Synaptic learning behavior and neuromorphic computing of Au/MXene/NiO/FTO artificial synapse

• DOI: 10.1063/5.0167497
• Publication Date: 2023-09-25T09:42:38Z
• Authors: Junlin Fang; Zhenhua Tang; Xi-Qi Li; Zhao-Yuan Fan; Yan-Ping Jiang; Qiu-Xiang Liu; Xin-Gui Tang; Jing-Min Fan; Ju Gao; Jie Shang
• Keywords: Neuromorphic engineering; Memristor; MNIST database; Non-blocking I/O

A traditional von Neumann structure cannot adapt to the rapid development of artificial intelligence. To solve this issue, memristors have emerged as preferred devices for simulating synaptic behavior and enabling neural morphological computations. In work, Au/NiO/FTO Au/MXene/NiO/FTO heterojunction were prepared on FTO/glass by a sol-gel method. comparative analysis was carried out investigate changes in electrical properties upon addition MXene films. synapses not only smaller threshold voltage, larger switching ratio, more intermediate conductivity states but also can simulate important behavior. The results show that memristor has better weight update linearity excellent modulation long data retention time characteristics. Utilizing convolutional network architecture, recognition accuracy MNIST Fashion-MNIST datasets improved 96.8% 81.7%, respectively, through implementation random adaptive algorithms. These provide feasible approach combining materials with metal oxides prepare neuromorphic computing.