A Neural Network Model for Estimating the PSD of a Dilute Latex from Multiangle DLS Measurements

GUGLIOTTA, L. M.; STEGMAYER, G. S.; CLEMENTI, L. A.; GONZALEZ, V. D. G.; MINARI, R. J.; LEIZA, J. R.; VEGA, J. R.

PARTICLE & PARTICLE SYSTEMS CHARACTERIZATION

WILEY-V C H VERLAG GMBH

Año: 2009 vol. 26 p. 41 - 52

0934-0866

The particle size distribution (PSD) of dilute latex wasestimated through a general regression neural network(GRNN) that was supplied with PSD average diametersderived from multiangle dynamic light scattering(MDLS) measurements. The GRNN was trained with alarge set of measurements that were simulated from unimodalnormal-logarithmic distributions representingthe PSDs of polystyrene (PS) latexes. The proposedmethod was first tested through three simulated examplesinvolving different PSD shapes, widths, and diameterranges. Then the GRNN was employed to estimatethe PSD of two PS samples; a latex standard ofnarrow PSD and known nominal diameter, and a latexsynthesized in our laboratory. Both samples were alsocharacterized through independent techniques (capillaryhydrodynamic fractionation, transmission electron microscopy,and disc centrifugation). For comparison, allexamples were solved by numerical inversion of MDLSmeasurements through a Tikhonov regularization technique.The PSDs estimated by the GRNN gave moreaccurate results than those obtained through other conventionaltechniques. The proposed method is a simple,effective, and robust tool for characterizing unimodalPSDs.