Narrow-band level set calculation method for collaborative optimization and semi-implicit discretization of light source mask

Abstract

The invention discloses a light source mask collaborative optimization semi-implicit discretization narrowband level set calculation method, and the method comprises the following steps: taking a discretized circuit board diagram variable and a discretized light source variable as observation matrixes, and taking a condition that a corresponding imaging matrix is equal to a discretized matrix of atarget pattern as a constraint condition; constructing a Lagrange equation containing a distance regularization level set item and a constraint condition to obtain a time-varying difference equation;implicitly and explicitly discretizing a diffusion term and a non-diffusion term respectively, constructing a three-diagonal linear equation set in a horizontal direction and a vertical direction byusing additive operator splitting, and performing efficient solution by using a Thomas method. According to the method, the optimization convergence efficiency can be improved while the optimization dimension is reduced.

Description

Technical field[0001]The invention relates to the technical field of optical proximity correction in the lithography resolution enhancement technology, and in particular to a narrowband level set calculation method for semi-implicit discretization of light source mask collaborative optimization.Background technique[0002]The projection lithography system is the core equipment for manufacturing micro- and nano-scale line-width ultra-large-scale integrated circuits. The projection lithography system is mainly composed of an illumination system, a mask, a projection objective, a pupil, and a silicon wafer coated with photoresist. The light waves emitted by the light source illuminate and pass through the mask to generate the near field of the mask, and project through the projection objective. The pupil’s low-pass filtering and photoresist etching transfer the mask pattern to the silicon wafer, but the loss of information in the image transfer process causes the shape on the silicon waf...

AI Technical Summary

Problems solved by technology

[0006] First, the above method optimizes and updates all observation points on the light source and mask layout during the iterative process. It needs to calculate the gradient of the spatial image and the cost function to the observation points of the light source and mask. The calculation efficiency is low, which is not conducive to large-scale Light source mask co-optimization simulation

[0007] Second, the above method uses a simple and widely used explicit discretization method, because of the limitation of the Courant-Friedrichs-Levy condition (CFL), the iteration step size is suppressed to produce excessive iterations, resulting in slow convergence; and Although the corresponding implicit discretization method overcomes the stability constraint and can use a sufficiently large iterative step size, it needs to solve a considerable scale of linear equations, which has a high computational complexity and is difficult to apply to the actual OPC technology.

[0008] In summary, the existing SMO methods need to be further improved and improved in terms of optimizing the mask frame, computational efficiency, and convergence efficiency.

Images