Development of the Model for Estimating Thickness and Optical Constants of Metallic Thin Films using Only Reflection Spectrum

Metallic thin films are essential for various device applications. Given that sensor performance highly depends on thin film thickness and optical constants, ellipsometry is an indispensable measurement technique. While a cost-effective alternative using a spectrometer exists, it is cannot simultaneously determine both thickness and optical constants. In this study, we develop and validate a machine learning model capable of simultaneously estimating film thickness and optical constants from a single reflectance spectrum, thereby extending its applicability to opaque substrates and materials for which conventional models are inadequate.