Performance of Passive Samplers Analyzed by Computer-Controlled Scanning Electron Microscopy to Measure PM10-2.5

We report on the precision and accuracy of measuring PM10-2.5 and its components with particles collected by passive aerosol samplers and analyzed by computer-controlled scanning electron microscopy with energy dispersive X-ray spectroscopy. Passive samplers were deployed for week-long intervals in triplicate and colocated with a federal reference method sampler at three sites and for 5 weeks in summer 2009 and 5 weeks in winter 2010 in Cleveland, OH. The limit of detection of the passive method for PM10-2.5 determined from blank analysis was 2.8 μg m(-3). Overall precision expressed as root-mean-square coefficient of variation (CVRMS) improved with increasing concentrations (37% for all samples, n = 30; 19% for PM10-2.5 > 10 μg m(-3), n = 9; and 10% for PM10-2.5 > 15 μg m(-3), n = 4). The linear regression of PM10-2.5 measured passively on that measured with the reference sampler exhibited an intercept not statistically different than zero (p = 0.46) and a slope not statistically different from unity (p = 0.92). Triplicates with high CVs (CV > 40%, n = 5) were attributed to low particle counts (and mass concentrations), spurious counts attributed to salt particles, and Al-rich particles. This work provides important quantitative observations that can help guide future development and use of passive samplers for measuring atmospheric particulate matter.