Model Refinement for the Classifier of the Spatial Pattern Randomness

• Main Authors: Ruei-Syuan Hou, 侯睿軒
• Other Authors: Chin Hsia
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/9w48yv

碩士 === 國立中央大學 === 電機工程學系 === 107 === In this thesis, we find the relationship between diesize and Boomerang Chart based on wafer map which random distribution of defects, build the model by the relationship that let user just provided diesize to get bound of wafer cause by random defects to achieve the aim of discrimination abnormal wafers fast. At first, we use the largest diesize to simulate bound accurately and carefully. We chose two linear regressions for this diesize. And by adjusting the offset of the center line of the bar graph with the wafer size, the center point at different sizes is corrected to prevent the center point deviation in the randomness analysis, and the offset of the center line for each diesize has a special relationship. Additionally, the standard deviation value is re-adjusted to meet the desired confidence interval percentage. Model based on the two factors to let user get bound of Boomerang Chart fast for full yield range, wider diesize and accurately has been provided. Then, we generate synthetic wafer by random number. Increasing the original standard deviation of 1.96 times (95% confidence interval) to 2.58 times (99% confidence interval) and 3.89 times (99.99% confidence interval). We use full-range analysis to enhance the verification of randomness. And then verify the normality of normalized NBD to observe whether distribution of B-score meets to our critical points. By these steps, verified B-score is a standard score. At last, the results of the previous thesis are compared with our method and our error function is used to quantify the difference between these to show the extent of improvement in this thesis.