Inverse Lithography with Adaptive Mask Complexity

Mask complexity is becoming a critical issue for advanced semiconductor nodes both in optical proximity correction (OPC) and post OPC manufacturing flows, especially when curvilinear features are introduced with inverse lithography technology (ILT) for pattern fidelity improvement. In this paper, we address the mask complexity by adaptively incorporating constant width sub-resolution assist features (SRAFs) into the level-set based inverse lithography. We first extract the SRAFs up to the 3
^rd
order with respect to the distance to the main features from a continuous transmission mask (CTM) and distance maps of the features skeleton are constructed. Mask complexity is regulated by applying constant width to the selected orders of SRAF features while the main features and the SARFs with other orders are optimized as freeform pixels. Simulation results show comparable pattern fidelity performance in terms of pattern error (PE) with adaptive constant width SRAF selection, and also mask complexity is reduced because constant width SRAFs, for which curvature and width mask rules are implicitly complied, can be manufactured by less circular shots.