Ultralow switching voltage slope based on two-dimensional materials for integrated memory and neuromorphic applications

Linfeng Sun, Genuwoo Hwang, Wooseon Choi, Gyeongtak Han, Yishu Zhang, Jinbao Jiang, Shoujun Zheng, Kenji Watanabe, Takashi Taniguchi, Mali Zhao, Rong Zhao, Young Min Kim, Heejun Yang*

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

56 Citations (Scopus)

Abstract

To realize ultrafast and energy-efficient electronic devices, reducing the switching voltage slope for ON and OFF states that scales the supply voltage and device dimensions is critical. Novel device architectures based on two-dimensional (2D) materials have overcome the fundamental thermionic limit of the switching slope (60 mV/dec); however, a versatile switching device required for highly integrated memory and neuromorphic applications has not been achieved with such exceptional switching slope characteristics. Here, we demonstrate a switching voltage slope down to 0.62 mV/dec in a threshold switching device based on a vertical heterojunction of silver/hexagonal boron nitride (h-BN)/graphene. The sub-1 mV/dec switching slope for the first time, maintaining a high ON/OFF ratio (up to 1010), originates from the unique coupling between the migrated silver atoms and the chemically-inert graphene electrode through the 2D insulating h-BN. Moreover, our original switching device enables the evolution from a conventional volatile (threshold switching) to non-volatile memristive state by adequate voltage spikes, which is ideal for selector applications in highly integrated crossbar array architecture and in a novel synaptic device for neuromorphic computing.

Original languageEnglish
Article number104472
JournalNano Energy
Volume69
DOIs
Publication statusPublished - Mar 2020
Externally publishedYes

Keywords

  • Graphene transistor
  • Memory device
  • Neuromorphic devices
  • Selector
  • Two-dimensional materials

Fingerprint

Dive into the research topics of 'Ultralow switching voltage slope based on two-dimensional materials for integrated memory and neuromorphic applications'. Together they form a unique fingerprint.

Cite this