Evaluation of Ambient Particulate Matter (PM) Sampler Performance Through Wind Tunnel Testing

Previous studies have demonstrated that EPA approved federal reference method (FRM) samplers can substantially misrepresent the fractions of particles being emitted from agricultural operations due to the relationship between the performance characteristics of these samplers and existing ambient conditions. Controlled testing in a wind tunnel is needed to obtain a clearer understanding and quantification of the performance shifts of these samplers under varying aerosol concentrations, wind speeds and dust types. In this study, sampler performance was tested in a controlled environment wind tunnel meeting EPA requirements for particulate matter (PM) sampler evaluation. The samplers evaluated included two low-volume PM10 and Total Suspended Particulate (TSP) pre-separators. The masses and particle size distributions (PSDs) obtained from the filters of tested samplers were compared to those of a collocated isokinetic sampler. Sampler performance was documented using two parameters: cut-point (d_50) and slope. The cut-point is the particle diameter corresponding to 50% collection efficiency of the pre-separator while the slope is the ratio of particle sizes corresponding to cumulative collection efficiencies of 84.1% and 50% (d_84.1/d_50) or 50% and 15.9% (d_50/d_15.9) or the square root of 84.1% and 15.9% (d_84.1/d_ 15.9). The test variables included three levels of wind speeds (2-, 8-, and 24-km/h), five aerosol concentrations varying from 150 to 1,500 mu g/m3 and three aerosols with different PSDs (ultrafine Arizona Road Dust (ARD), fine ARD and cornstarch). No differences were detected between the performance of the flat and louvered FRM PM10 samplers (a = 0.05). The mean cut-point of both the PM10 samplers was 12.23 mu m while the mean slope was 2.46. The mean cut-point and slope values were statistically different from the upper limit of EPA-specified performance criteria of 10.5 mu m for the cut-point and 1.6 for the slope. The PM10 samplers over-sampled cornstarch but under-sampled ultrafine and fine ARD. The performance of the dome-top TSP sampler was close to the isokinetic sampler, and thus it can be used as a reference sampler in field sampling campaigns to determine true PM concentrations. There were large variations in the performance of the cone-top TSP samplers as compared to the isokinetic sampler. Dust type and wind speed along with their interaction had an impact on sampler performance. Cut-points of PM10 samplers were found to increase with increasing wind speeds. Aerosol concentration did not impact the cut-points and slopes of the tested samplers even though their interaction with dust types and wind speeds had an impact on sampler performance.