Study of particulate matter from coke production and its distribution on soils using environmental magnetism, laser spectroscopy and multivariate statistical techniques

GARGIULO J.D., CHAPARRO M.A.E., BERTUCCELLI G., MARINELLI C.

Foz de Iguaçu, Brasil

Congreso; 2010 Meeting os the Americas; 2010

American Geopyisical Union

The coke production can lead to adverse consequences for the human health and the environment. The particulate matter (PM) emission of these factories comprises particles with a variety of grain size, morphology and composition, which may be inhaled by human and/or deposited in soils located up to several kilometers from the pollution source. In this work, we report magnetic and non-magnetic studies of PM emission from a small coke factory located in Tandil (37.28949° S; 59.19646° W, Argentina). The area of study and the dispersion of PM, up to 3 km from the factory, was preliminary studied using an air pollution dispersion model (Gaussian model), which takes into account meteorological data from the area and the emission parameters for this factory. Hence, in situ magnetic measurements (kis) were done in several sites regarding the prevalent wind directions. Furthermore, two samples (at 0 and 20 cm) were collected at each site for measurements in the laboratory. Selected elements, V, Ni, Cr and Zn, were determined using the LIBS technique (laser induced breakdown spectroscopy), showing a similar trend (with two peaks) for the first three metals. On the other hand, fine magnetite-like minerals were detected by rock-magnetic parameters (Hcr, S-ratio, kARM/k-ratio, SIRM/k). Among the magnetic concentration parameters, the mass-specific magnetic susceptibility showed two peak values (about 480 x10-8 m3 / kg) at 420 m and 1530 m from the source. This behavior is in agreement with the distribution of V, Ni and Cr, moreover, it is also in agreement with the PM distribution according to the different four (B, C, D and E) stability classes. In addition, there is a relationship between magnetic and chemical variables that was investigated using multivariate statistical techniques, in particular, Pearson correlation, redundancy analysis, and principal component analysis.