Semiconductor Hot Spot Defect Detection With Dynamic Threshold Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical inspection tools struggle to detect small defects in semiconductor manufacturing due to sensitivity limitations from wafer noise, as defect signals weaken with advancing technology, leading to gaps in defect detection.
Innovation Solution
A method and system for detecting hot spot defects using a dynamic mapping mechanism and machine learning to group and align defect images with a hot spot map, applying optimal filters and thresholding to enhance sensitivity and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical inspection tools are used to detect defects, then defect detection capability is provided, but sensitivity is limited by wafer noise
Solution Approach 1:
The patent segments the defect detection process into multiple stages: extracting local patterns (hot spots) from the semiconductor design, grouping similar hot spots together, and processing each group with optimized parameters. This segmentation allows the system to focus computational resources on detecting specific defect types while filtering out unrelated wafer noise, thereby improving sensitivity without being overwhelmed by overall noise levels.
Solution Approach 2:
The patent performs preliminary actions by extracting and grouping hot spots from the design before actual defect inspection. The hot spot map is pre-computed and stored, allowing the inspection tool to quickly reference known defect-prone areas. This preliminary preparation enables the system to enhance sensitivity for specific defect types while ignoring noise in areas where defects are not expected.
2Length of moving object
If defect size continues to decrease, then transistor size reduction is achieved, but defect signals become weaker than wafer noise
Solution Approach 1:
The patent applies local quality by extracting specific local patterns (hot spots) from the semiconductor design that are prone to defects. Instead of uniformly analyzing the entire wafer, the system identifies and focuses on specific local regions with characteristic patterns. This allows the inspection tool to optimize detection parameters for each local pattern type, enhancing sensitivity for detecting small defects in those specific regions while maintaining transistor size reduction.
3Measurement precision
If hot spot groups are extracted and grouped, then detection accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent merges similar hot spots into groups based on their local patterns. By combining multiple hot spots with identical or similar characteristics into single groups, the system reduces the total number of individual patterns that need to be processed during inspection. This merging strategy maintains high detection accuracy by preserving all relevant defect information while significantly reducing processing complexity through consolidation of redundant data.
Solution Approach 2:
The patent creates a universal hot spot map that can be applied across different semiconductor designs and inspection scenarios. The extracted and grouped hot spots form a reusable library of defect-prone patterns that enhances detection accuracy across multiple wafers and design iterations. This universal approach allows the system to maintain high accuracy without reprocessing the same pattern recognition logic repeatedly, thereby reducing overall processing complexity.
Data Source
AI summary
A hot spot defect detecting method and a hot spot defect detecting system are provided. In the method, hot spots are extracted from a design of a semiconductor product to define a hot spot map comprising hot spot groups, wherein local patterns in a same context of the design yielding a same image content are defined as a same hot spot group. During runtime, defect images obtained by an inspection tool performing hot scans on a wafer manufactured with the design are acquired and the hot spot map is aligned to each defect image to locate the hot spot groups. The hot spot defects in each defect image are detected by dynamically mapping the hot spot groups located in each defect image to a plurality of threshold regions and respectively performing automatic thresholding on pixel values of the hot spots of each hot spot group in the corresponding threshold region.