You are here:
INVERTING CASCADE IMPACTOR DATA FOR SIZE-RESOLVED CHARACTERIZATION OF FINE PARTICULATE SOURCE EMISSIONS
Citation:
Dong, Y., M D. Hays**, N D. Smith*, AND J S. Kinsey*. INVERTING CASCADE IMPACTOR DATA FOR SIZE-RESOLVED CHARACTERIZATION OF FINE PARTICULATE SOURCE EMISSIONS. JOURNAL OF AEROSOL SCIENCE. Elsevier Science Ltd, New York, NY, 35(12):1497-1512, (2004).
Description:
Cascade impactors are particularly useful in determining the mass size distributions of particulate and individual chemical species. The impactor raw data must be inverted to reconstruct a continuous particle size distribution. An inversion method using a lognormal function for particle size distribution is developed in this paper for fine particulate (PM2.5) source emissions studies where only nuclei and accumulation modes are of interest. For a cascade impactor of N stages, a set of N non-linear equations is established, in which the unknown parameters of particle size distribution are related to the N known stage mass concentrations, the calibrated stage collection efficiencies, and the measured particle losses in the units. The calculation of kernel functions from the experimentally calibrated stage collection efficiencies is provided, and the computation procedure is illustrated in detail. Application of the inversion method to source emissions characterization studies is demonstrated by inverting two sets of experimental impactor data: the stage particulate from an oil furnace and individually determined chemical species from a residential fireplace burning a wet oak wood fuel. Problems encountered in processing the data inversion are discussed, and solutions to them are offered. The demonstration shows that the developed inversion method is especially useful when source emissions contain a portion of particles that falls outside the measurement range of cascade impactors, or when a mass size distribution of an individual species is determined without the knowledge of the total mass concentration for that species, or when particle losses in the cascade impactor system are significant. Study shows that the stage measurement errors do not amplify the errors in inversion results. Limitations of the inversion method are also discussed.