Temperature-resilient solid-state organic artificial synapses for neuromorphic computing

• DOI: 10.1126/sciadv.abb2958
• Publication Date: 2020-07-03T23:15:35Z
• Authors: A. Melianas; T. J. Quill; G. LeCroy; Y. Tuchman; H. v. Loo; S. T. Keene; A. Giovannitti; H. R. Lee; I. P. Maria; I. McCulloch; A. Salleo
• Keywords: Neuromorphic engineering; Nanosecond; Resistive touchscreen; Linearity

Devices with tunable resistance are highly sought after for neuromorphic computing. Conventional resistive memories, however, suffer from nonlinear and asymmetric tuning excessive write noise, degrading artificial neural network (ANN) accelerator performance. Emerging electrochemical random-access memories (ECRAMs) display linearity, which enables substantially faster ANN training by array programing in parallel. However, state-of-the-art ECRAMs have not yet demonstrated stable efficient operation at temperatures required packaged electronic devices (~90°C). Here, we show that (semi)conducting polymers combined ion gel electrolyte films enable solid-state nearly temperature-independent up to 90°C. These linear over a >2× dynamic range, 20-nanosecond switching, submicrosecond write-read cycling, low low-voltage (±1 volt) low-energy (~80 femtojoules per write) excellent endurance (>109 operations 90°C). Demonstration of these high-performance is fundamental step toward their implementation hardware ANNs.