Source mask optimization using the covariance matrix adaptation evolution strategy.

Area of Science:

• Photolithography
• Optical Engineering
• Semiconductor Manufacturing

Background:

• Source Mask Optimization (SMO) is critical for resolution enhancement in advanced semiconductor lithography (2Xnm nodes and beyond).
• Existing SMO methods face challenges in optimization capacity and convergence efficiency, particularly for full-chip applications.
• Ensuring image fidelity and process robustness requires efficient and powerful SMO techniques.

Purpose of the Study:

• To propose a novel Source Mask Optimization (SMO) method that improves optimization capacity and convergence efficiency.
• To enhance the performance of lithographic processes for current and future technology nodes.
• To address the limitations of existing SMO techniques in terms of speed and effectiveness.

Main Methods:

• Utilized the Covariance Matrix Adaptation Evolution Strategy (CMA-ES) algorithm for SMO.
• Developed a new source representation method based on ideal point sources with adjustable positions and unit intensity.
• Employed a forward vector imaging formulation with specific encoding/decoding methods and a merit function for optimization.

Main Results:

• The proposed method demonstrates superior optimization capacity compared to existing techniques.
• The CMA-ES algorithm, combined with the new source representation, significantly improves convergence efficiency.
• Simulations and comparative analyses validate the effectiveness of the developed SMO approach.

Conclusions:

• The novel SMO method using CMA-ES offers enhanced optimization capabilities and faster convergence.
• The new source representation effectively identifies beneficial spatial frequency components for imaging performance.
• This technique provides a robust solution for advanced lithography, ensuring image fidelity and process robustness.