Source apportionment by multivariate analysis of fine and coarse airborne particulate matter in the Lazio region (Italy)

Assessment of source contributions to the atmospheric concentrations of airborne particulate matter (PM) affecting a defined area holds a key role, both in expanding knowledge of the influence of local/regional sources on the physical - chemical characteristics of collected particles and in providing scientific support to the effectiveness of measures to be taken by authorities towards air quality improvement. Source apportionment by statistical methods comprises the use of factor analysis (FA) and of multi-linear regression analysis (MLRA). The combined application of these methods to data of PM chemical composition helps, indeed, the identification of predominant particle source categories and the apportionment of their relative contributions to the collected PM. To this purpose, the analytical procedure to be used for the chemical characterization of PM samples has to provide the most possible complete information on particles composition, so that concentration data of both major and trace PM components can be included in the FA-MLRA analysis. Within the “Polveri Fini” (Fine Dust) project, aimed to investigate the contributions to PM in the Lazio region (Italy) from local sources and medium-long range transport, 24-h PM10 and PM2.5 samples were daily collected on quartz and Teflon filters over 1 year (Oct. 2004-July 2005). PM samples were characterized for chemical composition by a proper analytical scheme. By this scheme PM samples collected on quartz are analysed for the elemental and organic carbon content (EC/OC) by a thermal-optical analyser; PM samples collected on Teflon are first analysed by ED-XRF for major elements and then chemically leached into an extract and a mineralised residual fractions by a two-step leaching procedure, previously developed and validated by our lab. In the extract solution ions and soluble elements are determined by IC and ICP-AES, and the residual elements solution is also analysed by ICP-AES. A set of about 1000 PM10 and PM2.5 samples and 25 chemical components was investigated by FA-MLRA. Results about the chemical characterisation of the collected samples are discussed together with the main important sources affecting atmospheric particles in the Lazio region; these have been identified as crustal components, sea-salt, anthropogenic combustion, secondary compounds. For each main source category, the back-trajectories of the air masses have been determined and statistically evaluated