Distortion of cascade impactor measured size distribution due to bounce and blow-off.

Accurate measurement of particle size distribution by a cascade impactor requires that particles remain on the collection surfaces after impaction. Recent studies show that, when a solid aerosol is sampled, significant rebound and re-entrainment of particles occur in cascade impactors. A mathematical model of a cascade impactor including collection efficiency functions for solid aerosols on various impaction surfaces was developed and used to operate upon a series of log-normal size distributions. The impactor modeled was a six stage Andersen operated at 28.3 L/min. When bare metal or glass impaction surfaces are used, the impactor indicates non-log-normal distributions and the estimated distribution parameters are in error by as much as a factor of 7. When glass fiber impaction surfaces are used, although less serious errors are indicated, the deviation from log-normality is significant for aerosols with MMD larger than 3 micrometers. When oil covered impaction stages are used the indicated distributions parameters are acceptably close to the input parameters. The accuracy of the model was verified experimentally.