神经形态工程学
纳米线
材料科学
钙钛矿(结构)
纳米结构
纳米技术
同种类的
瓶颈
拓扑(电路)
计算机科学
人工神经网络
电气工程
物理
工程类
人工智能
嵌入式系统
化学工程
热力学
作者
Mei Huang,Shizhen Zhang,Peizhi Zhou,Zhishan Chen,Huiwang Lian,Bo Wang,Qiguang Li,Sareh Sadat Moshirian Farahi,Saravanakumar Subramanian,Qingguang Zeng,Yang Li
标识
DOI:10.1016/j.mtnano.2023.100449
摘要
Metal halide perovskites (MHPs)-constructed devices for artificial synapses have attractive and robust prospects in neuromorphic computing. However, the challenge is that their power-consumption does not compete with the biological synapses with a magnitude of fJ per spike. This is since hierarchical 0D/1D nanostructures generated from the phase separation in the liquid-phase reaction of MHPs intrinsically increase the power consumption, while homogeneous 1D MHPs may effectively break such bottleneck. Here, we propose the heterogeneous interface engineering to prepare homogeneous 1D MHPs CsPbI3 nanowires (NWs) with a very high morphological yield (almost 100 %) and diameter of 17.2 nm. We demonstrate the homogeneous NWs-mediated devices exhibit a low operating voltage (−0.7 V), over one order of magnitude lower than conventional flash memories. Our device also triggers the classic potentiation and depression processes exhibiting excellent electronic synaptic plasticity. This work could be conducive to synthesizing homogeneous 1D MHPs nanostructure in artificial synapses and inspire more research on various nanostructures for neuromorphic system application.
科研通智能强力驱动
Strongly Powered by AbleSci AI