Enhancement of resistive switching reliability and thermal stability in resistive random access memory and reliability effect on neuromorphic system
저항성 랜덤 액세스 메모리의 저항 스위칭 신뢰성 및 열적 안정성 향상과 소자 신뢰성이 뉴로모픽 시스템에 미치는 영향 연구
- 주제(키워드) Resistive random access memory (RRAM) , Neuromorphic system , Thermal stability , Reliability
- 발행기관 고려대학교 대학원
- 지도교수 유현용
- 발행년도 2022
- 학위수여년월 2022. 2
- 학위구분 박사
- 학과 대학원 나노반도체공학과
- 원문페이지 86 p
- UCI I804:11009-000000257279
- DOI 10.23186/korea.000000257279.11009.0001248
- 본문언어 영어
초록/요약
With the significant technological developments in recent times, memory technology has been rapidly developed to fulfill the demand for memory devices. However, as the silicon (Si) based memory technology has reached scaling limits of the nanoelectronic devices, it is difficult to expect performance improvement of memory devices only by scaling down. Along with the scaling limits, the decreasing system operation efficiency and increasing power consumption were caused by the structural limitations of von-Neumann-architecture-based digital computing. To overcome the scaling and strucuture limitation of current complementary metal-oxide-semiconductor (CMOS) technology, various new memory devices have been developed with their own advantages. As one of the new memory devices, a lot of research has been progressed on ‘memristor’ devices. Due to the structure characteristics like resistor, memristor is a promising candidate for the next generation device that can solve the scaling down problem with the potential of high-density integration and low fabrication cost. Additionally, memristor can perform data storage and operation together. Therefore, memristor can emulate the functions of synapses and neurons required to implement the neuromorphic system. However, most researches on memristors have focused on improving the electrical characteristics of the device and neglect the reliability of memristor operation. The reliability of the memristor is a very important factor because it is closely related to the stability of the neuromorphic system. In this thesis, reliability in RRAM-based memristor was analyzed from the viewpoint of resistance variability and thermal stability, and methods to improve it were suggested. RRAM-based memristor has many advantages such as high integration density, low power consumption, fast switching speed and simple structure. By applying nitrogen doping technique on RRAM-based memristor, low resistance variability and high thermal stability were achieved and reliability of RRAM-based memristor was improved. Additionally, the effect of reliability of RRAM-based memristors on the neuromorphic system was investigated. The reliability problem is more sensitive at the neuromorphic system level because the system is composed of a massive number of RRAM-based memristors.
more목차
Abstract i
Contents ii
List of Figures iii
List of Tables v
Chapter 1. Introduction 1
1.1 Background and Motivation 1
1.2 Thesis Organization 3
1.3 References 4
Chapter 2. Effect of Resistive Switching Stability 5
2.1 Introduction of Resistive Switching Stability 5
2.2 Enhancement of Resistive Switching Stability through Nitrogen Doping Technique 8
2.3 Effect of Resistive Switching Stability in Neuromorphic system 18
2.4 Summary 26
2.5 References 27
Chapter 3. Improvement of Thermal Stability with Active Layer Engineering 32
3.1 Introduction of thermal stability in RRAM 32
3.2 Thermal Stability Improvement with Nitrogen Doping Technique in Active Layer 35
3.3 Summary 47
3.4 References 48
Chapter 4. Effect of Thermal Stability in Neuromorphic system 54
4.1 Introduction of Thermal stability in Memristor and Neuromorphic system 54
4.2 Effect of Top Electrode Metal in Thermal stability of memristor 56
4.3 Effect of Thermal stability in Neuromorphic system 64
4.4 Summary 68
4.5 References 69
Chapter 5. Conclusions 75
Abstract in Korean 76
Curriculum Vitae 77
