Weibull Defect Time Grouping for TDDB Wafer Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods struggle to accurately classify and distribute defect times for various defect characteristics in semiconductor wafers, particularly for time-dependent dielectric breakdown (TDDB) defects, making it difficult to assess defect rates and types effectively.
Innovation Solution
The proposed solution involves a distribution output device that classifies defect times into data groups, calculates likelihood summations, determines optimal population parameters, and outputs a Weibull distribution for each group, allowing for a more accurate assessment of defect characteristics and vulnerability analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If defect times are classified into multiple data groups using Weibull distribution, then measurement precision of defect characteristics is improved, but device complexity increases
Solution Approach 1:
The patent segments defect time data into multiple data groups based on different defect characteristics (e.g., extrinsic defects, intrinsic defects, robust intrinsic defects). Each data group is analyzed separately using Weibull distribution to determine defect rates and vulnerability. This segmentation enables precise classification of defect characteristics while managing complexity through systematic grouping of similar defect types.
2Reliability
If likelihood summation calculations are performed for multiple data groups, then reliability assessment is improved, but loss of time increases
Solution Approach 1:
The patent performs preliminary classification of defect times into data groups based on defect characteristics before conducting likelihood summation calculations. By pre-organizing data according to defect types (extrinsic, intrinsic, robust intrinsic), the system reduces the computational complexity of subsequent reliability calculations and enables more efficient processing of multiple data groups.
3Measurement precision
If optimal population parameters are determined for each data group, then measurement precision of defect distribution is improved, but device complexity increases
Solution Approach 1:
The patent applies local quality by determining optimal population parameters (shape parameter and scale parameter) specifically for each data group rather than using a single set of parameters for all defects. This allows the defect distribution to be characterized with high precision for each defect type (extrinsic, intrinsic, robust intrinsic), capturing the unique statistical properties of each category while maintaining a systematic approach to parameter determination.
Data Source
AI summary
An operating method for a distribution output device includes; generating data number sets for data groups, grouping defect times according to an order in which the corresponding defects occurred in relation to each of the data number sets, calculating likelihood summations respectively corresponding to the data number sets in relation to defect times grouped in accordance with the data number sets, determining a maximum likelihood summation among the likelihood summations, determining optimal population parameter data for each of the data groups in relation to the maximum likelihood summation, and outputting a Weibull distribution for each of the data groups in relation to the optimal population parameter data for each of the data groups.