Wafer Similarity Analysis Using Normalized Component Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wafer analysis methods rely on manual and qualitative similarity determination, leading to increased turnaround time and lack of objective, quantified standards, which can result in misjudgments and inadequate model consistency.
Innovation Solution
A method and system for analyzing wafers that involves obtaining measurement data, normalizing it, separating into components, and calculating similarity based on these components to provide an objective and quantitative comparison.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual and qualitative determination of wafer similarity is used, then user judgment flexibility is maintained, but turnaround time increases excessively and full evaluation becomes impossible
Solution Approach 1:
The patent replaces manual qualitative judgment with automated image processing and analysis systems. The system uses computer-based image acquisition, preprocessing, feature extraction, and similarity calculation algorithms to objectively compare wafer maps, eliminating the time-consuming manual review process while maintaining or improving determination accuracy through quantitative metrics.
Solution Approach 2:
The patent transforms subjective qualitative similarity assessment into objective quantitative parameter-based comparison. By extracting specific features from wafer maps (such as defect density, spatial distribution patterns, and statistical characteristics) and calculating numerical similarity scores, the system enables rapid automated evaluation that can process multiple wafers simultaneously, dramatically reducing turnaround time.
2Ease of operation
If manual similarity determination is used, then subjective judgment is allowed, but objective and quantified standards are absent leading to misjudgments
Solution Approach 1:
The patent replaces subjective human judgment with automated image processing and analysis systems. The system uses computer-based image acquisition, preprocessing, feature extraction, and similarity calculation algorithms to objectively compare wafer maps, eliminating the time-consuming manual review process while maintaining or improving determination accuracy through quantitative metrics.
Solution Approach 2:
The patent implements a systematic feedback mechanism where similarity metrics are calculated based on extracted features and compared against predefined thresholds or reference values. This quantitative feedback loop provides objective criteria for determining whether wafers meet similarity requirements, ensuring consistent and reliable model selection without relying on variable human judgment.
3Measurement precision
If comprehensive wafer evaluation is performed manually, then thorough analysis is possible, but the process becomes excessively time-consuming and impractical
Solution Approach 1:
The patent divides the comprehensive wafer evaluation process into distinct automated segments: image acquisition, preprocessing (noise reduction, normalization), feature extraction (defect detection, pattern recognition), similarity calculation, and result output. This segmentation allows each step to be optimized and executed automatically, enabling thorough multi-dimensional analysis of wafer maps without manual intervention, thereby maintaining evaluation completeness while dramatically increasing analysis throughput.
Solution Approach 2:
The patent implements continuous automated processing where wafer maps are systematically acquired, analyzed, and compared in an uninterrupted sequence. The system maintains continuous operation by automatically transitioning between processing steps and handling multiple wafers in succession, eliminating the interruptions and delays inherent in manual evaluation, thus achieving both comprehensive analysis and high productivity.
Data Source
AI summary
A method for analyzing a wafer includes acquiring first measurement data for a first wafer and second measurement data for a second wafer, generating normalization data including first normalization data and second normalization data obtained by scaling the first measurement data and the second measurement data, respectively, separating each of the first normalization data and the second normalization data into at least one component to generate component data including first component data and second component data, and outputting a similarity of the first wafer and the second wafer calculated based on the component data.