Ferroelectric devices for advanced computing applications

Ryu, Hojoon

Permalink

https://hdl.handle.net/2142/117679 Copy

Description

Title

Ferroelectric devices for advanced computing applications

Author(s)

Ryu, Hojoon

Issue Date

2022-12-02

Director of Research (if dissertation) or Advisor (if thesis)

Zhu, Wenjuan

Doctoral Committee Chair(s)

Zhu, Wenjuan

Committee Member(s)

• Lyding, Joseph
• Zhao, Yang
• Rakheja, Shaloo

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• FERROELECTRIC DEVICES
• ADVANCED COMPUTING APPLICATIONS

Abstract

Ferroelectric materials offer a wide range of useful properties, including ferroelectric hysteresis, high permittivity, strong piezoelectric effects, high pyroelectric coefficients and strong electro-optic effects, which make them promising candidates for nonvolatile memories, capacitors, sensors, actuators and detectors. However, traditional ferroelectric materials such as perovskite oxides have a limitation in thickness scaling and are not compatible with CMOS processes. This dissertation investigates emerging ferroelectric materials, including doped hafnium oxide and van der Waals (vdW) ferroelectrics, which can retain ferroelectricity down to nanometer scale and can be seamlessly integrated with any substrates. Novel ferroelectric devices, including ferroelectric tunneling junctions (FTJs), ferroelectric high electron mobility transistors (Fe-HEMTs), and ferroelectric synaptic devices are also developed based on these ferroelectric materials. In this project, we systematically study the synthesis of ferroelectric doped HfO2 using atomic layer deposition (ALD). Various process conditions, including aluminum concentration, annealing temperature/time, and metal electrodes, are investigated. Based on the optimized process conditions, we demonstrate high-quality ferroelectric Al-doped HfO2 with remanent polarization up to 20 µC/cm2, endurance higher than 108 cycles, and retention time longer than 10 years. These materials will be very promising for non-volatile memory applications. Utilizing the high-quality ferroelectric doped HfO2, we develop a new type of FTJ based on dielectric/ferroelectric heterostructure. The interfacial Al2O3 layer and the semiconducting substrate enable sizable tunneling electroresistance (TER) ratios even when the thickness of ferroelectric layer is above 10 nm. We demonstrate FTJ synapses with symmetric potentiation and depression characteristics and widely tunable conductance. We also show that spike-timing-dependent plasticity (STDP) can be harnessed from the FTJs based on Zr-doped HfO2. These novel FTJs will have high potential in neuromorphic computing. Furthermore, we demonstrate Fe-HEMTs based on ferroelectric HfO2 on AlGaN/GaN substrates for high-temperature memory applications. The high curie temperature of doped HfO2 and the wide band gap of GaN channel enable the transistor to operate at high temperatures. Our Fe-HEMTs on AlGaN/GaN structures show stable operation up to 590 ℃ with a record high Ion/Ioff ratio (6.6x106) and good transconductance (20 mS/mm). Our FE-HEMT shows clear counterclockwise hysteresis at both room temperature and 200 ℃, which is consistent with the ferroelectric switching in HZO. Finally, we demonstrate low-thermal-budget FeFETs based on vdW CuInP2S6 (CIPS) and indium-zinc-oxide (IZO). The CIPS/IZO FeFET shows a non-volatile memory operation with a low IOFF current and high carrier mobility. The CIPS/IZO FeFETs exhibit high dynamic ratios (around 125), which are promising for neural network applications. The single-perceptron neural network based on CIPS/IZO FeFETs shows a recognition accuracy of up to 80 %. These new ferroelectric materials and devices will have high potential in both energy-efficient non-volatile memories and artificial neural networks. These ultra-low power electronics will have a wide range of emerging applications spanning mobile and wearable electronics, medical implant devices, signal processing, and image recognition systems, as well as computing systems in large data centers.

Graduation Semester

2022-12

Type of Resource

Thesis

Copyright and License Information

Copyright 2022 Hojoon Ryu

Owning Collections