decreases data quality, reduces instrument “uptime”, and increases costs. As advances in field-based instrumentation multiply the quantity and informational density of data generated, data processing will become a major limitation to collecting, analyzing, and interpreting chromatography data. We present here a new approach to processing chromatographic data that efficiently integrates chromatographic peaks by fitting idealized mathematical functions to both well-resolved and co-eluting peaks; this approach reduces analysis time by an order of magnitude. Peak fitting is shown to yield results comparable to traditional integration for 70,000 peaks representing a wide range of compounds measured by two different on-line instruments at four different locations. The quantitative parameters describing the fit (e.g. coefficients, residuals, etc.) are exploited to increase the efficiency of quality control, constrain integration of poor chromatographic peaks, and yield higher accuracy in the deconvolution co-eluting peaks than is possible by traditional methods. We will also explore additional opportunities to use these parameters to improve data analysis and use peak fitting to measure and analyze compounds that are otherwise un-resolvable or inaccessible. Current and planned implementation of the described approach for use by the broader community of atmospheric and environmental scientists will be discussed.