Mask Corner Rounding in OPC Modeling

The continued development of semiconductor manufacturing technology and the pressing demand for the accuracy and predictability of lithography simulators, it becomes increasingly important to understand actual mask condition. This paper aims to identify an effective and efficient method to model mask corner rounding (MCR) in computational lithography. First, we have found that Gaussian function can be used as a good convolution kernel for mask, which can represent the MCR, and we have found a good corresponding relationship. Second, we propose a modeling flow which integrates the MCR effect and optical simulation framework, and we have found that it seems to perform quite well.