- 學位論文
應用田口方法提升陣列製程之乾蝕刻良率
Enhancement of Dry Etching Yield for Array Processing Using Taguchi Method
摘要
由於臺灣上下游光電產業的垂直整合成功，帶動了液晶面板(TFT-LCD)產業的快速發展，LCD產業可說是台灣極具影響力的產業，不單其產值位居十大光電產品之首，相關項目累積投資額也接近數千億台幣以上。其可觀的經濟規模也促使政府在產業政策中，將面板產業與半導體產業列為「兩兆雙星」之林，企盼結合全國產業資源，再造另一次台灣經濟奇蹟。TFT LCD產品的要求朝向厚度薄、質量輕與高亮度等高品質邁進，高階產品著重的就是品質的穩定與良率的提升。 本研究對蝕刻製程中，因自發性發塵現象，所導致蝕刻良率降低的問題，利用田口式方法設計實驗參數，並找出少數關鍵的重要因子，分別為乾式幫浦背壓、上電極出氣孔大小、玻璃基板頂針材料及電漿聚集裝置與基板間之間隙。遂以重要因子後，再經田口實驗計畫後，尋找出適當反應之最佳設定水準組合，進行製程實驗，即得到最佳製程參數，可應用於生產之製程條件上。 本研究利用田口法實驗得知，最佳實驗參數是100(mT)的乾式幫浦背壓、上電極出氣孔大小是1.2(mm)、玻璃基板頂針材料需選用PBI、電漿聚集裝置與基板間之間隙是0.8(mm)。改善前Particle數量為單片30~70顆，分佈狀況為大量分佈型態；改善後Particle數量下降到20~50顆，分佈狀況改善成微量散佈形態，Particle實際改善率約為25（%）。 透過本研究可發現，傳統法實驗共進行23次的交叉比對才得到最佳實驗參數解，然田口法實驗只需要8次的實驗即可得到最佳實驗參數解。傳統法實驗其實驗成本高昂，且實驗過程冗長，過程耗時且浪費金錢費用，而實驗結果也不竟理想。然利用田口方法，實驗較簡易，且易於對參數作條件控制，有效節省調整參數的時間與成本，進而快速完成實驗。
並列摘要
In recent years, the supply chains in liquid crystal display (LCD) industry are flourished in Hsin-Chu city, Taiwan due to the successful integration of the opto-electronic industry. Nowadays not just it is one of the most significant industries in Taiwan but also its industrial value is the number one in the top ten of opto-electronic industries. The accumulative amount of relative investment is more than hundred billion Taiwanese dollars and the government has identified LCD and semiconductor industry as dual trillion-industries in Taiwan considering their enormous value. LCD companies and corporations are looking forward to another economic miracle by integrating and combining all resource of the industry. Recently the demand for TFT-LCD is growing because of the important role it plays in consumer electronic products such as cell phones and LCD TVs. People start to pay attention to the quality of LCD especially when this industry and market is competitive and quickly changing. TFT LCD products are tending to become thinner, lighter and brighter than before. The quality in the product and low deficient rate in the production are the keys for high-end products. This research describes the quality of substrate structure in the dry etching process and the two-method experiment design which is try-and-error solution and the Taguchi method. Then, the experiment analysis and error of data are mentioned in the paper, which also points out the research results. The result from the best experiment describes the amounts of particle are less by using the Taguchi method. By the result of Taguchi method, the best experiment parameter is Dry pump 100(mT), up electrode hold size is 1.2(mm)、glass lift pin materials use PBI、Clamp form glass gap is 0.8(mm).The original number of particles is 30-70 per, distribution condition improve form a great quantity to disseminate type. After improvement, distribution condition improve form small to disseminate type, Particle practice improve is 25（%）.The result of the following research shows that we can get the best parameters in 8 experiments by Taguchi method, which is better than 23 experiments by the old method. The old method has a longer process wasting more time and money. The result is worse, too. However, Taguchi method is easier. In other words, we are able to control experiment conditions easily according to the parameters. It spends less time and money. The experiment will be completed rapidly.