2020-08-012024-05-17https://scholars.lib.ntu.edu.tw/handle/123456789/695793摘要：氮化鎵(GaN)高頻高電子遷移電晶體（HEMT)可提升電源轉換效率，及滿足下世代 5G通訊需求．然其高功率應用上仍有瓶頸，如平面結構所導致的散熱、電流坍塌 、寄生元件能量儲存等．射頻元件的漏電流和高頻損耗亦是問題。目前，高頻幾乎 都磊晶在SiC基板上，雖然高電阻值Si基板是另一選擇，但其自然彎曲問題限制了 基板尺寸與後續製程。 本計畫提出矽穿孔(TSV)技術來解決上述問題，除提高崩潰電壓與幫助散熱外 ，緩衝層蝕刻與TSV可抑制漏電流與電流坍塌，並因應力改變，可增加傳導電流 ．而透過重新設計TSV圖樣與磊晶結構，可提生射頻增益等性能。此外，藉由 TSV，我們將製作新穎的垂直式GaN HEMT結構，並討論其電流機制。 在三年執行期間，我們將研究GaN高功率、射頻、及新的垂直式元件之物理及電 特性。(1)高功率元件，我們把TSV與傳統“同電極平面”元件整合，來降低介面溫度 與通道阻抗。我們將透過TSV將源極移至基板底部，實現類似“垂直”式結構，因此 不需要閘極(或源極)的互連橋接，可有效的利用元件的面積。(2)射頻元件，我們將 挖掘數個孔以阻擋射頻訊號透過Si基板平行傳導，並進一步填充射頻絕緣物，來改 善功率增益與雜訊。(3)我們將提出一種新的垂直式HEMT結構，不需要p-GaN的 “再成長”，而是以直接成長的p-GaN/2DEG來截止通道電流。<br> Abstract: GaN electronics for power and RF (radio-frequency) applications have attracted tremendous attentions in recent years. The improvement of power conversion efficiency for power supplies and power modules/systems is emerging and becomes critical topics for government regulations. And the demand of higher bandwidth, RF power, and linearity for next 5G communications has driven GaN HEMTs (high-electron mobility transistors) toward immediate commercial applications. However, the development of GaN HEMTs for power electronics isn’t all rosy. Disadvantages, such as thermal dissipation due to planar structure, current collapse, energy storage in the parasitic capacitor and inductor components and the resulting thermal accumulation issues, have hindered the progress for commercialization. As for GaN RF devices, low leakage current and low high-frequency loss during operation are key issues. Currently, extremely high RF performance GaN HEMTs are almost all fabricated on the expansive SiC substrate. Even though low-cost high-resistivity Si substrate is an alternative choice, it suffers from a naturally bowing problem which limits the substrate size in diameter and subsequent process handling. We propose TSV (Through-Si-Via) technique to tackle both power and RF device problems while still maintain cost efficiency. TSV technique has already been proved in power devices that can increase breakdown voltage and thermal dissipation. By redesigning the TSV layouts and epistructure, we expect to improve RF performance as well. The idea of employing both buffer etching and substrate removal techniques simultaneously in this proposal is believed to suppress leakage current and current collapse. In addition, the removal of Si substrate inspires us to come up with some new idea of realizing vertical GaN HEMT structure. Thus we will design and fabricate a novel vertical HEMT structure in this proposal. In the 3-year project execution period, we will investigate device physics of GaN power devices, RF devices, vertical devices based on the Si substrate and buffer removal technology. (1) For power devices, we will be integrating TSV with conventional “source, gate and drain on top “ device structure for reducing junction temperature and channel resistance. We will relocate source contact to the bottom of the sample by TSV to realize a semi- “vertical” structure. Gate (or source) interconnect bridge will not be needed and effective chip size will be better utilized. (2) For RF devices, we will dig several TSV holes to block parallel RF conduction through the Si substrate. We will further fill RF insulting dielectric or polymer layer for improving power gain and noise floor. (3) We will propose a novel vertical HEMT structure on the Si substrate without p-GaN “RE-growth”. The p-GaN/2DEG is able to pinch-off the designated channel. Further improvement on having GaN substrate can avoid high resistivity buffer layer for better on-off and current level. We will be investigating the device physics of above three types of devices. The advantages are beyond what are currently studied in the related field in academia and industry.氮化鎵高頻高電子遷移電晶體矽穿孔GaNhigh electron mobility transistorsthrough silicon via