Phase recognition in SEM-EDX chemical maps using positive matrix factorization

Area of Science:

• Materials Science
• Analytical Chemistry
• Data Analysis

Background:

• Scanning electron microscopy (SEM) coupled with energy-dispersive X-ray spectroscopy (EDX) provides detailed elemental composition of solid materials.
• Understanding the mixing state and interactions of elements is crucial for material characterization.

Purpose of the Study:

• To adapt and apply Positive Matrix Factorization (PMF) for analyzing elemental maps generated by SEM-EDX.
• To demonstrate the capability of PMF in identifying common elemental features and visualizing their spatial distribution.

Main Methods:

• Conversion of graphical and digital SEM-EDX images into PMF input files.
• Application of PMF, a multivariate factor analysis technique, to group elements based on common occurrences in elemental maps.
• Visualization of PMF-derived factor maps to illustrate chemical composition and interactions.

Main Results:

• PMF successfully grouped elements into distinct factors, representing common chemical associations.
• Visualized factor maps provided insights into ash interactions and the composition of different chemical layers in oxygen carrier materials.
• The method demonstrated effectiveness in handling various chemical mapping data, including large datasets.

Conclusions:

• PMF is a valuable tool for analyzing complex elemental mapping data from techniques like SEM-EDX.
• The adapted method enhances the understanding of material composition, mixing states, and interfacial chemistry.
• This approach is versatile and applicable to diverse chemical mapping datasets beyond SEM-EDX.