Hfo2-Based Rram with In-Situ Conductive Channels Induced by Nanoparticles to Improve Stability

Resistive random access memory (RRAM) has emerged as a novel memory technology due to its simple structure and compatibility with CMOS technology. In recent years, it was reported that HfO2-based RRAMs are promising for commercial applications but suffer from a relatively poor performance in stability and retention due to the random formation and breakage of the conductive channels. In this work, we introduced a layer of Ag nanoparticles embedded in the HfO2 resistive switching layer to address the randomness in HfO2-based RRAMs. It was found that the embedding of Ag nanoparticles enhances the local electric field in HfO2, as a result of which conductive filaments in the RRAM will preferably grow where nanoparticles are located. By controlling the uniformity and distribution of Ag nanoparticles, the resistive switching uniformity is significantly improved. The relative fluctuation of Vset is reduced from 13.85% to 6.96%, and the relative fluctuation of Vreset is reduced from 23.01% to 14.45%. More importantly, we also achieved strong stability and a large resistive window with an on/off ratio of up to 102. In the end, we proposed a resistive switching model to interpret the experimental observations.