Optoelectronic Synapses of Ni0.5Zn0.5Fe2O4 With Photoelectric Dual-Mode Response for Traffic Scenario Intelligent Detection.

Optoelectronic synaptic devices serve as promising hardware platforms for neuromorphic computing; however, achieving stable photoelectric dual-modal synaptic modulation and practical intelligent perception remains challenging. Herein, an Au/Au/Ni0.5Zn0.5Fe2O4 (NZFO)/FTO optoelectronic synaptic memristor was fabricated via a facile sol-gel method. The NZFO film exhibits a single-phase spinel structure with abundant intrinsic defect states, which contribute to defect-assisted carrier transport and conductance modulation under electrical and optical stimuli. The device demonstrates photoelectric dual-modal synaptic behavior, enabling short-/long-term plasticity, paired-pulse facilitation, and STM-to-LTM transition with wavelength-dependent optical modulation over 365-650 nm. The device achieves stable learning and relearning behaviors within 16 and 9.3 s, respectively, and demonstrates repeatable optical pulse modulation over multiple consecutive cycles. Integrated with a convolutional neural network, the NZFO synaptic device achieves recognition accuracies of 96.20% and 84.45% on the MNIST and Fashion-MNIST datasets, respectively. Furthermore, the device-assisted neuromorphic perception framework enables dynamic target perception and traffic signal recognition in an intelligent driving demonstration, highlighting its capability for integrated sensing, memory, and processing. This work demonstrates NZFO as a promising material platform for energy-efficient neuromorphic perception and intelligent computing applications.