神经形态工程学
光刻
材料科学
电解质
可扩展性
晶体管
纳米技术
计算机科学
串扰
光子学
制作
光电子学
电子工程
电气工程
电极
人工智能
人工神经网络
电压
工程类
化学
物理化学
医学
替代医学
病理
数据库
作者
Li Yuan,Tingting Zhao,Junshuai Dai,Longwei Xue,Xudong Zhang,Cong Peng,Pan Wen,Hai Liu,Honggang Hu,Longlong Chen,Hanshen Xin,Jun Li,Xifeng Li,Jianhua Zhang
标识
DOI:10.1002/adfm.202418052
摘要
Abstract High‐density bio‐electrolyte‐gated synaptic transistors (BEGTs) array are promising for constructing neuromorphic computing architectures. Due to the bulk ion conductivity and the crack sensitivity of the electrolyte film, patterning the electrolyte is an indispensable route to prevent spatial crosstalk and improve the flexibility of the device array. However, the susceptibility of bio‐electrolyte to organic solvents poses challenges in developing reliable all‐photolithography techniques for fabricating scalable, patterned, and high‐density BEGTs array. This study introduces an all‐photolithography method that adopts a photo‐crosslinker‐enabled electrolyte to create a high‐density (11846 devices per cm 2 ) multimodal BEGTs array. This array demonstrates essential neuromorphic behaviors without inter‐device crosstalk and maintains its flexibility, enduring 200 bending cycles at a 6 mm radius without significant performance degradation. Meanwhile, the BEGTs array exhibits multimodal synaptic behavior, not only successfully mimicking the biological visual memory system for sensing and processing images but also proving highly accurate in classifying handwritten digits, making it suitable for constructing neuromorphic computing systems. This work offers a dependable strategy for the scalable and stable fabrication of BEGTs array, providing valuable insights for advancing artificial neuromorphic systems.
科研通智能强力驱动
Strongly Powered by AbleSci AI