Applicability of analytical techniques to quantify trace metals in particulate matter for source apportionment in South Africa

Abstract

The impacts of atmospheric trace metal species in particulate matter (PM) on human health and the environment is well documented. Rapid advances in the development of analytical techniques used for characterisation and quantification of atmospheric trace metals have been made in recent years. Two analytical methods generally used for the analysis of trace metal species include wavelength-dispersive X-ray fluorescence (WD-XRF) and inductively coupled plasma mass spectrometer (ICP-MS) with each method associated with benefits and limitations. Limited research has been conducted to compare these techniques within the context of PM characterisation in the South African environment. Therefore, the aim of this study was to compare WD-XRF and ICP-MS for the analysis of trace metals in PM in order to assess the applicability of these techniques for source apportionment studies in South Africa. Atmospheric PM samples were collected at two sites in South Africa. Size segregated ambient aerosol samples were collected at a regional background site, i.e. Welgegund, while indoor samples were collected in an informal settlement in Agincourt. Weekly ambient PM samples were collected on filters for two months with a Dekati three-stage cascade impactor, while 24-hour indoor samples were collected with a GX Cyclone comprising of a PM4 inlet for four weeks. PM samples collected on filters were analysed with a WD-XRF spectrometer, after which aerosols were extracted with nitric acid digestion from these filters for analysis with an ICP-MS. In total, 31 trace metals were analysed by both analytical techniques. Only trace metal species that passed the assessment criteria were considered in the comparison of concentrations determined with WD-XRF and ICP-MS, which included 18 indoor trace metals (Na, Mg, Al, P, Ca, Ti, Cr, Mn, Fe, Co, Ni, Cu, Zn, Rb, Sr, Mo, Pd, Sb) and 16 ambient trace metals (Na, Mg, Al, P, K, Ca, Ti, Mn, Fe, Ni, Zn, As, Sr, Pd, Pt, Pb). The trace metal concentrations determined with both analytical methods indicated higher trace metal concentrations in indoor samples compared to ambient samples, which was attributed to ambient aerosols being more dispersed compared to particulates associated with indoor emissions and different sources impacting the two sites. Furthermore, the highest trace metal concentrations were determined for Na, Mg, Al and Ca. The percentages of samples for which concentrations of specific trace metals were above the limit of detection (LOD) of WD-XRF and ICP-MS differed for the two analytical techniques, while statistical correlations between trace metal concentrations determined with ICP-MS and WD-XRD also indicated very weak relationships between these species. In order to further assess these observed differences, concentrations determined for each trace metal species in each sample collected during the ambient and indoor campaigns with ICP-MS and WD-XRF were directly compared. These direct comparisons of concentrations determined for specific trace metals in each sample did not reveal distinct trends in the concentrations determined for most species with these two analytical techniques. However, it did seem that higher concentrations of trace metal species associated with crustal sources were determined with WD-XRF, which was expected due to the nitric acid digestion underestimating silicate minerals in dust during ICP-MS analysis. Lower trace metal concentrations determined with ICP-MS was also attributed to interferences during the analysis of extracted samples. Lower trace metal concentrations determined with WD-XRF compared to ICP-MS was generally attributed to most trace metal species occurring in the fine size fraction. Comparison of the concentrations determined for each trace metal species with the two analytical methods did not reveal distinct confirmations with regard to the suitability of ICP-MS and WD-XRF for determination of concentrations of specific trace metals in South Africa. Therefore, future studies in order to improve the comparison of these two analytical methods for atmospheric trace metal analysis in South Africa are strongly recommended.