10th International Aerosol Conference September 2 - September 7, 2018 America's Center Convention Complex St. Louis, Missouri, USA
Abstract View
Prediction of Atmospheric Organic Aerosol Concentrations From Carbonyl Absorption in the Mid-Infrared
Matteo Reggente, Robin Modini, Giulia Ruggeri, SATOSHI TAKAHAMA, Andrew Weakley, Alexandra Boris, Ann Dillner, Provat Saha, Andrew Grieshop, Christoph Hueglin, Christopher Pöhlker, Meinrat O. Andreae, Samara Carbone, Paulo Artaxo, EPFL
Abstract Number: 1313 Working Group: Carbonaceous Aerosol
Abstract Fourier Transform-Infrared (FT-IR) spectra of particulate matter is chemically informative with regards to molecular bonds and functional groups present in the sample. The process of building calibration models for various PM constituents (e.g., functional groups, carbon content) allows us to extract useful features for quantitative prediction, and inform us regarding the most explanatory vibrational modes (Dillner and Takahama 2015, Takahama et al. 2016).
We develop a new algorithm for eliminating uninformative wavelengths in spectroscopic calibration for prediction of organic carbon (OC) concentrations equivalent to that of thermal optical reflectance (TOR) measurements. Using this algorithm together with FT-IR spectra obtained from Teflon filters and TOR OC concentrations measured from collocated quartz fiber filters in 526 samples from 7 sites in the US IMPROVE monitoring network for 2011, we build a calibration model that accurately predicts TOR-equivalent OC mass concentrations in PM2.5 for more than two thousand samples collected across the same sites in 2011 and 11 additional sites in 2013. Surprisingly, this model uses 10 wavelengths spanning a very narrow region of ~0.6 micrometers in the mid-infrared region. Investigation with reference compounds in laboratory-generated aerosols suggests that this region is not associated with hydrocarbon C-H structures, which is typically prominent in atmospheric PM (more than 50% by mass in IMPROVE network samples; Ruthenburg et al. 2014). However, we find strong association with carboxylic acids, and weaker but non-negligible association with ketonic carbonyls. We further find that these 10 wavelengths can be used to predict PM2.5 thermal optical transmittance (TOT) OC in PM2.5 in Zurich, Switzerland, and submicron OM in the Amazon rainforest, Brazil, (GoAmazon campaign) and Centreville, Alabama, USA (Southern Oxidant and Aerosol Study campaign). We discuss further interpretation of this model and implications for organic aerosol modeling.
[1] Dillner, A.M., Takahama, S., 2015. Predicting ambient aerosol thermal-optical reflectance (TOR) measurements from infrared spectra: organic carbon. Atmospheric Measurement Techniques 8, 1097–1109. https://doi.org/10.5194/amt-8-1097-2015. [2] Ruthenburg, T.C., Perlin, P.C., Liu, V., McDade, C.E., Dillner, A.M., 2014. Determination of organic matter and organic matter to organic carbon ratios by infrared spectroscopy with application to selected sites in the IMPROVE network. Atmospheric Environment 86, 47–57. https://doi.org/10.1016/j.atmosenv.2013.12.034. [3] Takahama, S., Ruggeri, G., Dillner, A.M., 2016. Analysis of functional groups in atmospheric aerosols by infrared spectroscopy: sparse methods for statistical selection of relevant absorption bands. Atmospheric Measurement Techniques 9, 3429–3454. https://doi.org/10.5194/amt-9-3429-2016.