MASCOT : Mask pattern correction tool using genetic algorithm

We propose a new optical proximity correction (OPC) method using a polygonal mask model and a genetic algorithm (GA). With the polygonal mask model, our method can generate simple and practical masks that satisfy the constraints arising from manufacturing requirements. A robust optimization method, the genetic algorithm was adopted for obtaining adequate modifying ability for a given arbitrary mask pattern. For some of these mask patterns, the characteristics of distortion due to the optical proximity effects (OPE) are not well-known. GAs simulate the free-competition among animals through generation alternation is recognized as the algorithm that can avoid being stuck in local optima in multi-modal optimization problems. We developed our mask pattern correction tool (MASCOT) using the mask model and the optimization method. Experimental results showed great improvement in the generated resist pattern. Satisfactorily practical computational performance was obtained in our experiments with a workstation.