Complex Aerosol Nanostructures: Revealing the Phases from Multivariate Analysis on Elemental Maps Obtained by TEM-EDX

Transmission electron microscopy (TEM) is a popular off-line technique to study aerosol nanoparticles. Coupled with an X-ray detector, high-resolution elemental maps of the sample can be obtained by scanning the focused electron beam and collecting the emitted X-ray spectra, called energy dispersive X-ray spectroscopy (EDX). Interpretation of the acquired data can be difficult in the case of complex particles having regions of different and overlapping phases because the elemental distributions do not reveal phase information. However, the resulting large datacubes collected are well-suited for multivariate statistics techniques to reveal information clouded by the noise of raw spectra. In this work, we show the comparison between two multivariate techniques, hierarchical clustering and non-negative matrix factorization, to separate elemental maps into phase maps of Cu@Ag core@shell aerosol nanoparticles and aerotaxy nanowires. We compare the results of the Cu@Ag sample to X-ray photoelectric spectroscopy measurements as an independent measure of surface composition.