Parallel programming of an ionic floating-gate memory array for scalable neuromorphic computing

Abstract

Neuromorphic computers could overcome efficiency bottlenecks inherent to conventional computing through parallel programming and readout of artificial neural network weights in a crossbar memory array. However, selective and linear weight updates and <10-nanoampere read currents are required for learning that surpasses conventional computing efficiency. We introduce an ionic floating-gate memory array based on a polymer redox transistor connected to a conductive-bridge memory (CBM). Selective and linear programming of a redox transistor array is executed in parallel by overcoming the bridging threshold voltage of the CBMs. Synaptic weight readout with currents <10 nanoamperes is achieved by diluting the conductive polymer with an insulator to decrease the conductance. The redox transistors endure >1 billion write-read operations and support >1-megahertz write-read frequencies.

Bridging Biological and Artificial Neural Networks with Emerging Neuromorphic Devices: Fundamentals, Progress, and Challenges

“…[108][109][110][111] In addition, electrochemical random-access memory (ECRAM) based on ion intercalation has recently been reported as a promising synaptic cell, showing multi-states and incremental switching with near-ideal switching symmetry and linearity. [29,[112][113][114][115][116][117][118][119] The electrochemically driven ion intercalation process is more controllable than filament-related ion movements in RRAM; therefore, ECRAM also exhibits a much smaller stochasticity. In addition, by borrowing the battery concept, those devices successfully decouple the read and write operations and thus realize low programming energy and long retention time simultaneously.…”

Modeling the Metal–Insulator Phase Transition in LixCoO2 for Energy and Information Storage

“…The current state‐of‐the‐art memristors, e.g., phase‐change memories (PCM) or filamentary‐type metal oxide memories, face drawbacks such as stochasticity in state switching (high “write” noise) and “write” nonlinearities as well as large switching voltages and currents . To overcome some of these challenges, all solid‐state intercalation devices that use Li‐ion/proton intercalation have been proposed as alternative memristor candidates. These devices require low power for switching and maintain their resistive state over large periods of time, i.e., are nonvolatile, as the dopants (intercalated ions) cannot diffuse out of the system without applying an external bias voltage …”

2D MXene–TiO2 Core–Shell Nanosheets as a Data‐Storage Medium in Memory Devices

“…Nano‐floating‐gate transistor memories (NFGTMs) have shown significant promise for application as nonvolatile memories and artificial synaptic devices because of their advantages of nondestructive readout, high stability, and low operation current. [ 28,29 ] However, during miniaturization of NFGTMs aimed at increasing their storage capacity, they suffer from the drawbacks of capacitive coupling caused by the parasitic capacitance after charge trapping and a low gate coupling ratio. [ 30 ] Ti 3 C 2 T x MXene is a potential candidate material for overcoming these drawbacks on account of its intrinsic 2D structure, because the crosstalk between neighboring floating gates can be decreased significantly by reducing the floating‐gate thickness.…”