Ionic‐Migration‐Gated 2D Perovskite/MoS 2 Junction Field‐Effect Transistor for Ultralow‐Power Neuromorphic Vision

作者：Rui He, Xitong Hong, S. S. Chen, Shufen Zhang, Pingdan Xiao, Qianlei Tian, Xinpei Duan, Zhengdao Xie, Xia Liu, Yawei Lv, Chang Liu, Qinghui Hong, Woo Jong Yu, Lei Liao, Xuming Zou · 发表于：Laser & Photonics Review · 年份：2026 · DOI：10.1002/lpor.71149 · 研究领域：Advanced Memory and Neural Computing、2D Materials and Applications、Ferroelectric and Negative Capacitance Devices

ABSTRACT The continuous aging of global infrastructure urgently calls for autonomous monitoring systems to address the drawbacks of conventional methods, including manual operation, power‐intensive operation, and a lack of real‐time capability. Addressing this, we report a novel 2D perovskite/MoS 2 heterostructure‐based junction field‐effect transistor that harnesses neuromorphic computing for intelligent crack detection. The device operates via a dynamic ionic gating mechanism, where mobile ions within the 2D perovskite gate modulate the MoS 2 channel depletion region, providing a native analogy to neurobiological signal processing while eliminating intricate interface engineering. This yields exceptional electrical characteristics, including a high on/off ratio (>10 7 ), a near‐ideal subthreshold swing (63 mV/dec), and an ultrahigh carrier mobility (215 cm 2 V −1 s −1 ). The dual‐gate architecture enables reconfigurable logic circuits such as inverters with a voltage gain of 50, while light‐induced ion migration facilitates broadband photodetection (375–914 nm) with high detectivity (3.4 × 10 11 Jones). Furthermore, the ion‐migration mechanism underpins biomimetic synaptic plasticity at an ultralow energy consumption of 3.25 fJ per spike. Deployed in an in‐memory neuromorphic network, the system demonstrates 91.6% accuracy in concrete crack recognition through circuit‐level simulation, establishing a foundational building block for energy‐efficient, intelligent infrastru...