Multibit tribotronic nonvolatile memory based on van der Waals heterostructures

Abstract

Low power and multifunctional nonvolatile memories are promising candidates for processing massive data in the Internet of Things era. However, the storage states in conventional memory devices are under the restricted control by electrical or optical signals. Herein, a new multibit tribotronic nonvolatile memory (T-NVM) based on a graphene/hexagonal boron nitride/molybdenum disulfide van der Waals heterostructure and triboelectric nanogenerator (TENG) is proposed. The programming/erasing states can be modulated by the triboelectric potential, which is determined by changing the distance between the two triboelectrification layers. Under the modulation of external mechanical actions, the device exhibits a high on/off ratio of 105 via manipulating mechanical distance from − 0.2 mm to + 0.2 mm, a long retention time up to 6000 s, a stable switching behavior for over 100 cycles, and a multilevel data storage capability of 14 stages by different external stimuli. Furthermore, a memory inverter circuit employing the triboelectric potential as input signals can serve as the conversion of logical signals. This work proves the great potential of tribotronic devices for direct interaction with external environment in lower power and broadening diverse applications of human-robot interactions, self-powered wearable devices, and intelligent instrumentation.