銅導線與鋁合金銲墊銲接品質最佳化之研究

Abstract

導線銲接(Wire bond)在半導體封裝中，長久以來即扮演著將晶片與基板間訊號互連(Interconnect)的重要技術，此技術也已發展超過30年的時間。以往大多是使用「純金」的導線和鋁導線，作為主要的使用材料，純金與鋁料具備穩定的物理特性，較不易氧化與極佳的延展性等，因此，純金與鋁料過去均為半導體封裝製程中，銲接導線選用的最佳材料。但隨著黃金價格不斷地攀升，在過去的20年間飆漲約62%，這對於整個半導體封裝產業中的導線銲接，無疑是一項成本上的大挑戰。由於電子產品的功能日新月異，對於電性上的要求也越來越講究，因此，銅線優秀的電性特性及其機械性質，使得銅線被提出來研究，作為替代金線的材料；銅材料除具備較佳的導電性、熱傳導性及低阻抗之外，亦具備較慢的合金層生長速度(Inter-metallic compound growth)，最重要的是，銅線的成本較金線少90%。 然而，以銅線替代金線亦有其缺點，例如硬度較高，以及容易氧化等特性。也由於此特性的存在，使得在銲接製程上造成一些工程上的問題；使用銅線取代金線於半導體封裝製程中，將對銲墊下的電子線路的衝擊相對大很多。本研究主要針對20µm的鍍鈀(Pd)線，配合不同厚度的鋁銲墊(Bond pad) ，試著利用田口方法，配合統計軟體(JMP)進行分析，透過拉力與推球測試、OM以及SEM檢查，找出影響銅導線銲接的主要參數，並擷取銅導線與鋁合金銲墊銲接的最適化參數組合。

Wire bonding is major inter-connect technology to connect chips to substrate in electronic package industry. Traditionally, Au & Al are preferred material for wire bond process due to inherent advantage in material properties. Stable material physical properties(good oxidation prevention、good malleability and elongation). Above specific material properties keep Au material at industry as key material. Cu wire bonding is the trend in electronic package due to the cost and excellent electrical and physical performance compare with Au wire. However, Cu wire bonding still existing many challenges, one of them is bond pad damage during bonding process. High material hardness induce this kind of issue. Another one is pad splash of AlCu pad after Cu wire bonding, it squeezes out the pad-ring area, that is severe case can cause short problem on electric circuit. This study is focus on Cu wire bonding process used 80nm technology IC device with two kind of thickness of AlCu pad and try to understand the effect of bonding parameters CV, USG, Force & Time on bond pad, a detail experiment is conducted with JMP software and to setup two steps Design of Experiment method model. First step, to consider the all factors that will inference wire bonding process, apply the Factional Factorial method to find out the effective factors. Second step, apply 「Response Surface Method Design」 to find out the optimize parameters.