Correlative microscopy and machine learning –new tools for material characterization

The correlation of different microscopic techniques has seen increased interest in recent years due to the possibility of combining the strengths of multiple techniques.In addition to the practical challenges with regard to sample preparation, instrument design and the need for operators experienced in multiple techniques, uniquedata treatment challenges arise when combining data sets with different resolutions and contrast mechanisms. Using Raman-SEM-EDS as an example correlativetechnique we are discussing two approaches for correlative microscopy and data treatment thereof on the example specimen of a WO3-WS2 powder and a volcanicrock. The aim of both approaches is to translate microscopic images (or mappings) into quantified data. The first approach puts the focus on getting the mostinformation out of minimal experimental effort (WO3-WS2 powder). The second approach puts the focus on maximal analytical quality (volcanic rock).