Temperature-resilient solid-state organic artificial synapses for neuromorphic computing
Temperature-resilient solid-state organic artificial synapses for neuromorphic computing
复制标题
DOI：
10.1126/sciadv.abb2958
复制
发表时间：
2020-07-01
期刊：
影响因子：
13.6
通讯作者：
Salleo, A.
中科院分区：
文献类型：
--
作者：
Melianas, A.;Quill, T. J.;Salleo, A.
具有可调电阻的设备在神经形态计算中受到高度追捧。然而，传统的电阻式存储器遭受非线性和非对称电阻调谐以及过度的写入噪声，从而降低了人工神经网络（ANN）加速器性能。新兴的电化学随机存取存储器（ECRAM）显示写入线性，这使得通过并行阵列编程实现更快的ANN训练。然而，现有技术的ECRAM尚未在封装电子器件所需的温度（类似于90摄氏度）下表现出稳定和有效的操作。在这里，我们表明，（半）导电聚合物与离子凝胶电解质薄膜相结合，使固态ECRAM能够在高达90摄氏度的温度下稳定且几乎不受温度影响。这些ECRAM显示了> 2倍动态范围的线性电阻调谐、20纳秒开关、亚微秒读写循环、低噪声、低电压（+/- 1伏）和低能量（类似于每次写入80毫微微焦耳）操作以及出色的耐久性（在90摄氏度下>10（9）次读写操作）。这些高性能ECRAM的演示是在硬件ANN中实现它们的基本步骤。
Devices with tunable resistance are highly sought after for neuromorphic computing. Conventional resistive memories, however, suffer from nonlinear and asymmetric resistance tuning and excessive write noise, degrading artificial neural network (ANN) accelerator performance. Emerging electrochemical random-access memories (ECRAMs) display write linearity, which enables substantially faster ANN training by array programing in parallel. However, state-of-the-art ECRAMs have not yet demonstrated stable and efficient operation at temperatures required for packaged electronic devices (similar to 90 degrees C). Here, we show that (semi)conducting polymers combined with ion gel electrolyte films enable solid-state ECRAMs with stable and nearly temperature-independent operation up to 90 degrees C. These ECRAMs show linear resistance tuning over a >2x dynamic range, 20-nanosecond switching, submicrosecond write-read cycling, low noise, and low-voltage (+/- 1 volt) and low-energy (similar to 80 femtojoules per write) operation combined with excellent endurance (>10(9) write-read operations at 90 degrees C). Demonstration of these high-performance ECRAMs is a fundamental step toward their implementation in hardware ANNs.