Characterization and source apportionment of airborne particulate matter in Toronto.

摘要

This study was performed to physically and chemically characterize airborne particulate matter in the Toronto area and subsequently apply and develop receptor models. Aerosols were collected on Whatman #41 substrates at four sites in Toronto by high-volume and cascade impactor samplers. Accurate particle-size fractionation into six particle size classes ({dollar}<{dollar}0.5 {dollar}mu{dollar}m and ECD's: 0.5, 0.95, 1.5, 3.0, 7.2 {dollar}mu{dollar}m) was achieved using a Sierra 235 cascade impactor and confirmed via scanning electron microscopy of sample filters. Elemental analysis of collected aerosol was performed using thermal neutron activation (28 elements) and photonuclear activation (6 elements). Interelement correlation and factor analysis of unfractionated ambient aerosol composition data from a suburban site qualitatively identified six contributing sources of aerosol: viz. crustal dust, vehicle emissions, incineration, road salt, glass manufacturing and oil combustion. Chemical mass balance results, during the sample period, indicate urban dust constituted 69% of the ambient airborne particulate matter. Road salt (12.6%), automotive emissions (8.1%), incineration emissions (5.9%), glass manufacturing (2.9%) and oil combustion (1.2%) contributed lesser amounts. A new receptor model, the Size-Specific Elemental Mass Balance (SSEMB), was developed to incorporate particle-size information and hence allow accurate apportionment in complex industrial airsheds having multi-collinear sources. Following qualitative factor analysis of both fractionated and unfractionated data for an East Riverdale site, five major sources of inorganic aerosol were identified and their contribution to ambient inorganic suspended particulate matter subsequently quantified by SSEMB: viz. urban dust (74%), non-ferrous industry (16.2%), automotive exhaust (6.4%), de-icing salt (2.4%), and incineration (0.4%). The SSEMB enabled further separation of non-ferrous sources into casting and refining stack emissions, casting fume and fugitive process dust of which the latter alone contributed up to 53% of the total ambient airborne lead over the sample period. Only 21% of the airborne lead was attributable to vehicle exhaust and 1% to incinerator emissions. The average larger/smaller (L/S) ratio of predicted and observed concentrations for the SSEMB was 2.2. However, application of a conventional chemical mass balance model to the East Riverdale data yielded an average L/S ratio of 3.1. Predicted particle size distributions show that the CMB provides a poorer fit than the SSEMB, resulting in gross under-prediction, and inaccurate apportionment, of many of the elements having collinear sources.