Related Experiment Video
Updated: May 20, 2025

In Situ Transmission Electron Microscopy with Biasing and Fabrication of Asymmetric Crossbars Based on Mixed-Phased a-VOx
Published on: May 13, 2020
Strategic Material Design for Highly Reliable QLC 3D V-NAND Using Bypass Resistive Random Access Memory
Geonhui Han1, Jongseon Seo1, Junghoon Park2
1Center for Single Atom-based Semiconductor Device and the Department of Materials Science and Engineering, Pohang University of Science and Technology, Pohang 37673, Korea.
Abstract:
To overcome the limitation of conventional flash memory, electrochemical random-access memory (ECRAM)-based bypass memory (bypass RRAM) has been proposed as a potential candidate for V-NAND memory application. While bypass RRAM demonstrates excellent memory characteristics through ion hopping conduction, the key parameters governing multilevel cell (MLC) operation remain unexplored. In this study, we propose design guidelines for bypass RRAM, targeting highly uniform quadruple-level cell (QLC) operation by using quantized oxygen vacancy (Vo) injections. To achieve the uniform QLC operation, we precisely controlled ion migration using material engineering in the bypass RRAM. By leveraging the unique electrical characteristic of the WO resistive switching (RS) layer, we minimized Vo migration (from WO2.65 to WO2.73), which enabled low-voltage operation (<5 V) and a significant on/off ratio (>106) with a minimal stoichiometry (Δx < 0.08) change. Additionally, key parameters, such as ionic barrier (Ea,ion) in the electrolyte layer and ion diffusivity (Dion) in the RS layer, were identified to achieve both a high on/off ratio and a uniform sensing margin based on MATLAB simulations and experimental results. As a result, optimized parameters led to superior QLC performance, featuring a highly uniform distribution (σ/μ ∼ 0.01) and a uniform sensing margin (ΔG ∼ 4 μS) between each state without read disturbance issues. Finally, we also confirmed that the substantial reduction of the Vo migration at the nanometer scale suggests the potential for extending beyond QLC levels with quantized Vo injection, ensuring highly uniform switching for V-NAND memory.

