Process optimization of contact hole patterns via a simulated annealing algorithm in extreme ultraviolet lithography

Area of Science:

• Semiconductor manufacturing
• Advanced lithography techniques
• Materials science for microelectronics

Background:

• Critical dimension (CD), roughness, and sensitivity are key photoresist imaging performance indicators in extreme ultraviolet (EUV) lithography.
• Shrinking CDs necessitate tighter control for high-fidelity imaging, yet research often neglects two-dimensional (2D) patterns.
• Existing methods focus on 1D patterns, leaving a gap in optimizing 2D feature fabrication.

Purpose of the Study:

• To develop an image quality optimization method for 2D contact holes in EUV lithography.
• To address the need for simultaneous control of multiple imaging indicators in advanced semiconductor patterning.
• To accelerate the development of EUV lithography processes for next-generation devices.

Main Methods:

• Utilized machine learning to correlate process parameters with 2D contact hole imaging indicators: horizontal/vertical widths and edge roughness, alongside sensitivity.
• Employed a simulated annealing algorithm to efficiently search for optimal process parameters.
• Integrated multiple performance metrics for comprehensive pattern quality evaluation.

Main Results:

• Achieved highly accurate optimization of 2D contact hole imaging parameters.
• Demonstrated significant improvements in overall device performance through optimized lithography.
• Validated the method's effectiveness with rigorous experimental imaging results.

Conclusions:

• The developed method effectively optimizes 2D contact hole imaging quality in EUV lithography.
• Machine learning and simulated annealing provide a powerful approach for complex lithography process control.
• This technique accelerates the development cycle for advanced semiconductor devices.