應用二維小波轉換檢測晶圓晶粒之可見瑕疵

Inspecting Visual Defects of Wafer Die by Using Two-Dimensional Wavelet Transform

紀偉龍(Wei-Lung Chi)

指導教授 ： 葉繼豪

國立臺北科技大學/管理學院/工業工程與管理系所/碩士(2005年)

摘要

本論文係利用機器視覺(Machine Vision)與二維小波轉換(Two-Dimensional Wavelet Transform；2-D WT)檢測晶圓之表面可見瑕疵，如：刮痕、微粒、污染及墨點，本論文以晶圓之1/20晶粒次區域影像為待測影像，先將晶粒次區域影像經平滑化與Sobel邊緣化處理後，再經由二維小波轉換的多解析分解可計算出二維小波群聚能量(Wavelet Transform Modulus Sum；WTMS)與小波能量群聚比值(α)。由於晶粒表面可見瑕疵及墨點之邊緣像素點在某相鄰階層間之小波能量群聚比值與正常影像差異甚多，故藉此可判定刮痕、微粒、污染或墨點之瑕疵像素點位置，再進一步利用最適橢圓法(Best Fitting Ellipse)求出這些瑕疵素點群之幾何特徵以分類微粒及污染、刮痕與墨點。實驗結果顯示，本論文所提出之晶圓瑕疵檢測方法在特定小波基底與多解析分解之特定階數下，能有效且精確偵測出表面可見瑕疵(微粒及污染、刮痕)及墨點之瑕疵像素點位置，後續之最適橢圓法計算得出的幾何特徵(長短軸比值與面積值)亦可初步區隔微粒及污染、刮痕與墨點。本研究提出之方法使得一待測完整晶粒影像僅需比對約20000個像素點即可，故可提昇檢測速度與節省儲存空間，期能協助目前人工檢測瓶頸與自動化光學檢測設備高成本之問題。

關鍵字

機器視覺 ； 晶圓晶粒 ； 可見瑕疵 ； 墨點 ； 二維小波轉換 ； 最適橢圓法

並列摘要

This paper develops an automatic optical inspection (AOI) system to inspect the visual defects such as particle, contamination, and scratch and dot on wafer die by using two-dimensional wavelet transform (2-D WT) and machine vision. The potential pixels for visual defects and dot can be precisely captured by the wavelet transform modulus sum (WTMS) and across-level ratio (α) on adjacent decomposition levels. Once the potential pixels are addressed, best fitting ellipse algorithms are utilized for calculating the geometric features such as the length of major axis, the length of minor axis, and area to classify particle and contamination, scratch, and dot on wafer die. Experiment results show that the proposed method is able to precisely capture visually defective and dot pixels on a wafer die. Moreover, it is initially feasible to classify pixel candidates into particle and contamination, scratch, or dot based on best fitting ellipse algorithms. Since the number of pixels requiring inspection on a die is around 20000 pixels, the inspection can be speed up and the capacity of this stage can be increased.

並列關鍵字

Machine vision ； Wafer die ； Visual defects ； Dot ； 2-D Wavelet Transform ； Best Fitting Ellipse

參考文獻

[5]R. T. Chin and C. A. Harlow, “Automated visual inspection: a survey”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol.4, No.6, 1982, pp.557-573.

[10]S. Kubota, N. Eguchi, and H. Masuda, “DUV lasers applied to semiconductor inspection and optical disk mastering”, Advanced Semiconductor Lasers and Applications/Ultraviolet and Blue Lasers and Their Applications/Ultralong Haul DWDM Transmission and Networking/WDM Components, 2001 Digest of the LEOS Summer Topical Meetings, pp.2-6, 2001.

[13]許文輝，區間神經網路應用於半導體晶圓缺陷辨識及良率估計之設計，元智大學，工業工程研究所碩士論文，1998。

[15]曾紀綱，應用機器視覺方法於晶圓表面瑕疵檢測之研究，元智大學工業工程與管理學系碩士論文，2003。

[16]S. H. Guan, P. Xie, and H. Li, “A golden-block-based self-refining scheme for repetitive patterned wafer inspections”, Machine Vision and Applications, Vol. 13, No. 5-6, pp.314-321, 2003.