Full project title: “Bayesian inference in optical metrology”. The BayesOM project aims to develop a new field method for optical metrology in the semiconductor industry. Instead of the traditional approach requiring the acquisition of multiple phase-shifted interferograms, we propose using algorithmic phase demodulation based on a single fringe image. The innovative aspect lies in the application of Bayesian inference, enabling precise estimation of the geometric parameters of structures such as waveguides. Unlike classical methods based on Fourier or Hilbert transforms, our approach achieves sub-pixel accuracy while minimizing errors in areas with abrupt height changes. Combining this technique with a precision interferometer could theoretically provide sensitivity at the level of single angstroms. The planned research includes both simulations and experiments with various Bayesian models to improve measurement accuracy under challenging conditions.

Principal Investigator: Prof. Maciej Trusiak

This section is currently being updated.