Low-Power Artificial Neurons Based on Ag/TiN/HfAlOx/Pt Threshold Switching Memristor for Neuromorphic Computing
Yifan Lu, Yi Li, Haoyang Li, Tianqing Wan, Xiaodi Huang, Yuhui He, Xiangshui Miao
Abstract
Threshold switching (TS) devices are promising candidates to build highly compact and energy efficient artificial neurons. Here, we present a Pt/Ag/TiN/HfAlO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">x</sub> /Pt (PATHP) device with excellent TS characteristics, including a large selectivity(10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">10</sup> ), a wide range of operation current from 10 nA to 1 mA, an extremely steep slope (0.63 mV/dec) and fast turn-on speed (50 ns). The stable TS performance can be ascribed to the introduction of TiN buffer layer and the alternate atomic layer deposited HfAlOx layer. Further, we experimentally demonstrate the functions of leaky-integrate-and-fire neurons with low power feature based on a RC circuit and a single device, respectively, which are essential for constructing spiking neuromorphic systems.