Search

Organic carbon (OC) is a major component of atmospheric particulate matter (PM). Typically OC concentrations are measured using thermal methods such as thermal-optical reflectance (TOR) from samples collected on quartz filters. However, TOR measurements are destructive and expensive. We estimate TOR OC concentrations using Fourier transform infrared (FT-IR) spectra of ambient samples collected on Teflon filter. We have developed a sparse statistical calibration model (ElnetPLS), which excludes unnecessary wavenumbers from infrared spectra during the model building process, permitting the identification and evaluation of the most relevant vibrational modes of molecules in complex aerosol mixtures associated with reported TOR OC concentrations. The sparsest ElnetPLS model has similar model performances of the full (2784) wavenumber models while requiring only ten wavenumbers associated with carbonyl groups.

Atmospheric particulate matter (PM) has been associated with increased morbidity and mortality, reduced visibility, and is one of the least understood components of the climate system. We use the Fourier transform infrared (FT-IR) absorbance spectra of atmospheric aerosol (PM2.5) collected on Teflon filters to characterize the aerosol chemical composition using the functional group (FG) representation. Teflon filters have been collected daily at the National Air Pollution Monitoring Network (NABEL) station in Zurich (Switzerland) from the 1st of April 2016 until the 31st of March 2017. We quantify alcohol COH, carboxylic COOH, alkane CH, carbonyl CO, and amine NH functional groups of the ambient samples by fitting individual Gaussian line shapes to spectra. We compare our analysis with collocated black carbon measurements (BC) that have been apportioned to traffic emission and wood burning.

Organic carbon and elemental carbon are major components of atmospheric PM. Typically they are measured using destructive and relatively expensive methods (e.g., TOR). We aim to reduce the operating costs of large air quality monitoring networks using FT-IR spectra of ambient PTFE filters and PLS regression. We achieve accurate predictions for models (calibrated in 2011) that use samples collected at the same or different sites of the calibration data set and in a different year (2013).

Organic carbon (OC) and elemental carbon (EC) are major components of atmospheric PM. Typically OC and EC concentrations are measured using thermal optical methods from samples collected on quartz filters. However, these measurements are destructive and relatively expensive. In this work, we estimate TOR OC and EC using Fourier transform infrared (FT-IR) spectra from Teflon filters. Dillner and Takahama (2015a, b) showed the feasibility of accurate prediction developing calibration models for OC and EC using a subset of 2011 IMPROVE monitoring network samples. We evaluate the capability for the same calibration models to predict OC and EC in ambient samples collected in 2013, which includes the same six sites and 11 additional sites used in 2011. We also propose a statistical method by which we can compare the spectral similarity of new samples to those in the calibration set, and anticipate the magnitude of prediction errors prior to application of the calibration model.

Atmospheric particulate matter (PM) has been associated with increased morbidity and mortality, reduced visibility, and is one of the least understood components of the climate system. Organic matter (OM) constitutes a substantial fraction (up to 80%) of PM (Lim and Turpin 2002), and its quantification and characterization can help in understanding the impact of atmospheric aerosol on health and climate. OM is a complex mixture of thousands of different organic molecules that vary in structure and physicochemical properties (Hamilton et al. 2004). The functional group (FG) representation simplifies the chemical analysis of OM by providing information about properties like volatility and hygroscopicity and can be used for overall organic matter and organic carbon (OC) quantifications, and its apportionment by source class (Russell et al., 2011). We use the Fourier transform infrared (FT-IR) absorbance spectra of atmospheric aerosol (PM2.5) collected on Teflon filters to characterize the chemical composition of OM using the FG representation. Teflon filters are collected daily at the National Air Pollution Monitoring Network (NABEL) station in Zurich (Switzerland) from the 1st of April 2016 (until the 31st of March 2017). We processed the spectrum of each sample to correct the drift of the baseline and substrate interference using the method proposed by Kuzmiakova et al. (2016). We quantify the functional group composition of the ambient samples by fitting individual Gaussian line shapes (top panel in Fig. 1). We quantified alcohol COH, carboxylic COH, alkane CH, carbonyl CO, and amine NH functional groups, as described in Takahama et al. (2013). We use the FT-IR spectra to apportion OM and FGs associated with traffic emission and wood burning using collocated measurements of black carbon (BC – light absorption of PM2.5 measured at multiple wavelengths). An example of FG study is shown in Fig. 1 for a sample collected in November 2016. The FG distribution shows that alkane CH accounts the 39% of the total OM. The significant contributions of alcohol and carboxylic acid (16% and 29 respectively) exemplify the influence of processed aerosol from surrounding regions affecting the PM2.5 in Zurich. Moreover, the high OM/OC ratio (2.01) with substantial contributions from alcohol, and carbonyl FGs are consistent with those found in biogenic or wood-burning samples (Russell et al., 2011). These sources have also been reported in previous studies of Zurich PM by carbon isotope analysis Szidat et al. (2004) and aerosol mass spectrometry (Canonaco et al., 2015).

Organic carbon (OC) is a major component of atmospheric particulate matter (PM). Typically OC concentrations are measured using thermal methods such as thermal-optical reflectance (TOR) from samples collected on quartz filters. However, TOR measurements are destructive and expensive. We estimate TOR OC concentrations using Fourier transform infrared (FT-IR) spectra of ambient samples collected on Teflon filter. We have developed a sparse statistical calibration model (ElnetPLS), which excludes unnecessary wavenumbers from infrared spectra during the model building process, permitting the identification and evaluation of the most relevant vibrational modes of molecules in complex aerosol mixtures associated with reported TOR OC concentrations. The sparsest ElnetPLS model has similar model performances of the full (2784) wavenumber models while requiring only ten wavenumbers associated with carbonyl groups.