Work

Study on Noise in Metal Oxide-Based RRAM and Its Application

Year: 2023
Type: dissertation
Abstract: In the current pyramid-like structures memory hierarchy, it consists of, from top to bottom, a processing core, cache memory by static random access memory (SRAM), main memory by dynamic random access memory (DRAM), and storage memory by solid-state disk (SSD), or hard disk drive (HDD). In general, the closer to the processing core, the more high-speed operation is required, whereas the farther away from the core, the higher storage capacity is demanded. Consequently, the performance gap between DRAM and NAND Flash memory, which are currently major memory technologies, is continuously increasing. However, the need for new memory technology is increasing in order to solve the problem of data processing speed due to the explosive increase in the amount of data and the physical limitation of the existing memory technologies that has been raised for a long time. In addition, research and development on the storage class memory (SCM) technology is in progress as method of implementing In-Memory Process, a concept to solve the problem of Von Neumann architecture in various research groups. \nAmong the candidates on the SCM, which satisfies both the high speed of DRAM and the density of NAND Flash, the resistive switching random access memory (RRAM) has been widely investigated as a leading candidate for next generation nonvolatile memory applications due to RRAMs advantageous features such as simple structure, low cost, high density, fast operation, and CMOS compatibility. However, the reliability issues which PCM suffered from is also being reproduced in RRAM. RRAMs various issues such as endurance, retention, and uniformity stem from intrinsic variability because resistive switching mechanism of RRAM itself is fundamentally stochastic. \nThe main content of this dissertation is to develop a new electrical analysis technique to improve the reliability of RRAM. \nFirst, the elementary low frequency noise (LFN) characteristics of various RRAM devices were analyzed, and the correlation between LFN characteristics and the conduction/resistive switching mechanisms was experimentally verified. \nAlso, it was suggested that the LFN measurement can be an additional analysis technique for devices degradation mechanism and multi-level cell (MLC) operation. \nFinally, from the random telegraph noise (RTN) measurement, we conducted a study to extract the position and energy of traps that can cause cells failure. The experiment on the extraction of traps physical information using the RTN measurement was conducted for the first in this study, and then research findings provided researchers with guidelines for the RTN analysis of RRAM.
Author: 이정규
Language: English
Open Access status: green
OpenAlex ID: W6987375552
Cites: 85
Cited by:
Related to: 10