Measurements of Fine Particle Mass Concentrations Using Continuous and Integrated Monitors in Eastern US Cities

PM 2.5 mass concentrations were measured using multiple continuous samplers (a continuous ambient mass monitor (CAMM), a real-time ambient mass sampler (RAMS), and a drier-equipped tapered element oscillating microbalance (TEOM)) as well as a federal reference method (FRM) filter sampler in three eastern US cities. The effects of measurement systems, ambient conditions, and composition on the measured PM 2.5 mass concentrations were examined at each site. Different responses to PM 2.5 mass were observed by the continuous samplers with a variation of 10-30% in the average PM mass concentrations. The impacts of semivolatile particulate ammonium nitrate on PM 2.5 mass concentration was apparent for those time periods when high nitrate concentration >1.62 g m −3 was measured in Atlanta. Contribution of semivolatile particulate ammonium nitrate to the ambient PM 2.5 mass appeared more significant than semivolatile organic material in Atlanta. The effects of the ambient water were not negligible and were significant in Baltimore. The total ambient PM 2.5 mass by the CAMM was underestimated in all locations compared to the other continuous PM mass monitors, presumably due to the lack of equipments that collect semivolatile particulate ammonium nitrate and organic materials. The results suggest that short time intervals do not guarantee the measurement of total ambient PM 2.5 mass including semivolatile materials. In the comparison of each of the continuous versus integrated PM 2.5 mass over the sites, the drier-equipped TEOM presented the best agreement, with the FRM having r 2 values >0.90 and slopes close to 1 regardless of location and time period, suggesting that the same losses of semivolatile materials as seen in the FRM can occur in the sampling by the drier-equipped TEOM. The regressions between the CAMM or the RAMS and the FRM showed the variance of their relation, by site and time period, from 0.6 to 0.9 for r 2 and from 0.7 to 1.0 for the slope.