Computational techniques to incorporate shot count reduction into inverse lithography

We develop an inverse lithography method to tackle the shot count minimization in mask fracturing. The shot count minimization in model based fracturing is considered as a problem of finding a sparse combination of basis functions for the mask patterns, where the basis functions are defined as rectangles corresponding to the shots. This problem is formulated as a nonlinear least square problem, and a Gauss-Newton algorithm is proposed to solve it. The algorithm is modified to promote sparsity to reduce the shot count. Preliminary results of mask fracturing using the proposed algorithm is shown, and it is also incorporated into inverse lithography to show its effectiveness to reduce shot count.