DMD digital lithography optimization based on a hybrid genetic algorithm and improved exposure model

Abstract

An efficient method to optimize digital micromirror device (DMD) lithography was proposed using a hybrid genetic algorithm integrated with an improved exposure model. The improved exposure model significantly refines traditional approaches by incorporating advanced parameters not previously considered, including the cross-transfer coefficient, detailed light source functions, and impulse response functions. These enhancements provide a comprehensive assessment of the entire optical imaging system’s impact on lithography quality and more accurately simulate the interactions of light with the photoresist. The hybrid method combines the robust optimization capabilities of genetic algorithms (GA) with this sophisticated exposure model, facilitating precise micromirror configurations and optimizing light distribution for specific lithographic patterns. This integration results in substantial improvements in lithographic precision, with improvements of up to 84% for hexagonal star patterns, 83% for arrow patterns, and 85% for embedded figure patterns. These advancements enhance imaging quality, reduce optical proximity distortions, and improve overall lithography performance, offering crucial insights into the precision and efficiency improvement of microelectronics fabrication processes.