晶体管
材料科学
光电子学
二硫化钼
阈下摆动
兴奋剂
油藏计算
神经形态工程学
场效应晶体管
阈下传导
电导
计算机科学
人工神经网络
电气工程
物理
电压
人工智能
循环神经网络
工程类
凝聚态物理
冶金
作者
Yitong Chen,Rui Wang,Dingwei Li,Qi Huang,Yingjie Tang,Huihui Ren,Yan Wang,Guolei Liu,Fanfan Li,Hong Wang,Bowen Zhu
出处
期刊:Small
[Wiley]
日期:2025-06-25
卷期号:: e2503836-e2503836
被引量:2
标识
DOI:10.1002/smll.202503836
摘要
Molybdenum disulfide (MoS2) has drawn extensive interest due to its excellent performance as a 2D n-type semiconductor. However, achieving high-performance devices based on p-type MoS2, which is crucial for complementary field effect transistors (CFETs) and neuromorphic computing applications, remains limited. Here, an optoelectronic reservoir computing (RC) based on a p-type Nb-doped MoS2 device is reported. The device shows p-type transistor characteristics with a high on/off current ratio exceeding 106, low subthreshold swing (SS) of 234 mV dec-1, notable on-state current of 12 µA µm-1, along with robust cyclic performance and ambient stability. Importantly, the optically tunable facilitated and depressed conductance arises from Nb-vacancy defect states, confirmed by Kelvin probe force microscopy measurement. Furthermore, the device enables 100% motion recognition across eight directions and functions as a reservoir to map feature information from various sequential optical inputs, exhibiting computing capability in pattern recognition with an accuracy of 88%.
科研通智能强力驱动
Strongly Powered by AbleSci AI