應用失效模式與效應分析建構製程品質改善系統--以半導體產業為例

Applying FMEA to construct process quality improvement system--an example of semiconductor industry

張耀文

指導教授 ： 黃雪玲

國立清華大學/工學院/工業工程與工程管理學系/碩士(2014年)

摘要

台灣的半導體製造產業發展至今已至成熟期，各家業者的製程能力也相去不遠，產品的多樣性與製造流程的複雜化，數百道製造流程已無法容許任何一道製程步驟異常或人為疏失，稍有不慎就可能造成上百萬損失甚至影響公司聲譽，因此製程品質的控管就更加重要。面對產業國際化的競爭，建立一套適合半導體產業在製程品質方面的風險管理與監控機制，是企業面對這競爭激烈環境下的解決方案。
由前人發展的許多論證與案例來看，失效模式與效應分析技術已運用在各產業，而半導體產業在製程品質改善的部分有很多方法，由各方法比較得知，以 FMEA 技術最為適合應用在製程繁複的半導體產業。本文研究方法應用 FMEA 技術以製造流程步驟為基礎，定義各步驟的製程功能及目的，列出各製程站之潛在失效問題及原因，並分析失效效應之嚴重性、發生頻率及偵測的難易度，依評估分級表計算風險優先指數，若未達標準，提出改善計畫及完成日期，改善計畫執行後，需重新評估風險優先指數，直到風險優先指數達到標準。
本研究運用 FMEA 系統架構並配合其他品質改善手法，解決個案某產品高壓製程良率問題，歸納其主要結論如下：
1).訂定風險優先指數評估分級表，架構關鍵製程的風險管控的機制，同時提供一份完整的製程技術知識平台，也可作為新產品或製程快速導入量產之依據。
2).建立製程 FMEA 系統之品質改善流程，作為公司執行製程 FMEA 之運作規範。
3).解決個案製程良率問題，降低了因失效效應所造成的損害，亦提昇生產效率和產品良率。

關鍵字

失效模式與效應分析 ； 半導體製程 ； 品質管理

並列摘要

Taiwan's semiconductor manufacturing industry has been in maturity, each company's process capability is not so different, product diversity and complexity of manufactory processes, and hundreds of steps manufacturing processes are not allowed any one process step abnormalities or human error. The slightest mistake could result in the loss of millions and even affect the company's reputation. Therefore, the quality control process is very important. One of the solutions for companies facing fierce competition environment is to establish risk management and control mechanisms in the semiconductor industry.
Based on literature review, FMEA technique has been used in various industries. Among the many ways in the semiconductor industry for process quality improvement, FMEA techniques are best suited to applications in the complex manufacturing process of the semiconductor industry. In this study, FMEA techniques were applied to define the function and purpose of each manufacturing process, to list of the potential failure modes and root cause, and to analyze the severity of failure, frequency of occurrence and detection of difficulty. Then risk priority number (RPN) was calculated by assessment classification table, and improvement plan and completion date were proposed. After the implementation of the improvement plan, the risk priority index was reevaluated until it reached a standard level. Finally, the case study revealed the improvement of process quality.
In this study, FMEA was applied and quality improvement techniques were implemented to solve cases High Voltage process yield issues, and the main conclusions are summarized as follows:
1).Providing a process technology knowledge platform as a basis for a new product or process of rapid mass production.
2).Establishing quality improvement process as the company’s operation guidelines for process FMEA.
3).Solving process yield cases, and improving production efficiency and product yield.

並列關鍵字

Failure Modes and Effects Analysis ； Semiconductor process ； Quality control

參考文獻

曾耀群(2009)，應用醫療照護之失效模式與效應分析於醫療流程之改善，國立清華大學工業工程與工程管理學系碩士論文.

陳建福(2013)，應用萃思法於製程可靠性測試之改善，國立清華大學工業工程與工程管理學系在職專班碩士論文.

郭展銓(2011)，失效模式效應分析（FMEA）在扣件熱處理製程上之應用，品質月刊.

王嘉麟(2006)，運用失效模式與影響分析評估矽甲烷供應系統支安全性-以 TFT-LCD廠為例，國立交通大學工學院產業安全與防災學程碩士論文.

白志超(2007)，DRAM廠衍生產品製程開發之品質規劃─以QFD與FMEA展開，國立清華大學工業工程與工程管理學系.