標題: 田口方法於半導體封裝製程改善之應用-以焊線站之焊線強度最佳化為例
Application of Taguchi Methods to Improve the Semi-conductor Packaging Process - A Case Study for Wire Strength Optimization of Wire Bonding Process
作者: 張華基
Nathan Chang
蘇朝墩
Dr. Chao-Ton Su
管理學院工業工程與管理學程
關鍵字: 積體電路;拉力測試;參數設計;直交表;信號雜音比;田口方法;Integrated Circuit;Wire Pull Test;Parameter Design;Orthogonal Array;Signal to Noise Ratio;Taguchi Methods
公開日期: 2001
摘要: 半導體封裝製程的主要目的,是將積體電路(Integrated Circuit,IC )晶片上的線路引接至外部之導線架,並覆蓋上保護晶片線路之膠質,確保IC功能正常。在封裝製程中,焊線站是擔任串接內外部線路(晶片與導線架)的任務,因此,假如焊線強度不足,會導致後續的封膠作業將焊線沖斷或受損,或者在後段PCBA(PC Board Assembly)組裝過程中造成可靠度(Reliability)的問題,不但破壞了工廠的品質形象,影響公司的商譽,甚至還會遭致客戶索賠及後續訂單的損失。 本論文應用田口玄一(Genichi Taguchi)博士所發展的參數設計(Parameter Design)方法,找出焊線站作業最佳化的條件設定參數,使焊線強度可以穩定的改善,而不會在後續製程發生斷線或受損現象。針對焊線製程改善問題,本論文以焊線拉力測試(Wire Pull Test)強度為品質特性,配合直交表(Orthogonal Arrays)的配置,並利用望大特性(Larger-The-Better,LTB)信號雜音比(Signal-to-Noise Ratio,SN比)進行分析。研究結果顯示,此分析方法可成功獲得最佳參數設定,經最後進行確認實驗後,確認加法模型在本實驗中是合宜的,並穩定的改善了焊線強度。
The main purpose of semi-conductor packaging process is to ensure workable functionality and protect IC (Integrated Circuit) chip by connecting the electrical circuits of IC chip with external lead frame and injecting the molding compound. In packaging process, the operation of wire bonding is to link the internal/external electrical circuits between IC chip and lead frame. The bonding wire is possibly be broken or damaged during the following molding process if the wire strength is not strong enough. Furthermore, this defect will cause reliability concern in back-end PCB assembly process. Such defect of weak wire strength is not only damaged the quality image of packaging manufacturer, but also customers raised claim on this defect sometimes. Moreover, the afterward business orders will be deducted or cancelled. The firm applied Taguchi’s approach to the wire bonding process, so as to improve the wire strength in short time. The quality characteristic was decided as reading value by “Wire Pull Test”. The study was also analyzed along with appropriate orthogonal arrays to accelerate and simplify the process of experiment. Based on Larger-The-Better, Signal-to-Noise ratio (SN ratio) data to analyze, the result shows the method result in the optimal parameter design. Furthermore, the additive model is appropriate in the experiment by means of confirmation experiment. The proposed optimal conditions do improve the wire strength of wire bonding process.
URI: http://140.113.39.130/cdrfb3/record/nctu/#NT901031002
http://hdl.handle.net/11536/69567
顯示於類別:畢業論文