軟性薄膜電晶體技術在近年有越來越多的應用。軟性薄膜電晶體技術相較於傳統的矽技術有許多優點,例如較低的成本、較短的製作時間,以及其可撓的特性。目前軟性薄膜電晶體技術有三個較大的問題:(1) 製程中臨界電壓的變異、(2) 電路老化造成臨界電壓的改變、(3) 電路受到撓曲時漂移率的改變。電路設計者在設計電路時無法有效的考慮各種效應,但以上所提及的效應對電路良率有極大的影響。因此此篇論文提出一個軟性薄膜電晶體類比電路良率最佳化的自動化軟體,考慮以上所提的三種主要效應。此良率最佳化軟體所提出的演算法是建立在反應曲面法(response surface methodology)之上,並利用直交表(orthogonal array)判斷模擬的重要變數。本篇論文使用8微米a-Si與10微米IGZO薄膜電晶體製程下四種不同的有機發光二極體驅動電路,以及差動放大器,並且考慮兩種不同的材料技術。實驗結果顯示此最佳化軟體分別提升了a-Si與IGZO技術發光二極體驅動電路6.8%、12.0%的良率。此最佳化軟體提供使用者電路中各電晶體寬度的組合以達到在老化、撓曲,以及製程變異的情況下最佳的良率。
Flexible thin-film transistor (TFT) technology is widely used in recent years. Flexible TFT technology has many advantages over conventional silicon technology such as low cost, short manufacturing time, and flexibility. Flexible TFT technology has three important problem that may cause yield loss of flexible TFT circuits. They are 1) Process variation in threshold voltage, 2) aging effect results in threshold voltage shift, and 3) bending effect result in mobility change. It is hard for designer to consider all the effects when designing the flexible TFT circuits. Thus, this thesis proposes a yield optimization automation tool for analog flexible TFT circuits, considering the above three effects. Response surface methodology based optimization flow is proposed in this thesis, using orthogonal array to perform the screening experiment to identify important variables. Organic light-emitting diode (OLED) drivers and operational amplifier (OPAMP) using 8μm amorphous-silicon TFT technology and 10 μm Indium-Gallium-Zinc Oxide TFT technology are demonstrated. Experimental results show that this tool can promote the yield of a-Si and IGZO OLED driver by average 6.8% and 12.0%, respectively.