高电子迁移率晶体管
充电控制
材料科学
光电子学
波段图
接受者
兴奋剂
氮化镓
肖特基势垒
晶体管
肖特基二极管
拓扑(电路)
异质结
凝聚态物理
图层(电子)
电气工程
物理
电压
纳米技术
量子力学
功率(物理)
工程类
电池(电)
二极管
作者
Qianshu Wu,Jia Chen,Liang He,Jinwei Zhang,Qiuling Qiu,Chenliang Feng,Liuan Li,Taotao Que,Zhenxing Liu,Zhisheng Wu,Zhiyuan He,Yang Liu
标识
DOI:10.1109/ted.2021.3130848
摘要
The Schotty-type p-GaN gate high-electron-mobility transistors (HEMT) feature a unique gate structure. A comprehensive understanding of the charge control mechanism in the p-GaN gate region is a fundamental step for the optimization of this technology. In this work, a physics-based analytical model is presented which takes into consideration all the capacitive effects from gate metal deep into the GaN buffer. According to our analysis, the p-GaN layer can be either partially depleted by the metal/p-GaN Schottky junction or fully depleted, depending on the doping concentration and thickness of the p-GaN layer. Our model accurately captures the charge control properties under both conditions and is validated against TCAD numerical simulations. For a certain p-GaN thickness, a lightly doped p-GaN leads to a full-depletion condition, such that the acceptor concentration directly affects the band diagram at AlGaN/GaN interface. The ${V}_{\text {th}}$ of the HEMT increases quickly with acceptor concentration in p-GaN. With sufficiently high acceptor concentration in p-GaN, the device reaches the partial-depletion condition, the acceptor concentration loses its influence over the band diagram at the location of the AlGaN/GaN interface, since the Fermi-level at the AlGaN surface is pinned near the valence band of p-GaN. The ${V}_{\text {th}}$ starts to decrease with acceptor concentration, but at a relatively slow rate. The maximum ${V}_{\text {th}}$ is obtained near the boundary between partial-and full-depletion conditions. In consideration of the process margin, the device designed with a partially depleted p-GaN is preferable, since it ameliorated the ${V}_{\text {th}}$ sensibility against acceptor concentration.
科研通智能强力驱动
Strongly Powered by AbleSci AI