Strategy for the maximum extraction of information generated from combinatorial experimentation of Co-doped ZnO thin films

Abstract By expanding our level of understanding of structure–processing–property relationships through a data-mining methodology, this study demonstrates how to remove obstructions in complex high-throughput (HT) data analyses for developing new transparent conducting oxides. The demonstration is performed with principal component analysis (PCA) as an exploratory data analysis tool in the context of Co-doped ZnO (Co:ZnO) thin films generated from combinatorial HT syntheses. With the use of minimal available information, X-ray diffraction (XRD) patterns and their corresponding processing conditions, PCA enabled effective detection of pervasive changes in intensity and peak shifts as a function of composition, processing or a combination of both. These identifications are nearly impossible to detect via normal data interpretation methods. It was also possible to identify abnormal XRD patterns, unusual composition arrays (i.e. libraries), key chemistries in compositional arrays and critical peak occurrences.