PM 10 metal concentrations and source identification using positive matrix factorization and wind sectoring in a French industrial zone Laurent Y. Alleman ⁎, Laure Lamaison 1 , Esperanza Perdrix, Antoine Robache 2 , Jean-Claude Galloo Université Lille Nord de France, F-59000, Lille, France Ecole des Mines de Douai, Département Chimie et Environnement, 941 rue Charles Bourseul, BP 10838, 59508 Douai cedex, France article info abstract Article history: Received 24 July 2009 Received in revised form 8 February 2010 Accepted 24 February 2010 The elemental composition data of ambient aerosols collected upon selected wind sectors in the highly industrialised harbour of Dunkirk (France) were interpreted using pollution roses, elemental ratios, Enrichment Factors (EF), Principal Component Analysis (PCA) and Positive Matrix Factorization (PMF) receptor model. The objective was to identify the possible sources of PM10 aerosols, their respective chemical tracers and to determine their relative contribution at the sampling site. PM10 particles samples were collected from June 2003 to March 2005 in order to analyse up to 35 elements (Ag, Al, As, Ba, Bi, Ca, Cd, Ce, Co, Cr, Cs, Cu, Eu, Fe, K, La, Mg, Mn, Mo, Na, Ni, Pb, Rb, S, Sb, Sc, Si, Sm, Sr, Th, Ti, U, V, Zn and Zr) using Inductively Coupled Plasma (ICP)-Atomic Emission Spectrometry (AES) and ICP-Mass Spectrometry (MS). A significant effort has been made on estimating the total uncertainty of each result by regularly analysing blanks, quality controls and SRM NIST standards. Based on this procedure, a selected set of 24 "robust" elements was compared to the 35-element matrix in order to evaluate the sturdiness of our PMF statistical treatment. Eight source factors were resolved by PCA for all the wind sectors explaining 90% of the total data variance. The PMF results confirmed that eight physically interpretable factors contributed to the ambient particulate pollution at the sampling site: crustal dust (11%), marine aerosols (12%), petrochemistry activities (9.2%), metallurgical sintering plant (8.6%), metallurgical coke plant (12.6%), ferromanganese plant (6.6%), road transport (15%) and a less clearly interpretable profile probably associated to dust resuspension (13%). These weighted contributions against wind direction frequencies demonstrate that industrial sources are the most important contributors to this site (37%) followed by the natural sources (detrital and marine sources) (23%) and the road transport (15%). © 2010 Elsevier B.V. All rights reserved. Keywords: Trace element PM 10 Positive matrix factorization Source-receptor model ICP-MS 1. Introduction Aerosols originate from a variety of natural (soil dust erosion, sea salt, volcanism, natural forest fires …) and anthropogenic (industries, transports, biomass burning, combustion of fossil fuels …) sources. Atmospheric particles with an aerodynamic diameter smaller than 10 μm (PM 10 ) have been put under scrutiny in the past, being easily inhaled and deposited within the respiratory system (Pope et al., 1995). Studies show that PM 10 play a role in the incidence and severity of respiratory diseases (Pope and Dockery, 1999; Brunekreef and Holgate, 2002) and have significant associa- tions with decline in lung function and cardio-vascular pathologies. As these particles may be harmful to humans, their assessment level and chemical composition are signifi- cant from an environmental health perspective. In addition, Atmospheric Research 96 (2010) 612–625 ⁎ Corresponding author. Ecole des Mines de Douai, Département Chimie et Environnement, 941 rue Charles Bourseul, BP 10838, 59508 Douai cedex, France. E-mail address: alleman@ensm-douai.fr (L.Y. Alleman). 1 Present address: Laboratoire Traces Inorganiques, CARSO-LSEHL, 321, avenue Jean Jaurès, 69007, Lyon, France. 2 Present address: Direction Régionale de l'Industrie, de la Recherche et de l'Environnement DRIRE Bourgogne, Dijon, France. 0169-8095/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.atmosres.2010.02.008 Contents lists available at ScienceDirect Atmospheric Research journal homepage: www.elsevier.com/locate/atmos