- 學位論文
底膠製程優化與良率模型建立-以智慧型手機為例
Underfill Process Optimization and Yield Model for Smart Phone Application
摘要
因現今手持性電子產品功能需求性增加，亦需因應市場之輕、薄、短、小的趨勢，故半導體封裝技術方面日益複雜。封裝製程中，電子元件下方之錫球功能多為連接元件及基板或為訊號傳遞，但在實際應用之熱循環的過程中材料間會因不同的熱膨脹係數(Coefficient of Thermal Expansion； CTE)導致錫球發生熱疲勞現象而無法形成良好銲點，而底部填膠技術雖可保護錫球及減少元件所承受之熱應力，但通訊業者因應市場需求，需在手持性智慧型手機製造於表面黏著技術(Surface Mount Technology； SMT)製程中新增置放金屬蓋(metal shielding)之產品進行底部填膠，以減低電磁波，而此動作卻增加後續製程改善之繁雜性。 本研究針對電子通訊產品上之主要元件，首先利用特性要因圖探討影響氣泡、未填滿及溢膠等點膠品質之重要因子使用田口實驗設計進行製程參數優化，結果顯示最佳參數組合之總點膠量為74mg、間隔時間為0.35秒、點膠路徑為I-54637及板溫為85℃。再根據上述實驗所決定之最佳板溫(85℃)增減10℃條件下進行模擬膠材於晶片及PCB間的流速實驗，接著進行濕潤性實驗並進行流速修正，結果顯示，各溫度條件下膠材發生膠凝前流動距離平方與時間呈線性關係，且溫度高之流速最快。最後利用蒙地卡羅模擬方法(Monte Carlo Simulation)預測各溫度條件下完成填充所需時間及膠凝發生時間，以預測底部填膠製程良率，結果顯示板溫設定為85℃為最佳。
並列摘要
The trend of increasing miniaturization for handheld products has resulted in extreme complexity in the packaging and assembly technology. The solder balls located on the bottom side of component offers electrical connection. During service environment, the mismatch of Coefficient of Thermal Expansion (CTE) may result in fatigue failure of the solder interconnection. The encapsulation using underfill process help protect solder balls from damage. In the telecommunication industry, metal shielding is used to cover the electronics component to reduce the influence of electromagnetic wave. This further complicates the underfill process. This study investigates the critical factors that may influence the underfilling quality. Defects considered include voiding, uncompleted filling and overflow. The Taguchi based experimental design is used to optimize the process. Results show that the total dispensing volume of 74mg, awaiting time 0.35 seconds, dispensing pattern I-54637 and PCB temperature of 85℃ are desired. It is then followed by the flow experiments using glass slides to characterize the capillary flow of the underfill material. Results are corrected by the wetting performance of underfill on various flowing surfaces. Results show a linear dependency of the square of distance and the flow time. Finally, the Monte Carlo Simulation analysis is used to simulate and estimate the times to complete the flow and occurrence of gelling. This helps to predict process yields for different process temperatures.