Please use this identifier to cite or link to this item:
Title: 鐵電氧化鉿鋯之負電容效應及類神經元件應用
Ferroelectric HfZrO2 for Negative Capacitance and Neuromorphic Device Applications
Authors: 李敏鴻
Lee, Min-Hung
Siang, Guo-Yu
Keywords: 鐵電材料
Ferroelectric materials
deep learning
Issue Date: 2019
Abstract: 鐵電材料的遲滯現象(Hysteresis)具有雙穩態的特性,滿足記憶體對於信號的存取要求和負電容特性(Negative capacitance, NC)電壓放大的概念,因此近年來對於鐵電材料進行廣泛的研究。由於負電容特性改善次臨界擺幅(subthreshold swing, SS),使MOSFET的SS在室溫下克服Boltzmann tyranny 2.3kbT/decade的物理極限,另一方面具有穩定遲滯現象和非破壞性讀取的特性適合作為非揮發性記憶體(Non-Volatile Memory, NVM)。 本論文將針對鐵電材料氧化鉿鋯(HfZrO2, HZO)作為元件絕緣層的特性進行研究,首先將研究環繞式閘極場效電晶體搭載鐵電薄膜後,達到負電容效應,再來使用鐵電材料與非揮發性記憶體結合,研究應用於深度學習(Deep Learning, DL)且搭配不同結構與波型,尋找最佳的資料演算方式。
Bi-stable state nature feature of hysteresis loops by ferroelectric materials satisfies the demands of storage signal purpose for memory and voltage amplification concept for negative capacitance. Therefore, it has been extensively investigated in recent years. Benefiting from negative capacitance effect, subthreshold swing (SS) demonstrated with improvement on to overcome the physical limitation of Boltzmann tyranny 2.3kbT/decade for MOSFET at room temperature. On the other hand, the property of hysteresis loops and non-destructive reading are suitable as Non-Volatile Memory (NVM). In this research, we will study the characteristics of ferroelectric material HZO as the dielectric layer of the device. Firstly, we will study the GAAFET and carry up the ferroelectric thin film HZO to achieve a negative capacitance effect, and then use ferroelectric materials in combination with NVM, and apply it to Deep Learning (DL) with different structures and waveforms to find the best data calculation method.
Other Identifiers: G060648013S
Appears in Collections:學位論文

Files in This Item:
There are no files associated with this item.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.