請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/22835
標題: | 應用於微影成像的一個有效的輪廓生成演算法 An Efficient Contour Generation Algorithm for Microlithography Aerial Image |
作者: | Szu-Kai Lin 林斯鍇 |
指導教授: | 陳中平(Chung-Ping Chen) |
關鍵字: | 光學微影,阿貝成像方法,成像模擬,輪廓(等高線),光學鄰近修正, optical microlithography,aerial image simulation,contour,OPC (optical proximity correction), |
出版年 : | 2010 |
學位: | 碩士 |
摘要: | 隨著超大型積體電路技術的特徵尺寸 (feature size) 迅速縮小,已小於曝光光源的波長,光的繞射效應使得曝光後的圖像 (pattern) 明顯地偏離了原本設計的光罩圖像。因此,光學鄰近效應 (OPE, optical proximity effect) 需要在光罩設計和修正時廣泛地考慮和模擬。
為了評估微影結果的品質,往往產生了大規模的成像,用來應用於如光學臨近修正 (OPC, optical proximity correction)、微影互通檢查 (LCC, lithography compliance check) 等仔細檢查的應用程序。用於二維成像中的像素數量,其複雜度是 O(n^2) ,其中 n 與成像解析度相關,這表示執行的複雜度在 n^2 時間。然而,可以很容易發現大多數的品質指標,例如臨界尺寸 (CD, critcal dimension) 或邊緣置放錯誤 (EPE, edge placement error),只使用到了輪廓,且在一個具體輪廓的像素數量一般為 O(n) 左右。因此,大多數微影成像模擬工具在計算時間和記憶體有一個巨大的浪費(至少為 O(n))。現在的問題是:「如何在沒有精確的成像的情況下而產生一個成像輪廓?」。 在這篇論文中,我們表明它確實是可行的,不需精確的成像而得知成像的輪廓。這個概念是表示輪廓在低解析度成像的可預測性。在我們的演算法中,我們先降低成像的解析度,得到一個大略的輪廓形狀,再以一個輪廓追蹤的演算法,得到精確的輪廓。另外,我們使用查表 (LUT, look-up table)的方法,先將光罩上的多邊形分解成數個矩形,藉由查表得到每個矩形的結果。最後,我們也對光罩圖樣做分割,這種方式進一步實現加速和平行計算。因此,當我們擁有成像強度的資訊時,就可以快速獲得成像輪廓。初步實驗表明了這個演算法的可行性。 As the VLSI technology feature sizes quickly shrink smaller than the wavelength of exposure light sources, the diffraction effects have made the exposed pattern significantly deviated from the original intended mask patterns. As a result, optical proximity effects need to be extensively considered and simulated during mask design and corrections. To evaluate the quality of microlithography result, massive aerial images are often generated for careful inspection using applications such as OPC (optical proximity correction), LCC (lithography compliance check). The number of the pixels used in a 2D aerial image is in the order of O(n^2), where n is the image resolution, which means the runtime scales in a n^2 fashion. However, most of the quality indexes such as CDs or EPE (edge placement error) can be readily observed using contours only and the number of pixels in a specific contour is around O(n) in general. Therefore, there is a huge waste (at least O(n)) of both computation time and memory in most microlithography aerial image simulation tools. The question is: 'how to compute an image contour without explicitly generate images? ' In this thesis, we show that it is indeed feasible to know the image contour without explicit images. The concept is to represent the contour can be predictable in low-resolution images. In our algorithm, we first reduce the imaging resolution, to get a rough shape of the contour, then use a contour tracing algorithm to refine the contours. In addition, we use the lookup table for convolution operation. This work is first to on decompose a polygon on the mask into several rectangles, and then look up the result for each rectangle. Finally, we also implement the mask pattern partition. This approach further to achieve speed up and parallel computing. Therefore, when we have the information of imaging intensity, we can quickly generate image contours. The initial experiment demonstrates that this algorithm is feasible. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/22835 |
全文授權: | 未授權 |
顯示於系所單位: | 電子工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-99-1.pdf 目前未授權公開取用 | 2.11 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。