Source characterization and identification by real-time single particle mass spectrometry

A Real-Time Single Particle Mass Spectrometer, RSMS-3, was deployed to Wilmington, Delaware to study regional and local contributions to fine and ultra-fine urban particulate matter (PM). Approximately two-thirds of PM₁ consisted of internally mixed secondary aerosol. The remaining one-third was externally mixed including biomass burning (13%), fossil fuel combustion (7%) and various industrial sources (13%). In this last group, particle classes containing specific combinations of transition and/or heavy metals gave wind-rose plots consistent with specific point sources. For example, particles containing V and Ni were detected from different wind directions than those containing V and Fe. Samples from two industrial emission stacks, a steel manufacturing facility 10 km away and a coal-fired electrical power generation facility 5 km away, were analyzed and compared to the ambient data set. In each case, a direct correlation was found: a Pb–Zn–K–Na class for the steel manufacturing facility and an Fe–La/Ce class for the power generation facility. The ambient particle classes showed additional small signals from secondary components indicating atmospheric processing. Ambient particle classes containing only a subset of these elements, such as Zn only, Fe only and Pb–K only, were nonspecific, that is, the wind-rose plots were more diffuse and the particles could not be mapped to individual sources. The merits of stack sampling as an aid to interpreting single particle data sets are discussed.