1. F. Lucarelli 1, S. Becagli 2, G. Calzolai 1, M. Chiari 1, T. Martellini 2, S. Nava 1, L. Paperetti 1, R. Udisti 2 and E. Yubero 3 1. INFN and Dept. of Physics (University of Florence), Italy, 2. Dept. of Chemistry (University of Florence), Italy, 3. LCA - Miguel Hernández University, Spain lucarelli@fi.infn.it Since aerosol particles retain elemental compositions characteristic of their origin, the simultaneous detection of groups of elements by multi-elemental techniques, like PIXE, can be of great help in the study of aerosol sources by receptor models, like Positive Matrix Factorisation (PMF). Many studies have been devoted to the application of these models to 24-h data; however, since the impact of many sources can vary on a time scale of few hours or less, the use of hourly concentration data sets can be of great help. In the framework of the PATOS project, the first extensive field campaign for the aerosol characterisation in Tuscany (Italy), the application of PMF to daily and hourly concentrations in the most polluted of the sampling sites (Lucca-Capannori) allowed to identify the main aerosol sources and to estimate their impact. INTRODUCTION SOURCE APPORTIONMENT BY PMF ON DAILY SAMPLES PMF has been applied to daily samples collected in Capannori (Lucca) during the fist 9 months (Sept05-Jun06): 6 sources have been identified and PM10 source apportionment calculated. SOURCE IDENTIFICATION BY ONE-HOUR RESOLUTION SAMPLES Average source apportionment Due to the important contribution of biomass burning, we further investigate this source: Since K is also present in the crustal composition, we calculated its enrichment factor and looked at its level of solubility (comparing concentrations obtained by PIXE and IC). We found very high enrichment values with respect to the crustal composition and we also found that most of this element is soluble, thus confirming the burning origin. Looking at the concentrations of the elements/compounds characteristic of this source in the other sampling sites we found that it is a local source (see for example the graph of K concentration on the right) further information came from the analysis of hourly concentrations (see below) PIXE analysis of streaker samples X FC SM The beam size corresponds to one hour of aerosol sampling: analysing the streak point by point elemental concentrations are measured p PM2.5PM2.5-10 Traffic (fine+coarse fractions)Biomass burning (fine fraction) Periodic pattern with peaks starting at about 19:00 and lasting ~10 h: suggesting wood burning for domestic heating Peaks during traffic rush hours Appling PMF to the streaker data we identified the following sources: Fine fraction: soil dust, traffic, sea salt Coarse fraction: secondary aerosols, biomass burning, traffic, soil dust. THE PATOS PROJECT  six sites in Tuscany, representative of areas of different typology Prato Arezzo Lucca Firenze Livorno Grosseto  PM10 daily samples collected for one year (Spt.05-Spt.06) by CEN equivalent samplers, simultaneously on Teflon and Quartz fiber filters thus allowing the application of different analytical techniques: PIXE  elemental concentrations for Z>10 IC  inorganic ions soluble component ICPMS  metals soluble component CHNTA  total C (TC) GC and GC-MS  n-alkanes and PAHs  hourly samples (fine and coarse fractions) collected for shorter periods, by a streaker sampler PIXE  elemental concentrations for Z>10 Lucca-Capannori resulted the most pollutted area with very high PM10 concentrations during winter Aerosol source apportionment by Positive Matrix Factorisation applied to daily and hourly concentration datasets obtained by PIXE Correlation coefficients (sources - elements/compounds) Day by day source apportionment  The source ‘Biomass burning’, characterised by high loadings of K, Nitrate, Glycolate and TC, gives the main contribution to PM10 and it is the responsible of the very high PM10 values during winter months (Nov-March).  Important contributions are also given by secondary inorganic aerosols and traffic  The high contribution of soil dust during June has been ascribed to a strong Saharan dust intrusion (see poster PII20) Source profiles (  g/  g)