Field testing a low-cost passive aerosol sampler for long-term measurement of ambient PM2.5 concentrations and particle composition

Abstract While low-cost sensors can be valuable in assessing patterns of exposure to PM2.5 in cities, few sensors are optimal for monitoring long-term averages, the metric most relevant to long-term mortality risk, and none can readily determine particle composition. This study was designed to field-test a low-cost passive sampler (University of North Carolina Passive Aerosol Sampler, UNC-PAS) to measure long-term mass and elemental concentrations of PM2.5 in urban areas. We collected sequential 4-week and 12-week integrated samples over 1 year at eight sites in Boston, New York City and the San Francisco Bay Area with co-located federal reference and/or equivalent PM2.5 samplers. Field blanks and duplicates were included to assess sampling artifacts and precision, respectively. We analyzed filters for PM2.5 concentrations and elemental composition by site and by city using computer-controlled scanning electron microscopy coupled with energy dispersive spectroscopy, combined with theoretical sampling rates based on particle diameter. To improve UNC-PAS performance for submicron, carbonaceous particles typical of urban environments, we developed and evaluated refinements to standard particle analysis methods and deposition calculations. The analysis improvements yielded better performance for samples heavily impacted by wildfires, while the calculation improvements yielded better agreement with continuous and filter data across all sites. We observed good precision, with a single sample error coefficient of variation of 14.2%. Compared with co-located filter and continuous methods, the passive samplers were, on average, accurate to within 13.6% and −1.0%, respectively. Microscopic elemental and morphological analysis of individual particles showed contributions from vehicle combustion, salts, aged salt-carbon mixtures, biogenic spores, fly ash, and wildfire smoke. We conclude that the UNC-PAS can be a useful, complementary approach for long-term community monitoring, with the potential to improve spatial coverage and provide valuable information on composition and sources.