Performance evaluation of a fiber length classifier

A performance evaluation was conducted on a differential mobility classifie r that separates fibers according to length using dielectrophoresis. The cl assifier had been constructed and used for several applications in previous studies. The performance of the classifier was predicted using a two-dimen sional axisymmetric model of the flow field and then calculating particle t rajectories for a variety of conditions. Based on the flow calculations, se veral regions of the classifier mere improved to reduce likelihood of turbu lent losses. For a given total flow through the classifier and a maximum vo ltage across the electrodes, the performance of the classifier was found to depend on the ratios of the aerosol flow to the inner and the outer sheath flows. It was found that the minimum classifiable length, the minimum leng th distribution width, and the throughput of classified fibers can each be optimized, but not independently, Several approaches to testing the resolut ion of the classifier were tried The first was to measure the length distri bution of fibers passing through the classifier under different conditions using electron microscopy, However, this was a slow and imprecise measure o f performance, Two approaches using monodisperse latex spheres were used; o ne operated the instrument as an electrical mobility (electrophoresis) anal yzer and the other evaluated only the flow system accuracy, All measures in dicate that the classifier operates close to theoretical performance, but i mprovements are still possible. Suggested improvements require redesign of the flow system and improved electrode alignment.