Synaptic and Sensor Applications of Amorphous Oxide Thin-Film Transistors with Plasma Surface Process

다국어 초록 (Multilingual Abstract)

This study investigates amorphous oxide semiconductor based thin-film transistors (TFTs) and proposes various device applications utilizing their electrical characteristics. The fabricated TFTs were analyzed in terms of their chemical and electrical properties to evaluate performance enhancement and their potential as next- generation functional devices. All amorphous oxide channels were deposited using RF sputter, employing amorphous Zn–Sn–O (a-ZTO), amorphous Si–Zn–Sn–O (a- SZTO), and amorphous Si–In–Zn–O (a-SIZO). Among these materials, a-SZTO offers high stability without incorporating rare-earth elements such as In or Ga, while a-SIZO exhibits excellent electrical performance even after low-temperature annealing below 150 °C. The thin films and TFT characteristics were measured using semiconductor parameter analyzers (HP4156C and Keithley 4200A-SCS). In the first part of this research, the electrical property and stability of amorphous oxide TFTs were improved through structural modifications and post-annealing processes. In the second part, structural engineering of the TFTs was employed to further enhance electrical performance, and the electron injection mechanism of the metal capping (MC) layer was explored for application in touch sensing. A PDMS- based triboelectric nanogenerator (TENG) was fabricated and integrated with the MC- TFT device to demonstrate human finger-touch detection. Finally, a-ZTO and a-SZTO TFTs were combined with a ferroelectric poly(vinylidene fluoride-co-trifluoroethylene) (P(VDF-TrFE)) layer to fabricate ferroelectric field-effect transistors (FeFETs), enabling the evaluation of neuromorphic synaptic behavior. The devices were stimulated with electrical pulses, allowing the extraction of long-term potentiation (LTP) and long-term depression (LTD) characteristics, from which nonlinearity values were calculated. These experimentally obtained parameters were then used for system-level simulations. A multilayer perceptron (MLP) model was trained using the Modified National Institute of Standards and Technology (MNIST) handwritten dataset via NeuroSim V3.0. The a-ZTO device, which exhibited superior synaptic characteristics, achieved a maximum classification accuracy of 72% in the simulation. Keywords: Amorphous oxide semiconductor (AOS), Thin film transistor (TFTs), Plasma treatment, Metal capping layer, Touch sensor, Neuromorphic device

목차 (Table of Contents)

• CHAPTER 1. Introduction 1
• CHAPTER 2. Background 3
• 2.1 Background of Amorphous oxide semiconductors 3
• 2.2 Thin film transistor 6
• 2.3 Mechanism and effect of plasma treatment 10
• 2.4 Mechanism of MC structure on amorphous oxide semiconductor 13
• 2.5 Amorphous oxide semiconductors based neuromorphic devices 15
• CHAPTER 3. Experiment details 18
• 3.1 Fabrication process of device 18
• 3.1.1 Fabrication process of AOS TFT 18
• 3.1.2 Fabrication of MC TFT 18
• 3.1.3 Condition for plasma treatment 18
• 3.2 Fabrication process of touch sensor 22
• 3.3 Characterization of FeFET based neuromorphic device 24
• 3.4 Fabrication and measurement of touch sensor based on MC TFT 26
• CHAPTER 4. Results and discussion 28
• 4.1 Effect of plasma treatment on AOS 28
• 4.2 MC TFT using touch sensor 36
• 4.3 Neuromorphic device based on AOS 44
• CHAPTER 5. Conclusion 52
• References 53
• 국문초록 57