Inverse lithography source optimization via compressive sensing

Area of Science:

• Computational lithography
• Image processing and reconstruction

Background:

• Current source optimization (SO) methods in lithography are computationally intensive or offer limited process window (PW) extension.
• Pixel-based SO approaches often involve solving complex quadratic or linear programming problems.

Purpose of the Study:

• To develop an efficient and robust source optimization method for improving lithographic imaging.
• To leverage compressive sensing (CS) theory for accelerated and enhanced SO design.

Main Methods:

• Formulating source optimization as an underdetermined linear problem, enabling acceleration.
• Transforming the SO problem into an l1-norm image reconstruction problem based on CS principles.
• Applying the linearized Bregman algorithm to synthesize sparse optimal source patterns.

Main Results:

• The proposed linear SO formulation demonstrates superior aerial image contrast compared to traditional quadratic methods.
• Sparse regularization in inverse lithography effectively extends the process window (PW) of lithography systems.
• Simulations confirm the superiority of the proposed SO method over existing approaches.

Conclusions:

• Compressive sensing provides an efficient and robust framework for source optimization in lithography.
• The developed method enhances manufacturability and extends the process window, improving lithographic imaging performance.