Effects of aerosol particle size on the measurement of airborne PM2.5 with a low-cost particulate matter sensor (LCPMS) in a laboratory chamber

Abstract

Previous validation studies found a good linear correlation between the low-cost particulate matter sensors (LCPMS) and other research grade particulate matter (PM) monitors. This study aimed to determine if different particle size bins of PM would affect the linear relationship and agreement between the Dylos DC1700 (LCPMS) particle count measurements (converted to PM_2.5 mass concentrations) and the Grimm 11R (research grade instrument) mass concentration measurements. Three size groups of PM_2.5 (mass median aerodynamic diameters (MMAD): < 1 µm, 1-2 µm, and > 2 µm) were generated inside a laboratory chamber, controlled for temperature and relative humidity, by dispersing sodium chloride crystals through a nebulizer. A linear regression comparing 1-min average PM_2.5 particle counts from the Dylos DC1700 (Dylos) to the Grimm 11R (Grimm) mass concentrations was estimated by particle size group. The slope for the linear regression was found to increase as MMAD increased (< 1 µm, 0.75 (R² = 0.95); 1-2 µm, 0.90 (R² = 0.93); and > 2 µm, 1.03 (R² = 0.94). The linear slopes were used to convert Dylos counts to mass concentration, and the agreement between converted Dylos mass and Grimm mass was estimated. The absolute relative error between converted Dylos mass and the Grimm mass was smaller in the < 1 µm group (16%) and 1-2 µm group (16%) compared to the > 2 µm group (32%). Therefore, the bias between converted Dylos mass and Grimm mass varied by size group. Future studies examining particle size bins over a wider range of coarse particles (> 2.5 µm) would provide useful information for accurately converting LCPMS counts to mass concentration.

Keywords: Low-cost sensors; PM sensor calibration; Particle size; Particulate matter (PM).