Mask Inspection System Using MEEF-Based Regional Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mask inspection systems fail to accurately detect defects that may cause issues during pattern writing on wafers, despite conventional inspections deeming the patterns non-defective, especially for small pattern designs where high accuracy is required.
Innovation Solution
A mask inspection system that acquires optical image data, creates regional image data with multi-valued resolution based on importance level information, and uses a Mask Error Enhancement Factor (MEEF) to identify sections with a predetermined value, enhancing defect detection by comparing pixel values with threshold values or defect determination methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional inspection methods are used to determine patterns non-defective, then inspection speed is maintained, but defects that may cause issues during pattern writing on wafers are missed
Solution Approach 1:
The patent calculates MEEF values and creates regional image data with multi-valued resolution based on importance level information before the actual defect detection process. This preliminary classification of regions by their sensitivity to errors allows the inspection system to focus computational resources on critical areas, improving defect detection accuracy without proportionally increasing overall system complexity
Solution Approach 2:
The patent applies different inspection thresholds and resolution levels to different regions of the mask based on their MEEF values and importance levels. Critical regions with high MEEF values receive stricter inspection criteria and higher resolution analysis, while less critical regions use standard inspection parameters. This localized differentiation improves overall defect detection accuracy while managing computational complexity
2Measurement precision
If high resolution inspection is applied to all patterns, then defect detection accuracy is improved, but processing time and computational load increase significantly
Solution Approach 1:
The patent creates regional image data with multi-valued resolution where different resolution levels are assigned to different regions based on their importance and MEEF values. High-resolution inspection is applied only to critical regions with high MEEF values, while standard resolution is used for less critical regions. This selective approach maintains defect detection precision for important patterns while significantly reducing overall processing time and computational load
Solution Approach 2:
The patent applies excessive inspection rigor (high resolution and strict thresholds) only to specific critical regions rather than uniformly across the entire mask. By identifying regions where MEEF values indicate high sensitivity to errors, the system concentrates computational resources on partial areas that require the most careful inspection, avoiding the time penalty of applying high-resolution inspection to all patterns
3Manufacturing precision
If uniform inspection thresholds are used across all regions, then system complexity is reduced, but defects in critical small pattern designs are not detected
Solution Approach 1:
The patent determines multiple threshold values corresponding to different pixel value ranges in regional image data, where each threshold is tailored to the specific importance level and MEEF characteristics of that region. This allows the inspection algorithm to apply appropriate sensitivity levels to different pattern types - stricter thresholds for critical small patterns and more lenient thresholds for less critical areas - thereby improving manufacturing precision without requiring a single overly complex uniform algorithm
Solution Approach 2:
The patent changes the inspection parameters (threshold values and resolution levels) based on the regional characteristics derived from MEEF calculations and importance level information. By dynamically adjusting these parameters according to the specific region being inspected, the system achieves high inspection accuracy for critical patterns while managing algorithmic complexity through systematic parameter selection rather than ad-hoc complex algorithms
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach reduces the likelihood of defects being missed during inspection, allowing for the identification of potentially problematic patterns on wafers, thereby improving manufacturing yield and reducing costs associated with mask repair.
Implementation Method 1
Light transmitted through or reflected from the mask reaches an image sensor, thereby forming an image thereon
Data Source
AI summary
Optical image data of a mask is acquired. Reference image data associated with the optical images is created from design pattern data. Regional image data that includes pixel values denoted by multi-valued resolution based on importance level information of the patterns is created from region data including at least one portion of the patterns defined in the design pattern data. Defect determination is conducted on a pixel-by-pixel basis by comparing the optical image data with the reference image data, by means of either a plurality of threshold values determined by each pixel value within the regional image data or a plurality of defect determination methods. Image data of a section whose Mask Error Enhancement Factor (MEEF) is equal to or greater than a predetermined value is created from the region data including at least one portion of the patterns defined in the design pattern data.