Efficient nonlinear resist modeling by combining and cascading quadratic Wiener systems

Optical proximity correction (OPC) is currently a crucial technique for improving photolithographic image quality, with its correctness and efficiency relying heavily on resist modeling. The resist exposure and development processes involve various complex nonlinear physicochemical reactions, which impose a great challenge to model in a fast and accurate manner that meets the requirements of OPC. In this paper, we propose a set of solutions for nonlinear resist modeling, including a multi-stage cascaded quadratic Wiener model network, a simulation acceleration method using eigendecomposition, and an efficient method of model calibration utilizing the projected Landweber method. Various experiments are conducted for validation, which demonstrate that a single-stage Wiener model may already meet the usual requirements for production worthiness, while a multi-stage model achieves even better performance in terms of model accuracy, generality, and speed. The proposed resist modeling method holds great potential in advancing photolithography modeling and OPC for modern semiconductor manufacturing.