Passive Aerosol Sampler. Part II: Wind Tunnel Experiments

Abstract

Wind tunnel experiments have been performed on a passive aerosol sampler. The sampler estimates average concentrations and size distributions using a deposition velocity model and the measured particle flux to the sampler. The small-scale wind tunnel incorporated a high-output aerosol generator that produced nonvolatile, polydisperse particles. An eight-stage impactor was connected to the tunnel with an isoaxial, isokinetic probe and was equipped with polycarbonate-membrane substrates saturated with oleic acid to minimize particle bounce. Before performing experiments, the tunnel's test section was characterized. Aerosol concentrations were determined to have a CV < 6%. The friction velocity, an index of turbulence, was found to range from 0.09 to 0.25 m/s for wind speeds of 1.5 to 5 m/s. The empirical portion of the deposition velocity model, γm, was determined as a function of particle size by minimizing the sum-of-squares difference between impactor and passive sampler across all size bins and all experiments. The relatively simple correlation is a function of the particle Reynolds number only. Precision was assessed by running three passive samplers simultaneously in each experiment. The tests yielded CV PM2.5 = 18.1% and CV PM10 = 32.2%. ANOVA tests were conducted on accuracy and precision to see whether they depended on wind speed, relative humidity, or aerosol concentration, and accuracy was tested with respect to particle size. No significant trends were observed. Sensitivity analysis showed that the volume shape factor is the most important of the mass and shape conversion factors. If SEM is used, the passive sampler will exhibit some error when sampling volatile aerosols. Because concentrations fluctuate over time, long-term exposures measured by the passive sampler should be more accurate than conventional averages based on short-term samples.