Key findings from the USEPA coarse PM pilot speciation study

One year of PM_c speciation measurements were conducted in Phoenix and St. Louis using dichotomous samplers and the FRM by difference method. At both sites, PM_c gravimetric mass from the dichots was biased low compared to the FRM by difference method (PM_10-2.5) with dichot-to-FRM mean ratios in the range 0.76-0.90 depending on site and dichot sampler. Data from collocated dichots were used to evaluate measurement precision. For PM_c gravimetric mass at Phoenix the collocated precision was 12% which included a systematic bias of ∼5% between the samplers rather than purely random measurement error. Crustal elements (Al, Ca, Fe, Si, Ti) showed a bias similar to that for gravimetric mass and their measurement errors were highly correlated. PM_c mass closure between the gravimetric mass and mass reconstructed from the species data was met for St. Louis using an OM/OC ratio of 1.4-1.6. In contrast, an OM/OC less than unity, which is physically unrealistic, was required for mass closure on the Phoenix data. Sources of uncertainty are being examined including the multipliers to convert crustal elements to oxides and the multipliers to account for XRF attenuation for light elements.