Title: 利用實驗設計法改善廢氣處理設備排放濃度 最佳參數設定
Using Design of Experiments to Determine Local Scrubber Optimal Parameter Setting
Authors: 楊東霖
唐麗英
Yang, Tung-Lin
Tong, Lee-Ing
管理學院工業工程與管理學程
Keywords: 實驗設計;變異數分析;experimental design;analysis of variance
Issue Date: 2017
Abstract: 由於半導體業的蓬勃發展,大量的化學物質被應用於半導體業上,而CVD化學薄膜製程一直是半導體廠中一道重要的製程,被用來沉積出某種薄膜(film) 的技術,而此道製程必須使用大量NF3 gas clean chamber,NF3在半導體業是被定義必須處理的氣體,而處理方式為利用local scrubber作為污染防治設備。但半導體業遇到另一個問題,在廢氣處理設備處理NF3過程反應會產生F2副產物而最終將經由煙囪排放到周界環境,F2不僅為毒性氣體更會產生惡臭其嗅覺閥值為0.1ppm,因此我們在半導體廠區周圍常常被反應異味時此F2就是主要兇手之一。 為了協助A公司解決local scrubber outlet & stack F2排放問題,本研究針對電熱式local scrubber處理NF3過程影響去除效率因素,並利用2K-p部分因子實驗設計法規劃實驗,以變異數分析分析實驗數據,找到關鍵因子之後再進行3K全因子實驗數據得到最佳因子水準組合設定,能又效的降低local scrubber outlet F2進而使煙囪排放濃度降低使異味問題減少,此效果對環境有更大的保障並且降低這些毒性氣體排放到環境影響人體健康。本論文以台灣某半導體廠為例(A公司)提供之G廠牌local scrubber安裝於CVD區域的污染防治設備做為實驗機台,驗證了本研究方法確實有效可行,本研究方法將可提供給其他公司做為參考學習的範例,並也期望應用於不同process local scrubber。
Due to the flourishing development of the semiconductor industry, a large number of chemical substances are used in the semiconductor industry. CVD chemical film process has been an important process in the semiconductor factory, which is a technology to deposit some kind of film. The CVD chemical film process uses a lot of NF3 gas to clean its chamber and NF3 is defined as must deal with gas in the semiconductor industry. The method of NF3 treatment is using local scrubbers as the pollution control equipment, but there occurs another problem. F2 by-products are produced during the treatment of NF3, and these by-products will eventually be discharged into the perimeter environment via the stacks. F2 is not only a toxic gas, but also produces stench with an olfactory threshold of 0.1ppm, which makes it one of the main causes why the semiconductor factory is often reflected for the odor around it. In order to assist Company A to solve the problem of F2 emission of local scrubber outlet and stack, this study aims at the factors that affected the removal efficiency of NF3 in thermal type local scrubbers. Experiments were carried out by using 2K-p part factor experimental design method, and the experimental data were analyzed by variance analysis to find the key factors. Through the 3K factor experimental data analysis of the key factors can find the optimum combination of factor level, which can effectively reduce the local scrubber outlet F2 value and thus lower the stack emission concentration to improve the odor problem. This effect has greater environmental protection and avoids the release of these toxic gases into the environment to affects human health. In this paper, the G-brand local scrubber, provided by Taiwan Semiconductor Factory Company A and installed in the CVD area as a pollution control equipment, is used as an experimental object to verify that this research method is effective and feasible. The proposed method can also be provided to other companies as a reference and learning examples, and expect to apply to different process of local scrubbers.
URI: http://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT070263315
http://hdl.handle.net/11536/142377
Appears in Collections:Thesis