矽晶柱氧化疊差（OISF）之自動視覺檢驗

陳佳玲(Chia-Ling Chen)

指導教授 ： 蔡篤銘

元智大學/工程學院/工業工程與管理學系/碩士(2001年)

摘要

目前半導體業在長晶（Crystal growth）製程中最常檢驗的是氧化造成的疊差（Oxidation Induce Stacking Fault, 簡稱OISF），基本上OISF現象可分為棒狀及半月形兩種，這兩種OISF現象主要是在長晶及矽晶柱後製程中產生。半導體業對於OISF影像中棒狀或半月形之個數密度量測（單位面積內垂直水平個數總合），一般是透過光學顯微儀器放大影像，對員工做特殊訓練後以人工目測求得，計算過程中消耗大量的人力資源與時間。利用自動視覺技術於OISF之檢測，可以減少人力資源與時間之浪費，且無人工作業疲勞所造成的誤判、標準不一的主觀因素，同時可提高檢驗品質、降低生產成本。
本研究針對客戶要求做特殊處理的OISF檢驗，採用傳統Sobel邊緣偵測運算子（Sobel operator）與全尺寸小波轉換（Wavelet frame）兩種方法來分離OISF中棒狀與半月形之垂直與水平邊緣，找出垂直水平能量值並求得垂直水平個數。對於一般OISF影像垂直水平個數之計算，其實驗誤差百分比在5%以內。對於高重疊OISF影像垂直水平個數之計算，因重疊部份為製程不可控制因素，而容易產生較高的實驗誤差百分比，因此本研究透過邊緣點數之臨界值設定，有效的選取影像中垂直水平邊緣重疊的部份，提高計算垂直水平個數的精確度，將實驗誤差百分比控制在3%以內。

關鍵字

矽晶柱氧化疊差 ； 自動視覺檢驗 ； Sobel邊緣偵測運算子 ； 全尺寸小波轉換

並列摘要

Oxidation Induce Stacking Fault (OISF) is frequently inspected in the crystal growing process of semiconductor industry. The appearance of OISF has “bar” and “half-moon” types that are mainly produced in crystal growth and silicon crystal post-manufacturing process. OISF image are generally enlarged by an optical electron microscope, and the density measure (total number of vertical and horizontal OISFs in a unit area) is manually obtained by a specially trained inspector. Number counting is a tiresome task, and very subjective from inspector to inspector. In this research we use machine vision to automatically compute the density measures of both “bar” and “half-moon” OISFs in a crystal surface. The inspection quality and time can be greatly improved with the automated visual inspection system.
This research uses Sobel edge operator and wavelet frame to separate vertical and horizontal edges of “bar” and “half-moon” OISFs in a image. The number of horizontal and vertical OISFs in each separated image can be easily evaluated accordingly. Experimental results have shown that the average error of the proposed method is smaller than 5%, which is competitive with human inspectors.

並列關鍵字

Oxidation Induce Stacking Fault ； Silicon Crystal ； Automatic visual inspection ； Sobel edge operator ； Wavelet frame

參考文獻

Ahmet, A. L., Ertuzun, A., and Ercil, A., 1998, “Texture defect detection using subband domain co-occurrence matrices”, Image Analysis and Interpretation, Vol.1, pp. 205-210.

Bin, W., Motomura, Y., and Ono, A., 1997, “Texture segmentation algorithm using multichannel wavelet frame”, IEEE International Conference on Computational Cybernetics and Simulation, Vol. 3, pp. 2527-2532.

Boles, W. W., and Boashash, B., 1998, “A human identification technique using images of the iris and wavelet transform”, Security and Detection, Vol. 46, No. 4, pp. 244-251.

Dash, W. C., 1958, “The growth of silicon crystals free from dislocations”, in Growth and Perfection of Crystals（R.H. Doremus, B.W. Roberts, and D. Turnbull, eds.） Wiley, New York, pp. 361-382.

Fournier, A., 1995, “Wavelet and their application in computer graphics”, SIGGRAPH95, British Columbia.