Automated lamellar block copolymer process characterization

Lamellar CD-SEM image analysis is one of the key step for the development of new polymer formulations. We present in this paper a new approach for the analysis of lamellar CD-SEM that can be extended to any type of other pattern (contact…) with a machine learning approach. We will also focus on the roughness analysis and specifically the Line Edge Roughness (LER) and Power Spectral Density (PSD) with a robust estimation that takes into account curvature of the line. The last part is dedicated to the introduction of a process optimisation technique using machine learning to optimize process parameters from a first design of experiment.