es404024h_si_001.pdf (196.68 kB)

Gas-Phase CO2 Subtraction for Improved Measurements of the Organic Aerosol Mass Concentration and Oxidation Degree by an Aerosol Mass Spectrometer

Download (196.68 kB)
journal contribution
posted on 17.12.2013 by S. Collier, Q. Zhang
The Aerodyne aerosol mass spectrometer (AMS) has been widely used for real-time characterization of the size-resolved chemical composition of sub-micrometer aerosol particles. The first step in AMS sampling is the pre-concentration of aerosols while stripping away the gas-phase components, which contributes to the high sensitivity of this instrument. The strength of the instrument lies in particle phase measurement; however, ion signals generated from gas-phase species can influence the interpretation of the particle-phase chemistry data. Here, we present methods for subtracting the varying contributions of gas-phase carbon dioxide (CO2) in the AMS spectra of aerosol particles, which is critical for determining the mass concentration and oxygen-to-carbon (O/C) ratio of organic aerosol. This report gives details on the gaseous CO2 subtraction analysis performed on a high-resolution time-of-flight aerosol mass spectrometer (HR-ToF-AMS) data set acquired from sampling of fresh and diluted vehicle emissions. Three different methods were used: (1) collocated continuous gas-phase CO2 measurement coupled with periodic filter tests consisting of sampling the same particle-free air by the AMS and the CO2 analyzer, (2) positive matrix factorization (PMF) analysis to separate the gas- and particle-phase signals of CO2+ at m/z 44, and (3) use of the particle time-of-flight (PTOF) size-resolved chemical information for separation of gas- and particle-phase signals at m/z 44. Our results indicate that these three different approaches yield internally consistent values for the gas/particle apportionment of m/z 44, but methods 2 and 3 require certain conditions to be met to yield reliable results. The methods presented are applicable to any situation where gas-phase components may influence the PM signal of interest.