Abstract
This paper presents a new method of Inverse Lithography Technology (ILT) to accelerate its optimization process and convergence, which involves two steps in iteration: the first step is to find hot-spots in the layout in the optimization process and the second step is to optimize on the hot-spot areas. Information in optimization process such as gradient can be used in the first step to help finding the hot-spots. Optimized surroundings have been used as environment for local hot-spot fixing in the second step. In order to reduce ripple effects, specific penalty terms are used in the second step, thus the local fixing for hot-spots would minimally affect the surroundings through the procedure. This method focuses on the most un-optimized places and thus accelerates the optimization. Experiments show that this hot-spots aware ILT outperforms traditional ILT.