Development of a novel online sequential extraction method for the characterisation of fine dust sources using inductively coupled plasma mass spectrometric detection

Abstract

Multi-elemental analysis in combination with statistical models has the potential to support the apportionment of numerous sources of airborne dust. Lack of selectivity can be overcome by the application of sequential extraction schemes. However, classical offline extraction schemes are designed for >0.5 g of sample and are not suitable for particulate matter in (sub-)mg range.Therefore, a novel online sequential extraction (OSE) scheme was developed based on the well-established Tessier and BCR schemes. The solvent composition (deionised water, ammonium nitrate, acetic acid and hydroxylammonium chloride) was adapted to allow direct online introduction of the extracts into sector-field inductively coupled plasma mass spectrometry (ICP-MS). The OSE setup was optimised regarding sample containment, flow rates, transfer lines and ICP-MS transient data recording. Four model dust source samples (soil, coal, mine overburden and road dust) were screened by OSE for elemental fractionation patterns. Characteristic fractograms were obtained demonstrating the difference in mobility of the investigated elements Na, Mg, Al, K, Ca, V, Cr, Mn, Co, Ni, Cu, Zn, Rb, Sr, Cd, Ba and Pb in the chosen sample matrices. Results for the road dust, a very fine and homogenous matrix, showed high repeatability with most relative standard deviations from quadruplicate analysis <10%. Standard deviations for the other model samples were partly higher due to lower homogeneity of these matrices at the applied sample mass of only 5 mg.Verification of the elemental fractionation via the novel OSE was performed for the road dust model sample. A flow injection setup was applied to check the accuracy of the applied post-cartridge calibration approach. Recoveries for the QC standard were 90% with SD 4%. Bracketing the OSE runs with flow injection of the QC standard demonstrated the stability of the OSE operation over several days of measurement. Variation of the sample mass revealed a significant increase of the extracted mass fractions at lower mass of 2 mg. The results of this optimised road dust OSE data set were compared to offline sequential extractions of the same sample. Average recovery for the first fraction (water-soluble) was 87% with SD 21% indicating comparable extraction efficiency by OSE. Recoveries for the other extraction steps were partly lower due to lower reagent concentrations and much shorter extraction times. However, the OSE provided repeatable results matching the extraction pattern of the offline extraction. Overall, the summed extracted mass fractions of all 4 OSE extraction steps equalled 72% of the same fractions obtained with offline extraction. Considering the operationally defined character of all sequential extraction schemes these results demonstrate the reliability of OSE for elemental fractionation of fine particulate samples at mg range sample mass.