Dynamic Threshold Adjustment for Photomask Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional pattern inspection apparatuses face challenges in accurately detecting defects in ultrafine patterns due to the generation of pseudo defects, particularly with optical proximity correction (OPC) patterns, leading to inefficient use of resources and repeated inspections.
Innovation Solution
A pattern inspection apparatus that compares partial optical image data using region image data to dynamically adjust the decision criterion, reducing pseudo defects by varying the inspection threshold based on region-specific information, thereby improving detection accuracy for high-precision patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed inspection threshold is used for all regions, then the inspection process is simple and fast, but pseudo defects are generated in certain regions reducing measurement precision
Solution Approach 1:
The patent applies local quality by dividing the inspection target into multiple regions and assigning different inspection thresholds to each region based on its characteristics. This allows the inspection system to use stricter thresholds for critical regions and more lenient thresholds for non-critical regions, thereby improving overall measurement precision while avoiding pseudo defects in specific areas.
Solution Approach 2:
The patent implements dynamics by making the inspection threshold variable rather than fixed. The threshold is dynamically adjusted according to the regional characteristics of the inspection target, allowing the system to adapt its sensitivity to different areas. This dynamic approach resolves the contradiction by enabling high precision where needed while maintaining simplicity in regions where strict inspection is not necessary.
2Measurement precision
If a strict inspection criterion is applied, then defect detection accuracy is improved, but pseudo defects increase leading to repeated inspections and resource wastage
Solution Approach 1:
By applying different inspection criteria to different regions, the system maintains strict inspection only where necessary to detect actual defects, while using more lenient criteria in regions prone to pseudo defects. This local differentiation improves defect detection accuracy in critical areas without causing the widespread pseudo defects that would reduce overall productivity.
Solution Approach 2:
The patent changes the inspection parameter (threshold value) based on regional characteristics. By adjusting this parameter dynamically, the system achieves high defect detection accuracy in regions where it matters most while avoiding the generation of pseudo defects that would trigger unnecessary re-inspections and reduce productivity.
3Measurement precision
If region-specific inspection thresholds are implemented, then pseudo defects are reduced and detection accuracy is improved, but the inspection process complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining regional characteristics and corresponding inspection thresholds before the actual inspection process. This preparation work is done in advance, allowing the inspection system to simply apply the appropriate threshold based on the identified region, thereby reducing the complexity during the actual inspection while still achieving improved detection accuracy.
Data Source
AI summary
A pattern inspection apparatus which compares a plurality of partial optical image data of a same target plate to be inspected, the target plate being formed as a pattern, includes an optical image data acquiring unit which acquires optical image data of the target plate to be inspected, and a comparing circuit which compares the plurality of partial optical image data. In the comparing circuit, region image data generated based on information of region pattern representing a predetermined region is input, and a decision criterion is changed with reference to the input region image data when the plurality of partial optical image data are compared with each other to decide the presence/absence of a defect on the target plate.