| 摘要: | 隨著半導體、LED、太陽能等產業蓬勃發展，產品應用越來越廣泛，在全球的需求下，促使企業不斷擴建新廠區，提升產能，滿足市場需求。但隨了需求量提升，矽原料的消耗量上升也相對的快速提升成本，若不以永續發展的角度發展下去，此企業很快的不能負荷沉重的成本壓力，將會對企業帶來嚴重的後果。
本論文在進行產能規劃，需考量生產晶圓時，各個製程都必須得監控，確認製程之穩定、良率之確保，才能生產出客戶所需要之產品，因此在滿足市場需求前提下，就必須實際考慮生產晶圓時所需投入測試晶圓之成本，透過測試晶圓的再生，達到節本節省、再製造之理想，而測試晶圓又稱再生晶圓，再生晶圓的使用時機：IC晶圓造製場中的控片(Control wafer)及擋片(Dummy wafer)的使用時機為監控機台參數、當機維護、預防保養、機台熱機與條件化及製程改善實驗使用等。主要目的為及時偵測機台製程參數和維持加工環境，持續掌控為維持良率之必須工具。再生晶圓除了厚度因研磨拋光而變薄之外與原始晶圓並無太大的品質差異，使用再生晶圓作為空檔片可以減少晶圓廠對晶圓新片的消耗，有助於降低成本。
論文內容將探討企業如何在有限時間及產能條件下，充份運用現有產能並進行生產晶圓之成本控制等相關問題，並以混合整數規劃求出利潤最大化及最佳的產能調度，並依求解結果予以分析探討，提供面臨類似情況的企業做產能總合及再生晶圓規劃之重要參考依據。最後以三期再生晶圓的回收再利用的方式作為監控的依據，並可節省企業相當大的成本，其成本高達4,301千美元。
;As the semiconductor, LED, solar energy and other industries growing up, enterprises encouraged to expand of the new plant, and increasing production capacity to meet market demand. But with the demand increasing, consumption of raw materials rose relatively quickly make costs higher, but if the enterprise do not development in sustainable angle, it can’t take such fast and heavy cost pressure.
The paper explores capacity planning when producing wafers, each process must have to monitored, verifying the process, and keeping yield rate, in order to produce products for customers and reach the market demands, we need to consider the cost which come from monitoring wafer when it had produced. The use of reclaim wafer timing：Control wafer and Dummy wafer used to monitor the machine parameters, crash maintenance, preventive maintenance, heating the machine and getting conditional experiment and process improvement, etc. The main purposes are monitoring machine process parameters, maintaining process of environment and the most important to maintaining yield instruments. In addition to reclaim wafer polishing and thinning, the thickness than the original wafer is not much difference, as a reclaim wafer used in recycle can reduce the cost and consumption of new wafer.
The paper will fill the demands when explore the companies in the limited time and capacity conditions. Using mixed integer programming to maximize profits and find the best production scheduling, and analyze and discuss the consequence for enterprises to do aggregate capacity planning and regeneration wafers. |
