Long-Term Strategy for Assessing Carbonaceous Particulate Matter Concentrations from Multiple Fourier Transform Infrared (FT-IR) Instruments: Influence of Spectral Dissimilarities on Multivariate Calibration Performance

Abstract

Matching the spectral response between multiple spectrometers is a mandatory procedure when developing robust calibrations whose prediction is independent of instrument-related signal variations. A viable alternative to complex calibration transfer methods consists of matching the instrument spectral response by controlling a set of key instrumental and environmental parameters. This paper discusses the applicability of such an approach to three Fourier transform infrared (FT-IR) spectrometers used for the routine assessment of carbonaceous particulate matter concentrations in the Interagency Monitoring of PROtected Visual Environments (IMPROVE) speciation network. The effectiveness of the proposed matching procedure is evaluated by comparing the spectral response for each individual instrument in order to characterize the extent, and nature, of the remaining inter-instrument spectral dissimilarities. Instrument-related contributions to the signal were determined to be small compared with the spectral variability induced by the filter type used for sample collection. The impact of spectral differences on prediction was addressed through the comparison of model performance derived from multiple calibration scenarios. A hybrid model yielding accurate and homogeneous prediction regardless of the instrument was proposed for organic carbon (OC) and elemental carbon (EC), two major constituents of atmospheric particulate matter. Coefficients of determination of 0.98 (OC) and 0.90 (EC) with median biases not exceeding 0.20 µg (OC) and 0.07 µg (EC) are reported. The long-term stability, assessed from weekly measurements of reference samples, shows a deviation in predicted concentrations of less than ±5% over a 2.5-year period for most of the data collected. Extending OC and EC hybrid models to the prediction of ambient samples collected during the two subsequent years provides satisfactory performance. The proposed instrument matching procedure coupled with the relative simplicity of the hybrid model is an alternative to computationally advanced calibration transfer methodologies for the characterization of carbonaceous particulate matter using multiple FT-IR instruments.

Supplementary material to "Quantifying organic matter and functional groups in particulate matter filter samples from the southeastern United States, part I: Methods"

“…The majority of the duplicate FT-IR spectrometry measurement differences were due to slight changes in water vapor and PTFE absorption (in agreement with our previous work documenting typical differences between FT-IR spectra; Debus et al, 2018). Only minimal changes were observed for duplicate FT-IR analyses: 100% of duplicate OM concentration and ≥95% of duplicate functional group concentration biases were within sampling uncertainty (although 86% for naCO).…”

Quantifying organic matter and functional groups in particulate matter filter samples from the southeastern United States – Part 1: Methods

“…A strong, broad C=O stretching band at ∼ 1700-1800 cm −1 is observed in ambient OM spectra (Takahama et al, 2013). In particular, molecules containing carboxylic acids may contribute the majority of OM mass (Decesari et al, 2007), the most abundant of which are typically the C 2 -C 4 dicarboxylic acids (Kawamura and Bikkina, 2016). Six carboxylic acids were included in the calibration models.…”

Atmospheric particulate matter characterization by Fourier Transform Infrared spectroscopy: a review of statistical calibration strategies for carbonaceous aerosol quantification in US measurement networks

“…Quality control filters -blank filters and ambient samples -are analyzed weekly to monitor instrument stability. Debus et al (2018) and that observed changes are small. These data enable us to track instrumental changes that will require re-calibration.…”

“…The quality control data show that the TOR OC and EC measurements obtained from multiple FT-IR instruments in one laboratory are precise, stable (over the 2.5 year period evaluated) and agnostic to instrument used for analysis (Debus et al, 2018).…”

“…The total measurement time for one filter is 5 minutes. Additional details on the FT-IR analysis are described by Ruthenburg et al (2014) and Debus et al (2018).…”