Reference Image Generation for Pattern Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In pattern inspection methods, uniformity of pattern modes on masks leads to difficulties in generating high-accuracy reference images, resulting in potential false defect detection during D-DB inspections.
Innovation Solution
A method that reads design data and characteristic information, divides the pattern formed region into regions with different characteristics, calculates parameter information, and generates a reference image using specific filter coefficients tailored to each region based on pattern type, density, and contour information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a set of filter coefficients common to the entire mask region is used to generate a reference image, then the generation process is simple and fast, but the accuracy of the reference image deteriorates when patterns have non-uniform modes
Solution Approach 1:
The mask region is divided into multiple sub-regions based on pattern density characteristics. Each sub-region is then processed with its own optimized filter coefficients rather than applying a single universal set of coefficients to the entire mask. This segmentation allows the reference image generation to adapt to local variations in pattern modes, significantly improving accuracy while maintaining computational efficiency through region-based processing.
Solution Approach 2:
Different filter coefficients are assigned to different regions of the mask based on their specific pattern characteristics, particularly pattern density. Each region receives a tailored filter configuration that optimizes the reference image generation for its local pattern mode. This local quality approach ensures that dense patterns, sparse patterns, and intermediate patterns each receive appropriate processing parameters, resolving the contradiction between universal simplicity and local accuracy.
2Productivity
If a reference image is generated using uniform filter coefficients for the entire mask, then processing time is reduced, but false defect detection increases due to reduced inspection sensitivity
Solution Approach 1:
The mask is segmented into multiple regions based on pattern density, with each region processed using optimized filter coefficients. This segmentation enables parallel processing of regions, maintaining high productivity while improving the accuracy of each local reference image. The region-based approach allows faster convergence to accurate reference images compared to iterative global optimization, thus preserving processing speed while enhancing defect detection sensitivity.
Solution Approach 2:
Filter coefficients are dynamically adjusted based on local pattern density parameters. Each region's filter coefficients are optimized according to its specific pattern characteristics, allowing the system to adapt processing parameters to local conditions. This parameter adaptation improves reference image accuracy and defect detection sensitivity without significantly increasing overall processing time, as the parameter selection is based on pre-calculated density metrics rather than iterative optimization.
Data Source
AI summary
To include reading design data of a plurality of patterns formed on a sample and characteristic information indicating characteristics of each of the patterns from a storage device, the characteristic information being additionally written in the design data, dividing a pattern formed region of the sample on which the patterns are formed, into a plurality of regions where the characteristics are different from each other on a basis of the characteristic information, calculating parameter information according to the characteristics with respect to each of the regions, where the parameter information is provided for generating a reference image from the design data to be used in an inspection of the patterns, and generating the reference image from the design data on a basis of the calculated parameter information.