Analysis of functional groups in atmospheric aerosols by infrared spectroscopy: method development for probabilistic modeling of organic carbon and organic matter concentrations

Abstract

Abstract. The Fourier transform infrared (FTIR) spectra of fine particulate matter (PM2.5) contain many important absorption bands relevant for characterizing organic matter (OM) and obtaining organic matter to organic carbon (OM/OC) ratios. However, extracting this information quantitatively – accounting for overlapping absorption bands and relating absorption to molar abundance – poses several challenges. For instance, a subset of model parameters lead to calibrations that test almost indistinguishably well against laboratory standards generate substantially different predictions in ambient samples. Furthermore, additional parameters related to molecular structure are required to estimate carbon content from functional group (FG) abundance. However, since many carbon atoms can be branched (not fully functionalized) or polyfunctional, these parameters are not well constrained for ambient sample mixtures. In this work, we present a probabilistic framework to characterize combinations of these parameters that are consistent with field measurements of organic carbon (OC), for which estimates from thermal optical reflectance (TOR) measurements are used. Uncertainties in this probabilistic framework characterize the plausibility of many different parameter values that yield acceptable predictions (to the extent that they can be evaluated) neglected in conventional estimates of statistical uncertainties. Based on calibrations of aliphatic CH, alcohol COH, carboxylic acid COO, carboxylate COO, and amine NH2, we find model parameters for approximately homogeneous groups of samples determined from cluster analysis of FTIR spectra available for 17 sites in the Interagency Monitoring of Protected Visual Environments (IMPROVE) monitoring network (7 sites in 2011 and 10 additional sites in 2013). These groups are interpreted as being predominantly influenced by dust, residential wood burning, wildfire, urban sources, and biogenic aerosols. The resulting calibrations reproduce TOR OC concentrations (R2 = 0.96) and provide OM/OC values consistent with our current best estimate of ambient OC. The mean OM/OC ratios corresponding to sample types determined from cluster analysis range between 1.4 and 2.0, though ratios for individual samples exhibit a larger range. Trends in OM/OC for sites aggregated by region or year are compared with another regression approach for estimating OM/OC ratios from a mass balance of the major chemical species contributing to PM fine mass. Differences in OM/OC estimates are observed according to estimation method and are explained through the sample types determined from spectral profiles of the PM.

Estimating mean molecular weight, carbon number, and OM/OC with mid-infrared spectroscopy in organic particulate matter samples from a monitoring network

“…In contrast, organic aerosols usually undergo several steps of oxidation and receive substantial condensation of oxidized vapors, which results in higher OM/OC ratio at rural and remote sites. Previous studies using several different methods (including FT-IR and AMS) show the same trend in urban and rural sites (Ruthenburg et al, 2014;Zhang et al, 2007;Simon et al, 2011;Bürki et al, 2019). In addition, the majority of the samples are in the range that is usually considered for OM/OC ratio, i.e.,1.4-1.7 (Russell, 2003).…”

“…This observation is consistent with the presence of molecules with longer chains, as observed for laboratory samples. Bürki et al (2019) showed that urban samples (in the same dataset) have their highest average OM/OC ratio in summer which is concurrent with the their highest A 1 /A 2 ratio which suggests shorter chain length. The highest A 1 /A 2 ratio for rural samples is observed in spring when the aerosols are highly oxidized (Bürki et al, 2019).…”

“…Bürki et al (2019) showed that urban samples (in the same dataset) have their highest average OM/OC ratio in summer which is concurrent with the their highest A 1 /A 2 ratio which suggests shorter chain length. The highest A 1 /A 2 ratio for rural samples is observed in spring when the aerosols are highly oxidized (Bürki et al, 2019). This suggests that aged aerosols have lower carbon number probably due to the fragmentation process.…”

“…For this purpose, atmospheric samples are separated in to urban, rural and burning samples. The burning category constitutes 95 samples of urban or rural sites and is taken from clusters 9a, 9b and 10 of Bürki et al (2019) based on their spectral similarity. These samples are believed to be influenced by residential wood burning or wildfires since they were usually taken during a known fire period (Rimfire in California in 2013) or during winter months when residential wood burning typically occurs.…”

“…The OM/OC ratio is the first parameter that we investigate here since it has been studied extensively in atmospheric aerosols (Bürki et al, 2019;Hand et al, 2019;Ruthenburg et al, 2014;Takahama et al, 2011;Simon et al, 2011;Aiken et al, 2008).…”