Retrieving the quantitative chemical information at nanoscale from SEM EDX measurements by Machine Learning

摘要

The quantitative composition of metal alloy nanowires on InSb(001)semiconductor surface and gold nanostructures on germanium surface isdetermined by blind source separation (BSS) machine learning (ML) method usingnon negative matrix factorization (NMF) from energy dispersive X-rayspectroscopy (EDX) spectrum image maps measured in a scanning electronmicroscope (SEM). The BSS method blindly decomposes the collected EDX spectrumimage into three source components, which correspond directly to the X-raysignals coming from the supported metal nanostructures, bulk semiconductorsignal and carbon background. The recovered quantitative composition isvalidated by detailed Monte Carlo simulations and is confirmed by separatecross-sectional TEM EDX measurements of the nanostructures. This shows that SEMEDX measurements together with machine learning blind source separationprocessing could be successfully used for the nanostructures quantitativechemical composition determination.