Energy efficient multi-level memory using paper based second order MLGraphene/MoO3 - Aloe vera/MLGraphene memristor device for emulating synaptic functionalities

Abstract

Neuromorphic computing is an emerging architype representing a cutting-edge approach to computing that emulates the structure and function of human brain, leveraging neuroscience concepts to develop efficient, adaptive, and power conscious computing system surpassing the von Neumann architecture. Herein, we report artificial synaptic device defined on a paper using MoO3 embedded Aloe vera matrix as an active material. The multilayer graphene electrode (MLG) is drawn using pencil-on-paper (PoP) method. Devices could be programmed for multi bit-states to avail several conducting states (2n with n = 1,2,3,4). Further, the devices can be operated at low energy consumption (∼ pJ) stable at ambient conditions. Activity dependent measurements show that the synaptic weight update depends on the history of activity. The potentiation and depression can be tuned by properly choosing the prior activity. The threshold frequency at which transition into potentiation occurs is shifted towards lower frequency and depends on the number of prior activities. The potentiation and depression curves indicate that the nonlinearity can be controlled by utilizing non-identical pulse sequences. The pencil-on-paper (PoP) method could represent a new frontier in electronic devices leading to the development of portable, environment friendly, and flexible synaptic devices for versatile synaptic and memory applications.