Enhancing Stability and Performance of Conductive Bridge Random Access Memory: Use of a Copper-Doped ZnO...

Computing-in-memory (CIM) technology for edge computing systems demands memory devices that are not only fast but also capable of being easily integrated into stackable manufacturing processes. To address these requirements, conductive bridge random access memory (CBRAM) devices, which are compatible with copper-wiring processes, have emerged as promising candidates. However, the practical implementation of CBRAM is hindered by challenges associated with physical defects in its operating mechanism, leading to issues such as non-uniform switching voltages and high energy consumption in high-density storage applications. This study introduces a layer of copper-doped zinc oxide (ZnO) nanorods as the switching medium in the CBRAM device. The nanostructures play a critical role in regulating copper ion diffusion, thereby facilitating the formation of uniform conductive filaments. The influence of ZnO nanorods on the copper ion diffusion behavior and filament morphology was analyzed through characteristic current fitting of CBRAM devices. Experimental results demonstrate that ZnO nanorod-embedded CBRAM exhibits significantly enhanced memory performance, characterized by superior switching uniformity compared to devices lacking ZnO nanorods. Furthermore, CBRAM devices incorporating copper-doped ZnO nanorods demonstrate even greater memory performance, achieving exceptional uniformity. This approach not only reduces operating voltage and energy consumption but also improves switching uniformity and enhances the overall stability of CBRAM devices, making them more viable for advanced CIM applications.