Evaluation of quantitative procedures for X-ray microanalysis of environmental particles

Scanning electron microscopy combined with energy-dispersive X-ray spectrometry is particularly suited to characterizing morphology and elemental composition of individual microparticles. Although not straightforward, quantitative X-ray microanalysis of low-Z-containing particles is achievable using atmospheric thin-window X-ray detectors. A critical aspect of light element analysis is the choice of substrate material. In this work, particles were deposited on specially developed boron substrates. Three case studies were investigated successively in the order of increasing difficulty. Firstly, hundreds of calcium carbonate (CaCO(3)) particles ranging in size from 0.3 to 10 microm were analyzed. Three quantitative procedures were tested: the "k-ratio" method, conventional ZAF correction, and Monte Carlo simulations. Average relative errors obtained by the reverse Monte Carlo quantitative program named CASINO were better than 2.5 wt %, carbon included. Secondly, further evaluation was carried out on a finely crushed biotite mineral, containing more than nine elements. Finally, airborne particulate matter, consisting of a complex heterogeneous mixture of particles, was investigated. By applying the Monte Carlo quantitative procedure, the observed particles were easily classified into particle types. Pure compounds (e.g., CaSO(4).2H(2)O, SiO(2), CaCO(3), etc) were directly assigned according to stoichiometry. In some cases (marine-derived particles), a partial reactivity of atmospheric particles was demonstrated by quantitative analysis.