Robust pixel-based source and mask optimization for inverse lithography

摘要

A robust pixel-based simultaneous source and mask optimization (SMO) method is proposed. A three dimensional (3D) partially coherent imaging model is used in this method. The process variations, such as defocus, exposure dose, aberrations, radiometric correction and apodization are incorporated in the 3D imaging model and the optimization framework. The pattern error is calculated with respect to the 3D image in the photoresist. The expectation of the pattern error is used as the cost function. The sensitivity of the cost function is employed to guide cost function in the decedent direction during optimization. The framework is finally solved by joint optimization of the source and mask patterns. Numerical results verify the validity of the proposed method. Influences caused by the aberrations are also shown by simulations.