Practical applications of inverting spectral turbidity data to provide aerosol size distributions

Abstract

Retrieving the distribution of droplet size in a poly-dispersed system from spectral turbidity data is essentially a process of finding a solution to the integral equation involved. Three well-established methods are available: modeling the unknown distribution with an empirical function, an analytical method where the light-scattering kernel is approximated by a simple formula, and the matrix inversion method with smoothing constraint. These methods are reviewed by application to the measured spectral turbidities of poly-dispersed fogs of water droplets formed by spontaneous condensation in supersonic steam flows. On the basis of the examples considered, which include highly skewed monomodal and bimodal distributions, it was concluded that the matrix inversion method offered the best chance of obtaining solutions in the absence of any a priori information.