Determination of organic matter and organic matter to organic carbon ratios by infrared spectroscopy with application to selected sites in the IMPROVE network

Mass of ambient particulate organic matter (OM) is often estimated by multiplying the organic carbon (OC) mass by a fixed factor that typically ranges from 1.4 to 1.8. In this paper, we develop a nondestructive, mid-infrared spectroscopic (MIR) technique to measure OM in PM2.5 collected on PTFE filters (commonly called teflon filters). MIR techniques measure absorption by functional groups within organic molecules; that is, carbon atoms bonded to oxygen (O) or hydrogen (H), O bonded to H and other elements bonded together in organic molecules that comprise OM. We developed laboratory standards of atmospherically relevant organic compounds as the basis for calibrating the MIR absorption to the moles of functional groups. A multivariate regression technique was used to develop calibrations for quantifying alkane CH, alcohol OH, carboxylic acid OH, and carbonyl, which likely comprise the bulk of OM in most ambient samples. OM is estimated as the sum of masses attributed to these functional groups. The precision of the OM measurement is 6.9 mu g (relative precision is 7%) and the minimum detection limit is 4.8 mu g, corresponding to an ambient concentrations of 0.21 mu g m(-3) and 0.15 mu g m(-3) respectively, for the sampling protocol used here. The method was employed to estimate OM, OC (estimated as the sum of the mass of carbon in the alkane CH and carbonyl functional groups) and OM/OC from one year of routinely collected filters at seven Interagency Monitoring of Protected Visual Environments (IMPROVE) sites. In the IMPROVE network, OC is measured using a thermal optical method and OM is estimated for regional haze tracking simply as 1.8 x DC. Using the MIR technique on one year of samples from 7 IMPROVE sites, the median OM/OC ratio was 1.69 with the 10th and 90th percentiles of 1.46 and 2.01. Phoenix, AZ had the lowest annual-median value of 1.56 and Mesa Verde, CO had the highest of 1.83. For more than half of the sites, the lowest monthly-median OM/OC ratio occurred in winter. The sample, site median and seasonal median OM/OC ratios obtained from MIR analysis indicate that a single estimate of OM/OC does not represent the observed variability in OM/OC. (C) 2014 Elsevier Ltd. All rights reserved.