Local maxima of wafer signature via clustering for metrology guided inspection

By using computer system analysis and machine learning models to predict defect density, combined with histogram equalization technology, the complexity of setting inspection parameters in semiconductor manufacturing has been solved, enabling more efficient and accurate defect detection and improving the performance of inspection tools.

CN120457335BActive Publication Date: 2026-06-26KLA CORP
View PDF 9 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
KLA CORP
Filing Date
2024-04-15
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

In the semiconductor manufacturing process, existing inspection tools are unable to effectively set inspection parameters based on the complex design and noise characteristics of samples, leading to difficulties in defect detection. This is especially true when sample design complexity increases and defect size decreases, making it difficult to predict the output of inspection tools and affecting the efficiency and accuracy of inspection formula setting.

Method used

The computer system analyzes the predicted defect density on the sample, performs bare die clustering based on the measurement results, generates initial and final bare die clusters, and stores this information to guide the inspection process. The machine learning model is used to predict the defect density distribution, and histogram equalization technology is combined to optimize color assignment to improve the detection effect.

Benefits of technology

It enables more accurate defect detection on samples and more efficient setting of inspection parameters, improving the efficiency and accuracy of the inspection process. In particular, it enhances the sensitivity and reliability of defect detection when the sample design is complex and the defect size is small.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120457335B_ABST
    Figure CN120457335B_ABST
Patent Text Reader

Abstract

Methods and systems for generating information for use in setting up a process performed on a specimen are provided. One method includes clustering dies on the specimen by color assigned to the dies based on predicted defect density in the dies determined in response to measurements performed on the specimen, thereby generating initial die clusters. The method also includes analyzing the initial die clusters in location space to determine whether any of the initial die clusters contain two or more die clusters. In addition, the method includes designating the initial die clusters that do not contain two or more die clusters, and the two or more die clusters contained in any of the initial die clusters, as final die clusters. The method further includes storing information for the final die clusters for use in setting up a process performed on the specimen.
Need to check novelty before this filing date? Find Prior Art