In this paper the overview of the most important approximate methods for the optical characterization of inhomogeneous thin films is presented. The following approximate methods are introduced: Wentzel–Kramers–Brillouin–Jeffreys approximation, method based on substituting inhomogeneous thin films by multilayer systems, method based on modifying recursive approach and method utilizing multiple-beam interference model. Principles and mathematical formulations of these methods are described. A comparison of these methods is carried out from the practical point of view, ie advantages and disadvantages of individual methods are discussed. Examples of the optical characterization of three inhomogeneous thin films consisting of non-stoichiometric silicon nitride are introduced in order to illustrate efficiency and practical meaning of the presented approximate methods.
Dispersion models are necessary for precise determination of the dielectric response of materials used in optical and microelectronics industry. Although the study of the dielectric response is often limited only to the dependence of the optical constants on frequency, it is also important to consider its dependence on other quantities characterizing the state of the system. One of the most important quantities determining the state of the condensed matter in equilibrium is temperature. Introducing temperature dependence into dispersion models is quite challenging. A physically correct model of dielectric response must respect three fundamental and one supplementary conditions imposed on the dielectric function. The three fundamental conditions are the time-reversal symmetry, Kramers-Kronig consistency and sum rule. These three fundamental conditions are valid for any material in any state. For systems in equilibrium there is also a supplementary dissipative condition. In this contribution it will be shown how these conditions can be applied in the construction of temperature dependent dispersion models. Practical results will be demonstrated on the temperature dependent dispersion model of crystalline silicon.