Two approaches to mass closure analysis for carbon-rich aerosol in Metro Manila, Philippines
Abstract
In this paper, we investigate physico-chemical properties of particulate matter (PM) at an urban mixed site (UB) and two roadside (RS) sites during the 2015 Metro Manila Aerosol Characterization Experiment (MACE). Aerosol particle number size distributions (0.01–10 µm diameter) were measured using a combination of mobility particle size spectrometer and aerodynamic particle size spectrometers. PM2.5 filter samples were analyzed for total mass, organic carbon (OC), elemental carbon (EC), water-soluble inorganic ions, and elemental species. Mass closure between the gravimetric mass, chemical composition, and mass concentration derived from the number size distribution was performed. We found that the bulk PM2.5 mass was dominated by carbonaceous materials, followed by secondary inorganic aerosols and crustal matter at all sites. The average OC/EC ratios at the RS sites (0.16–1.15) suggest that a major fraction of the aerosol mass at these sites derives from traffic sources, while the OC/EC ratio at the UB site (2.92) is indicative of a more aged aerosol. The ultrafine particles (UFPs, diameter < 100 nm) dominated (89–95%) the total particle number concentration at the three sites highlighting the importance of such measurements in this region. However, UFPs have low mass contribution to PM2.5 (7–18%), while particles in the accumulation mode (diameter < 1000 nm) made up for most of the number-derived PM2.5 mass concentration (74–80%). On average, strong agreement between the chemically-derived mass and the gravimetric mass was found (slope =1.02; r2 = 0.94). The number-derived mass concentration correlated well with the gravimetric PM2.5 mass (slope = 1.06; r2 = 0.81). Mass closure between the gravimetric mass and the number-derived mass presented in this work provides a framework for a conversion between number size distributions and PM2.5 mass concentration in real time in an environment with similar characteristics.