Calibration Studies of the Drum Impactor*

A b s t r a e t ~ l i b r a t i o n studies were performed on the Davis Rotating-drum Universal-size-cut Monitoring (DRUM) sampler utilizing fluorescent monodisperse aerosols of polystyrene latex. The DRUM sampler is an eight-stage Lundgren-type rotating drum cascade impactor utilizing a series of single round jets for each stage including low-pressure stages. The DRUM sampler operates continuously for up to 2 weeks with particle collection on slowly rotating, greased substrate-coated cylindrical drums. The resulting narrow linear deposit for each stage can be evaluated with time resolution as short as 4 h utilizing Particle Induced X-ray Emission (PIXE) spectroscopy and other analytical methods. The DRUM sampler is particularly useful for evaluating environmental aerosols for size-related chemical properties. Aerosols of monodisperse fluorescent polystyrene latex spheres were generated by nebulization from dilute aqueous suspensions, dried with a diffusion drier, and reduced in electrostatic charge with a krypton-85 discharger. The aerosols were characterized by electron microscopy and with a Climet208 light scattering particle counter; they were quantified utilizing the fluorescent label. DRUM samples were collected at selected flow rates up to the critical 18.3 cm 3 s1, and the square root of the non-dimensionalized Stokes' number was plotted vs collection efficiency for the impactor stages. The results were in reasonable agreement with the expected effective cut-offaerodynamic diameters based on the non-dimensionalized theory of Marple and with the data of Mercer. All the stages demonstrated dynamic similarity when the Stokes' numbers were calculated at the reduced pressure and temperature at orifice outlets with orifice areas corrected for contraction at sonic and near sonic speeds. The measured cut-offvalues for another low pressure impactor were verified with these results. A wall loss study was conducted using polydisperse respirable aerosols of ammonium fluorescein generated by nebulization with a Lovelace nebulizer, dried with a diffusion drier, and reduced to Boltzmann equilibrium with a krypton-85 discharger. The results show wall losses to be about 10.7 % of the sample.