Evaluation of Sensitivity of Multivariate Statistical Analysis on STEM Spectrum-Imaging Datasets and its Improvement

Aberration-corrected scanning transmission electron microscopy (STEM) is the essential approach for atomic-scale characterization towards advanced materials developments. Not only atomic-scale imaging but also chemical analysis via electron energy-loss spectrometry (EELS) and X-ray energy dispersive spectrometry (XEDS) can be performed routinely by using the latest aberration-corrected STEM instruments [e.g. 1]. It would no longer be a dream to apply chemical analysis at atomic scales with single-atom detection sensitivity. In combination with the latest hardware, the advances in the recent software developments allow us to acquire large-scale datasets such as multidimensional image series and spectrum images (SIs). Although the trend to acquire large-scale datasets is desired, it is more challenging to deal with the large-scale datasets, e.g. extraction of unknown features and estimation of dominant trends. If the datasets are relatively noisy, which is very common for atomic-resolution EELS/XEDS SIs, data analysis would be much harder tasks. Multivariate statistical analysis (MSA) is one of efficient approaches to analyse the large-scale datasets in terms of feature identification and extraction [2, 3]. Since a use of MSA is relatively straightforward, various MSA approaches have been applied to SIs as data-mining and noise-reduction tools [e.g. 4]. Despite that the MSA approach is very efficient and useful, it may create unexpected artifacts especially in higher noise conditions [5]. Since these artifacts might mislead results, it is essential to explore the sensitivity of MSA. In this study, the MSA sensitivity will be evaluated by using SIs simulated to match atomic-resolution analysis conditions. Then, procedures to improve the MSA sensitivity will be proposed.