Determination of rare-earth elements in medium- and high-entropy ceramics by WDXRF and handheld XRF. Critical evaluation of the need for deconvolution

Abstract

In this paper, we discuss strategies for analysing medium- and high-entropy ceramics (REE₃NbO₇) that contain 4 and 5 different rare-earth elements. In the coming years, it may be necessary to shift from laboratory-based analysis of ceramic powders to on-site examination of thermal barrier coatings, particularly those employed in aircraft engines. To address this, we employed a portable X-ray fluorescence spectrometer capable of in-field non-destructive simultaneous quantification of Gd, Er, Tm, Yb and Y. The major problem with portable systems is the overlap of Gd, Er, Tm, and Yb lines. Therefore, we conducted a complete comparative study of the strategies that allow for the acquisition of analytical signals in the energy-dispersive spectra. In order to critically evaluate the need for deconvolution, a technique often suggested in the literature, we used wavelength-dispersive X-ray fluorescence. This XRF variant was used to obtain spectra with higher resolution, and thus different levels of line overlap. In this work, we highlighted peculiarities of the deconvolution process in higher- and lower-resolution X-ray fluorescence spectra. The study revealed that the portable spectrometer is capable of accurately quantifying the concentrations of Gd, Er, Tm, Yb, and Y in both medium- and high-entropy ceramics. A comparative analysis showed that quantitative analysis of medium- and high-entropy ceramics does not require deconvolution of energy-dispersive spectra. It is shown that the samples synthesised using the reverse precipitation method allow one to obtain reliable calibration curves and avoid matrix effects using a portable X-ray fluorescence spectrometer, despite the fact that the concentrations of the major components vary 3 times. However, it was found that experimental conditions in a vacuumed wavelength-dispersive X-ray fluorescence spectrometer reduced trueness by approx. 9%. Nevertheless, this system proved to be a suitable substitute for ICP-OES in laboratory practices. The portable version, on the other hand, holds promise for in-field use.