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 infrastructure monitoring.
He et al. (Sat,) studied this question.