在目前全球半導體封裝製造專業代工的模式中,如何保持高品質與快速交貨將主導著業者的生存空間,進而支援客戶發展新產品並快速達到規模經濟,同時,降低生產製造成本並適時回饋顧客,如此將可創造雙贏的局面。客戶時常將製程尚未成熟之新產品交予代工製造商,冀望自己的產品能以最快的速度在市場銷售。新產品生產的過程中,必須將製程的限制與不確定性因子予以考量,並且,以最快速的方式導入大量生產。若一昧地以經驗值判斷製程參數,將可能導致因製程不穩定而造成訂單延遲,嚴重地影響到企業的商譽。因此新產品的製程最佳化顯的相當重要。 本研究針對銅導線晶圓封裝之銲線製程問題加以探討,並建立兩階段實驗設計模式,第一階段將考量影響銲線之因子並予以納入實驗範圍,以部分因子實驗法找出影響銅導線晶圓銲線之顯著因子。第二階段就影響銲線之顯著因子,以反應曲面法加上中央合成設計求解出銲線之最佳解組合。其中第一階段將可求得影響銲線之因子組合;第二階段將可求得各顯著因子之最佳參數值,並可求得所有納入考量影響銲線之因子最佳參數值。至此,後續之新產品導入將可依據本研究之規劃方法將製程做最適化之改善。
For the worldwide semiconductor packaging sub-contactor market, how to maintain high quality and short cycle time will be the key performance for each sub-contactor, and also need to support customer on new product develop for high volume mass production, meanwhile, continuing to provide cost down program to customer achieving Win-Win situation. Customer would like to provide the pro-type non-mutual product for process characterization and preproduction preparing, and hope the product could spread out and win the market share quickly. At this critical position, factory need to consider the process constraint and nonconformity factor, then release to high volume mass production in a very short time. If just to follow up parameters from the previous experience, there must be high risk to impact the business due to the nonconformity factor. From the above shows, how to optimize the process for the new product will be the key point. This research is focus on Cu wafer packaging wire bonding process, and to set up two steps Design of Experiment method model. The first step is just considering the all the factors that will inference wire bonding process, apply the Factional Factorial method to find out the effective factors. Step two: apply Response Surface Method with Central Composite Design to find out the optimize parameters. The goal for the first step is just finding the effective facts for process, the second step is being focus on the how to optimize effective factors. Finally, need to come out the all the optimize parameters. At this moment, the coming products could follow up this best practice from this research to improve or optimize the process.