Elemental composition of mineral aerosol generated from Sudan Sahara Sand

Eighteen soil samples from central Sudan were fractionated by dry sieving in a size fraction from < 45 mum to > 300 mum while aerosols generated from these soils were fractionated in the particle size range from 0.25 mum to > 16 mum. The elemental concentrations of soil samples were determined by energy-dispersive X-ray fluorescence, while the elemental concentrations of generated aerosols were analysed by particle-induced X-ray emission. The elements Al, K and Rb show a slight positive fractionation with decreasing particle size throughout the particle size range studied. The concentrations of Ca, Mn, Fe, Sr and Y are maximum in the small soil size fraction (< 45 mum) and decrease for the coarse soil size fractions, while in the mineral aerosol particle sizes (0.25-> 16 mum) the concentrations remain more or less constant. The size distributions for Cr, Ti and Zr show a maximum in the particle size range 45-100 mum and the concentrations of these elements decrease sharply in the aerosol fraction down to 16 mum to remain constant in the smaller aerosol fractions. Enrichment factors for the elements were calculated relative to five reference materials: average crustal rock, average soil, the investigated Sahara bulk soil, the finest fraction of this soil and the aerosol generated from this soil, and using four reference elements: Al, Si, Ti and Fe. The enrichment factors were found to vary significantly depending on the choice of the reference material or the reference element. The enrichment factors for the Sudan mineral aerosol were almost identical to those for Khartoum atmospheric aerosol but different from those for Namib mineral aerosol and Israel atmospheric aerosol following dust storms. Multivariate display methods (cluster analysis, principal component analysis and linear discriminant analysis) were applied to the element ratios in the mineral aerosol from the Sahara and Namib and this showed that these mineral aerosol can be differentiated into different groups. An attempt was also made to relate the mineral aerosol to its parent soil through the use of these multivariate techniques and the elemental ratios in both the mineral aerosols and the bulk soils (Namib and Sahara). It was also possible using the elemental ratios and the multivariate display methods to associate the crustal component to the mineral aerosol generated from the Sahara.