Photolithography Process Variation Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Process variations in semiconductor manufacturing lead to differences between actual and desired integrated circuit characteristics, causing manufacturing problems and reducing yield and performance, necessitating a method to determine the effect of these variations on critical dimensions without requiring multiple prints of the layout.
Innovation Solution
A system that creates a layout with multiple pattern instances, corrects them using photolithography process models calibrated under different exposure and focus conditions, and then prints them under nominal conditions to determine the effect of process variations by measuring critical dimensions, thereby reducing time and resources required.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple layouts are printed under different process conditions to determine critical dimensions, then measurement precision is improved, but loss of time and resources increases
Solution Approach 1:
The patent applies preliminary action by pre-calibrating photolithography process models under different process conditions (exposure and focus variations) before actual manufacturing. Multiple virtual layouts are corrected using these pre-calibrated models to simulate different process conditions, eliminating the need to physically print multiple times while maintaining measurement precision
Solution Approach 2:
The patent uses virtual copies of layouts corrected with different photolithography process models to represent different process conditions. Instead of creating multiple physical prints, the system creates multiple corrected layout versions that mimic the effects of different exposure and focus conditions, allowing critical dimension determination from a single physical print
2Measurement precision
If multiple layouts are printed under different process conditions to determine critical dimensions, then measurement precision is improved, but resource expenditure increases
Solution Approach 1:
The patent performs preliminary calibration of photolithography process models under various process conditions before production. This pre-work allows the system to use computational corrections on a single layout rather than physically printing multiple layouts under different conditions, significantly reducing resource expenditure while maintaining measurement precision
Solution Approach 2:
The patent creates virtual copies of the layout corrected with different process models instead of creating multiple physical prints. Each corrected layout version represents a different process condition, allowing the system to evaluate multiple scenarios from a single physical print, thereby conserving manufacturing resources
3Productivity
If process variations are not accounted for, then manufacturing speed is maintained, but manufacturing precision deteriorates
Solution Approach 1:
The patent uses parameter changes by varying the photolithography process model parameters (exposure and focus conditions) during the correction phase to account for process variations. This allows the system to evaluate how different process parameters affect critical dimensions without actually changing physical process conditions during manufacturing, maintaining both speed and precision
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
Embodiments of the present invention provide systems and techniques for determining the effect of process variations. During operation, the system can receive a layout which includes multiple instances of a pattern. Next, the System can correct the pattern instances using different photolithography process models which model the photolithography process at different exposure and focus conditions. Next, the corrected layout can be printed on a wafer. The system can then perform electrical tests on the wafer, or it can measure the critical dimensions of the features on the wafer. The yield loss or the exposure focus matrix can then be generated by using the test data or the measurement data.