Correction of matrix effects via scattered radiation in X-ray fluorescence analysis of samples collected on membrane filters

Abstract

An empirical model for correction of matrix effects in samples collected on a membrane filters is proposed. Since the matrix correction is performed using the ratio of coherent (Rayleigh) to incoherent (Compton) X-ray intensities, the matrix elements need not be quantified. Contrary to intermediate-thickness samples without substrate, the ratio of radiation coherently and incoherently scattered from the sample collected on the filter depends on sample thickness. This relationship is discussed in this paper. The proposed model can be used for the determination of both medium- and heavy-Z elements in light matrix materials e.g. in environmental, biological or geological samples. The results are satisfactory, even if a minor matrix element, such as Fe, has its absorption edge energy between the energy of the characteristic radiation of an analyte element and the energy of the scattered primary radiation. The validity of the proposed algorithm was verified with certified geological materials (the matrices were composed of silicate, aluminosilicate, phosphate) and fly ash collected on a membrane filter. The elements from Z = 19 (K) to Z = 92 (U) were determined in samples of mass per unit area between 1 and 3 mg cm⁻². The agreement between XRF analysis and certified values is satisfactory and indicates the usefulness of the proposed matrix correction method, e.g. for the determination of trace and minor elements in environmental or geological samples. The measurements were performed using a wavelength-dispersive X-ray spectrometer (WDXRF) with a molybdenum target X-ray tube.