Abstract
Controllable multilevel resistive switching (RS) and neuromorphic characteristics emerges as a promising paradigm to build power-efficient computing hardware for high density data storage memory and artificial intelligence. Nevertheless, the current nonvolatile memory still endures from reliability and variability of the memristors. In this work, Pt/Al2O3/HfO2/HfAlOx/TiN multilayer memristor was prepared by using atomic layer deposition (ALD) to examine the well-regulated multilevel RS and neuromorphic properties. The memristor was found to demonstrate admirable RS properties, including forming-free, low operating voltage (Set/Reset), high switching ratio (>100), multi-level retention time (104 s), and good durability (1000 switching cycles). Furthermore, seven and four resistance states can be accomplished by modulating CC through set-operation and stop-voltage during the reset-operation. By modulating the multi-level resistance state, the electronic synapse can simulate synaptic plasticity, such as potentiation/depression, paired pulse facilitation (PPF) and spike-timing-dependent plasticity (STDP). Results show that a multilayer memristor has potential in the application of multilevel data storage memory and bionic portable electronic devices.
NSTL
Elsevier