Logo image
A Vacuum Gap Selector With Ultra-Low Leakage for Large-Scale Neuromorphic Network
Journal article   Peer reviewed

A Vacuum Gap Selector With Ultra-Low Leakage for Large-Scale Neuromorphic Network

Xinglong Ji, Song Hao, Khin Yin Pang, Kian Guan Lim and Rong Zhao
IEEE electron device letters, Vol.41(3), pp.505-508
01/03/2020

Abstract

Computer architecture Leakage currents low leakage current Memristors Microprocessors neuromorphic computing Neuromorphics Switches Thin-film selector Threshold voltage vacuum gap structure
In this work, we proposed a thin-film selector with a vacuum gap structure for neuromorphic application, which demonstrates outstanding performance including ultra-low leakage current (~0.20 pA), high ON/OFF ratio (>10 9 ), record-high turn-on slope (<; 0.31 mV/dec.), excellent endurance (> 10 8 ) and low turn-on energy (113.37 pJ). Our selector could minimize sneak currents in crossbar array, enabling terabits scale up capability. Moreover, we have integrated the selector with memristors to form a 2-selector-1-memristor structure and demonstrated several learning rules. These outstanding characteristics indicate that our selector has the potential to enable large scale neuromorphic networks.

Metrics

1 Record Views

Details

Logo image