Accelerating Flux Calculations Using Sparse Sampling.

Area of Science:

• Materials Science
• Computational Physics
• Electrical Engineering

Background:

• Miniaturization in electronics presents design and fabrication challenges.
• Computer simulations are crucial for optimizing fabrication processes but face computational complexity with 3D shapes.
• Re-evaluating particle flux on the substrate surface is a major bottleneck in topography simulations.

Purpose of the Study:

• To develop a method for enhancing the performance of particle flux re-evaluation in simulations.
• To reduce the computational cost of high-resolution topography simulations for electronic device fabrication.
• To maintain accuracy while improving simulation speed for complex 3D structures.

Main Methods:

• Calculating particle flux on a selected subset of surface elements instead of the entire surface.
• Employing an advanced multi-material iterative partitioning scheme for subset selection.
• Considering local flux differences and geometrical variations in the partitioning scheme.
• Validating the approach using an etching simulation of a dielectric layer in a multi-material stack.

Main Results:

• Achieved simulation speedups ranging from 1.8 to 8.0.
• Maintained surface deviations below two grid cells (0.6-3% of feature size) across all configurations.
• Demonstrated the feasibility and effectiveness of the proposed performance enhancement method.

Conclusions:

• The developed method significantly improves the efficiency of electronic device fabrication simulations.
• The approach offers a practical solution for handling complex 3D geometries in simulations.
• This technique enables faster optimization of fabrication processes for advanced electronic devices.