Programmable Electrofluidics for Ionic Liquid Based Neuromorphic Platform
Programmable Electrofluidics for Ionic Liquid Based Neuromorphic Platform
复制标题
DOI：
10.3390/mi10070478
复制
发表时间：
2019-07
期刊：
影响因子：
3.4
通讯作者：
Walker L Boldman;Cheng Zhang;T. Ward;Dayrl P. Briggs;B. Srijanto;P. Brisk;P. Rack
中科院分区：
文献类型：
--
作者：
Walker L Boldman;Cheng Zhang;T. Ward;Dayrl P. Briggs;B. Srijanto;P. Brisk;P. Rack
由于冯·诺伊曼瓶颈导致的计算能力的限制，正在开发通过将设备特征与神经元的突触重量相关联来模拟大脑中的神经生物处理的计算设备。该平台结合了离子液体浇注和电润湿，实现了离子液体的可编程放置/连接。在该平台中，短期增强(STP)和长期增强(LTP)分别通过对非晶态氧化铟镓锌氧化物(AIGZO)的静电和电化学掺杂来实现，并且为了更低的功耗考虑采用了脉冲偏置测量。虽然与电阻元件兼容，但我们展示了一种基于传递性非晶氧化铟镓锌(AIGZO)像素元件的平台。使用锂基离子液体，我们展示了增强(降低器件电阻)和抑制(增加器件电阻)，并提出了一种2D平台阵列，它将通过有源矩阵电润湿实现更高的像素计数。
Due to the limit in computing power arising from the Von Neumann bottleneck, computational devices are being developed that mimic neuro-biological processing in the brain by correlating the device characteristics with the synaptic weight of neurons. This platform combines ionic liquid gating and electrowetting for programmable placement/connectivity of the ionic liquid. In this platform, both short-term potentiation (STP) and long-term potentiation (LTP) are realized via electrostatic and electrochemical doping of the amorphous indium gallium zinc oxide (aIGZO), respectively, and pulsed bias measurements are demonstrated for lower power considerations. While compatible with resistive elements, we demonstrate a platform based on transitive amorphous indium gallium zinc oxide (aIGZO) pixel elements. Using a lithium based ionic liquid, we demonstrate both potentiation (decrease in device resistance) and depression (increase in device resistance), and propose a 2D platform array that would enable a much higher pixel count via Active Matrix electrowetting.