Vertical organic synapse expandable to 3D crossbar array

Abstract

Recently, three-terminal synaptic devices have attracted considerable attention owing to their nondestructive weight-update behavior, which is attributed to the completely separated terminals for reading and writing. However, the structural limitations of these devices, such as a low array density and complex line design, are predicted to result in low processing speeds and high energy consumption of the entire system. Here, we propose a vertical three-terminal synapse featuring a remote weight update via ion gel, which is also extendable to a crossbar array structure. This synaptic device exhibits excellent synaptic characteristics, which are achieved via precise control of ion penetration onto the vertical channel through the weight-control terminal. Especially, the applicability of the developed vertical organic synapse array to neuromorphic computing is demonstrated using a simple crossbar synapse array. The proposed synaptic device technology is expected to be an important steppingstone to the development of high-performance and high-density neural networks.

Dendritic Network Implementable Organic Neurofiber Transistors with Enhanced Memory Cyclic Endurance for Spatiotemporal Iterative Learning

“…This result is comparable to that for the previously reported planar P3HT‐based OECT neuromorphic devices. [ 16 ] Notably, an increase of the OFF‐current ( I off ) state with repeated operation cycles was unexpectedly observed. This increase of I off during cyclic operation was also observed in the typical planar P3HT‐OECT devices, as shown in Figure S3 (Supporting Information), which indicates that this phenomenon is not attributable to the neurofiber‐OECT device architecture.…”

“…[ 8–15 ] Basic circuit simulations for learning, such as handwritten(MNIST) classification and Pavlov associative based on the synaptic functions of OECTs, have demonstrated preliminary feasibility of using the OECT neuromorphic devices in applications. [ 12,16 ] Even though such significant advances in OECT‐based neuromorphic devices have been reported, the electrochemical doping mechanism of conjugated polymers in an OECT synaptic device—including how the ionic species at the conjugated polymer/electrolyte interfaces interact with the conjugated polymer under synaptic functional conditions—has not been fully elucidated. Understanding this mechanism is essential for further improving the performance of OECT neuromorphic devices.…”

Organic Synaptic Transistors: The Evolutionary Path from Memory Cells to the Application of Artificial Neural Networks

“…Given its working mode, EDLTs can be made into multi‐terminals synaptic transistors with several side gates, as well as vertical structure organic synaptic transistors. [ 40,48,49,95,96 ] In particular, the generation of side gate devices has been reported in recent years. For instance, several solid ion electrolyte materials, such as chicken albumen, chitosan, ion‐gel, and wood‐derived cellulose nano‐papers (WCNs), have been used in multi terminal synaptic or vertical structure devices to form EDL structures.…”

“…In addition, the three‐terminal devices can be designed as a vertical structure to form a highly interconnected neural network system. [ 48 ] At the same time, the postsynaptic received signal is not easily disturbed by the presynaptic input signal in three terminal devices, which is conducive to the design of multi‐input devices. [ 40,49 ] Another evolution of organic synaptic transistors is to make the function of simulating synaptic behavior more comprehensive and to increase diversity of input signals, from the previous simulation of an information processing system of human brains, to the current simulation of the visual nervous system of human eyes, tactile nervous system of human skin, auditory nervous system of human ears, and so on.…”

Non‐Volatile Electrolyte‐Gated Transistors Based on Graphdiyne/MoS2 with Robust Stability for Low‐Power Neuromorphic Computing and Logic‐In‐Memory

“…In the era of the "big data," massive unstructured data including images, sound, video, and text need to be handled efficiently, which is becoming a severe challenge for conventional von Neumann computers with physically separated processing and memory units. [1][2][3] Human brain-inspired neuromorphic computing, which can parallelly process unstructured information in an energyefficient manner, has attracted great attention in recent years. [3][4][5][6][7][8][9][10][11][12] Various artificial synapses, such as two-terminal memristors [13][14][15][16][17] and multiterminal neuromorphic transistors, [18][19][20][21][22][23][24] have been proposed to construct hardware artificial neural networks (ANNs) for neuromorphic computing.…”