Spectral Reflectometry-Based Film Thickness Measurement Using Particle Swarm Optimization

To meet the requirement for high-precision thickness measurement of silicon-based thin films in integrated circuits and the semiconductor industry, spectroscopic reflectometry has been widely used due to its non-contact and non-destructive characters. However, in such methods, the film thickness inversion relies on the spectral fitting, which often needs too much computational cost to be utilized in real-time applications. This study proposes a film thickness inversion approach to integrate a chromatic confocal system with a Particle Swarm Optimization (PSO) algorithm for efficient and accurate data fitting. By leveraging the global convergence properties of PSO, the method enables rapid optimization of reflectance model parameters and high accuracy. Experimental validation on a custom-built measurement platform with standard samples demonstrates the effectiveness and feasibility of the proposed approach in thin film thickness measurement.