Rapid Control of submicrometer periodic structures by a neural inversion from ellipsometric measurement

Subwavelength diffraction gratings created the need for rapid and reliable characterization techniques in particular in the microelectronic industry. Scatterometry proves to be the only alternative for the future among the current methods. It is based on the reconstruction of the grating profile from its optical response. Ellipsometry seems to be a powerful technique for the optical measurement and neural network is receiving increased attention as a powerful optimization method. The goal of this work is to show that the combination of ellipsometry and neural treatment can lead to a rapid and robust method for the prediction of grating geometrical parameters.