Halide perovskite based synaptic devices for neuromorphic systems
Halide perovskite based synaptic devices for neuromorphic systems
复制标题
DOI：
10.1016/j.mtphys.2022.100667
复制
发表时间：
2022-03-26
影响因子：
11.5
通讯作者：
Newman, Nathan
中科院分区：
文献类型：
--
作者：
Beom, Keonwon;Fan, Zhaoyang;Newman, Nathan
具有模仿生物神经网络的体系结构的信息系统之所以引起人们的兴趣，是因为它们可以有效地执行自适应学习和记忆功能，并立即处理大量信息。卤化物钙钛矿(HPS)由于具有传统半导体和金属氧化物所没有的独特性质，最近被用来制造记忆电阻、记忆电容器和光电晶体管，作为神经形态器件用于这些系统。在这篇综述中，我们介绍了人工神经网络(ANN)的基本原理，强调了HPS在这种背景下的独特性质，讨论了适用于ANN的不同的基于HP的神经形态装置，重点介绍了它们的初步性能演示的例子，并对它们的问题和未来的前景进行了评论。(C)爱思唯尔有限公司出版的《2022年》。
Information systems with architectures that mimic biological neural networks are of interest because they can efficiently perform adaptive learning and memory functions and process vast amount of in-formation instantly. Halide perovskites (HPs) have been recently explored to fabricate memristors, memcapacitors, and phototransistors as neuromorphic devices used in these systems, thanks to their unique properties, which have not been seen in conventional semiconductors and metal oxides. In this review, we introduce fundamentals of artificial neural networks (ANNs), emphasize unique properties of HPs in such a context, discuss different HP-based neuromorphic devices suitable for ANNs, highlight examples on their preliminary performance demonstration, and comment on their issues and future perspectives. (c) 2022 Published by Elsevier Ltd.