Tailoring Dynamic Synaptic Plasticity in FeTFT Optoelectronic Synapse for Associative Learning
Peng Yang, Hui Xu, Xiaopeng Luo, Shihao Yu, Yang Liu, Yefan Zhang, Xu Guo, Bing Song, Zhiwei Li, Sen Liu, Qingjiang Li
Abstract
Abstract Neuromorphic hardware with dynamic synaptic plasticity presents fascinating applications in advanced artificial intelligence. However, the development of low‐cost, CMOS (Complementary Metal‐Oxide‐Semiconductor)‐compatible, and dynamically tunable synaptic devices is still nascent. Notably, the spontaneous polarization of hafnium oxide‐based ferroelectric materials, combined with the persistent photoconductivity effect of indium‐gallium‐zinc‐oxide (IGZO) semiconductors, provide a potential solution. In this paper, a novel optoelectronic synaptic device based on ferroelectric thin‐film transistors (FeTFTs) is proposed to achieve dynamic synaptic plasticity through the co‐modulation of light and electrical signals, which can effectively adjust the dynamic range of synaptic weights and emulate complex biological behaviors. The effective dynamic synaptic plasticity of FeTFTs is quantified under different light power intensities and verified through the emulation of complex biological behavior, such as classical conditioning experiments and environmental adaptive behavior. Furthermore, a 3 × 3 FeTFT array is constructed to demonstrate its potential applications in memory functions. This CMOS‐compatible optoelectronic synaptic device with dynamic synaptic plasticity provides a robust hardware foundation for the future development of artificial intelligence, enabling it to adapt to more complex environments and perform tasks efficiently.