Application of Spectral Ellipsometry for Dielectric, Metal and Semiconductor Films in Microelectronics Technology

Abstract

The article reviews model-based and model-free approaches to solving problems of spectral ellipsometry related to the measurement of thicknesses and optical parameters of thin layers of dielectrics, metals and semiconductors in microelectronics application. Model-based approaches employ a priori information about the dispersion relation in form of the Cauchy, Drude, Drude—Lorentz and Tautz—Lorentz. Model-free approaches can use any smooth multivariate functional dependence describing a smooth spectral curve. Also, machine learning can be used to implement the model-free approach, which is well suited for determining the thickness of multilayer structures and their optical characteristics and allows to significantly increase the speed of data processing.