Learning-based compressive sensing method for EUV lithographic source optimization.

Area of Science:

• Semiconductor Manufacturing
• Nanotechnology
• Optical Engineering

Background:

• Extreme ultraviolet (EUV) lithography is crucial for fabricating advanced integrated circuits.
• Improving imaging fidelity in EUV lithography requires effective source optimization (SO).
• Current SO techniques face challenges in compensating for imaging distortions.

Purpose of the Study:

• To develop an efficient, learning-based source optimization approach for EUV lithography.
• To enhance imaging fidelity and computational efficiency in EUV lithography systems.
• To address the dimensionality and complexity of the EUV-SO problem.

Main Methods:

• A compressive sensing (CS) framework is employed to reduce the dimensionality of the EUV-SO problem through sparse sampling.
• The EUV-SO problem is formulated as an l1-norm inverse reconstruction problem utilizing sparse priors.
• A learning-based method is proposed to jointly optimize the source dictionary and projection matrix based on CS theory.

Main Results:

• The proposed method significantly reduces the dimensionality of the EUV-SO problem.
• The learning-based approach optimizes dictionary and projection matrices for improved performance.
• Simulations demonstrate enhanced imaging fidelity and computational efficiency.

Conclusions:

• The developed learning-based SO method offers a robust solution for EUV lithography.
• This approach effectively compensates for imaging distortions, leading to higher fidelity.
• The CS framework and optimized matrices contribute to computational efficiency in next-generation IC fabrication.