Source identification of inorganic airborne particle fraction (PM10) at ultratrace levels by means of INAA short irradiation

Many studies have focused their attention on the determination of elements of toxicological and environmental interest in atmospheric particulate matter using analytical techniques requiring chemical treatments. The instrumental nuclear activation analysis technique allows achieving high sensitivity, good precision, and excellent limit of detection without pretreatment, also considering the problems related to the radioisotope characteristics (e.g., half-life time, interfering reactions, spectral interferences). In this paper, elements such as Al, As, Br, Cl, Cu, I, La, Mg, Mn, Na, Sb, Si, Ti, and V are studied in atmospheric PM10 sampled in downtown Rome: The relative radionuclides after activation of the sample are characterized by very short (ranging from 2.24 to 37.2 min) and short (ranging from 2.58 h to 2.70 days) half-lives. Furthermore, As, Br, La, Mn, and Sb were also determined for evaluating the aerosol characteristics. The results, elaborated considering the matrix effects and the interfering reaction contribution to the radioisotope formation (e.g., Al-28 generated by both (n,gamma) reaction from Al-27 and (n,p) reaction from Si-28), show interesting values of As (0.3-6.1 ng m(-3)), Cu (22-313 ng m(-3)), Mn (17-125 ng m(-3)), V (7-63 ng m(-3)), higher than those determined in an area not influenced by autovehicular traffic, and significant levels of I (1-11 ng m(-3)) and Ti (25-659 ng m(-3)) in Rome PM10. The other elements show a pattern similar to the very few data present in the literature. It should be underlined the good correlation (r (2)) of Al vs. Mg (0.915) and Al vs. La (0.726), indicating a same sources for these species as well as Br-Sb showing a little lower correlation (0.623). This last hypothesis is confirmed by the study of the enrichment factors: Sb and Br may be attributed to anthropogenic sources; Cu, Cl, and I show a mixed origin (natural and anthropogenic), whereas Al, Si, Ti, Mn, Na, Mg, and As are of crustal origin. For having more information, a statistical approach based on the principal component analysis and the canonical discriminant analysis has been performed: All the samples (except one) are grouped in a cluster, and elements such as As, Br, Cu, I, La, Mn, Sb, Ti, and V are highly correlated, whereas Na and Cl and Mg and Al assemble in two different clusters. Finally, a comparison with other similar studies is reported showing interesting values for Al, As, Mg, Mn, and Ti.