Phosphor informatics based on confirmatory factor analysis.

The theoretical understanding of phosphor luminescence is far from complete. To accomplish a full understanding of phosphor luminescence, the data mining of existing experimental data should receive equal consideration along with theoretical approaches. We mined the crystallographic and luminescence data of 75 reported Eu(2+)-doped phosphors with a single Wyckoff site for Eu(2+) activator accommodation, and 32 descriptors were extracted. A confirmatory factor analysis (CFA) based on a structural equation model (SEM) was employed since it has been helpful in understanding complex problems in social sciences and in bioinformatics. This first attempt at applying CFA to the data mining of engineering materials provided a better understanding of the structural and luminescent-property relationships for LED phosphors than what we have learnt so far from the conventional theoretical approaches.