Near-Field Lithography Optical Proximity Correction Based on Covariance Matrix Adaptation Evolution Strategy

Near-field lithography (NFL) can achieve super-resolution patterning beyond the optical diffraction limit by using evanescent waves. However, optical proximity effects in NFL can lead to nonuniform near-field intensity distribution, reduced image contrast, and increased pattern distortion in the photoresist. Optical proximity correction (OPC) is required to improve pattern fidelity. This work investigates a pixelated OPC method based on the covariance matrix adaptation evolution strategy (CMA-ES) under a rigorous coupled-wave analysis (RCWA) imaging model. The mask pixels are iteratively optimized according to the photoresist pattern error obtained from RCWA forward simulation. To improve mask regularity, total variation and connected component area penalties are introduced into the objective function to suppress fragmented pixel structures. Representative 110 nm half-pitch axis-symmetric layouts and additional two-dimensional layouts are used for evaluation. The results show that optimization driven only by pattern error can reduce proximity-induced pattern distortion, but it tends to generate fragmented pixel structures in the corrected mask layouts. After the regularization terms are introduced, the optimized mask layouts become more regular, with a moderate tradeoff in pattern error. The proposed method shows potential applications in mask layout optimization and hotspot repair for NFL in the future.