Skip to main navigation Skip to search Skip to main content

A Stacked Hybrid Organic/Inorganic Electrochemical Random-Access Memory for Scalable Implementation

Research output: Contribution to journalArticlepeer-review

15 Scopus citations

Abstract

Organic semiconductors have emerged as an attractive class of materials for neuromorphic computing applications, particularly in crossbar arrays for artificial neural network (ANN) accelerators. Here, one of the last persistent challenges facing organic materials for adoption in these applications is addressed by developing a fabrication process capable of lithographically patterning vertical, three-terminal electrochemical random-access memories (ECRAMs). Central to the realization of this device architecture is the development of a hybrid electrolyte system: a porous inorganic matrix permeated with an ionic liquid, which enables vertical stacking of the organic semiconductor channel and gate. The resulting stacked hybrid organic/inorganic ECRAMs (SHOEs) exhibit superior dynamic range to comparable lateral devices (>2×), exceptional cycling endurance (>109 Write–Read Cycles), low energy switching (2.7 pJ), and can be fabricated with dimensions limited by lithographic resolution. The fabrication process developed allows for independent control over device channel, electrolyte, and gate dimensions, and by reducing channel lengths down to a single micron, the fabricated devices can operate (Write+Read) at >MHz speeds. Further, the hybrid electrolyte design provides an effective means to confine an ionic liquid for use in other electrolyte-gated devices.

Original languageEnglish
Article number2100426
JournalAdvanced Electronic Materials
Volume8
Issue number8
Early online dateJul 17 2021
DOIs
StatePublished - Aug 2022
Externally publishedYes

Keywords

  • artificial synapse
  • ionic liquid
  • nanofabrication
  • neuromorphic computing
  • organic semiconductor
  • PEDOT:PSS

Fingerprint

Dive into the research topics of 'A Stacked Hybrid Organic/Inorganic Electrochemical Random-Access Memory for Scalable Implementation'. Together they form a unique fingerprint.

Cite this