Title: Fully automated spectroscopic ellipsometry analyses of crystalline-phase semiconductors based on a new algorithm

Abstract: One significant drawback of a spectroscopic ellipsometry (SE) technique is its time-consuming and often complicated analysis procedure necessary to assess the optical functions of thin-film and bulk samples. Here, to solve this inherent problem of a traditional SE method, we present a new general way that allows full automation of SE analyses for crystalline-phase semiconductors exhibiting complex absorption features. In particular, we have modified a scheme established in our previous study, which performs a non-linear SE fitting analysis only in a low energy region at the beginning, while the analyzed energy region is gradually expanded toward higher energy by incorporating addition optical transition peaks. In this study, we have further developed a unique analyzing-energy search algorithm, in which a proper analyzing-energy region is determined to incorporate the feature of a new transition peak. In the developed method, a drastic improvement over the previous simple approach has been confirmed for expressing complex dielectric functions consisting of sharp and broad absorption peaks. The proposed method (Delta M method) has been applied successfully to analyze perovskite-based crystalline samples, including hybrid perovskite (CH3NH3PbI3) and chalcogenide perovskites (SrHfS3 and BaZrS3). In the automated analyses of these semiconductors, 7-8 transition peaks are introduced automatically to describe sample dielectric functions, while structural parameters, such as thin-film and roughness thicknesses, are also determined simultaneously. The established method can drastically reduce an analysis time to a level that allows the automatic inspection of daily varying material optical properties and expands the application area of spectroscopic ellipsometry considerably.