Resistive Switching and Synapse Properties of Bilayered CuO|MAPbI 3 Nanometer-Thick Films

Brain-inspired computing can be effectively realized by using oxide-based memristors that replicate biological synaptic functions. CuO nanoparticles synthesized via the sol–gel route were employed for photosynaptic memristive devices aimed at neuromorphic applications. The CuO exhibited uniform grain morphology, confirmed by field emission scanning electron microscopy and showed a direct band gap of 2.18 eV by Tauc’s plot. Electrical characterization revealed stable bipolar resistive switching with current levels ranging from 10–10 to 10–5 A and reliable retentivity up to 350 min. We report the enhanced device performance for a hybrid CuO/MAPbI3 heterostructure by exploiting the complementary roles of oxygen vacancies in CuO and iodide ion migration in MAPbI3. This synergistic mechanism lowered operating voltages, improved cycle stability, and enhanced uniformity compared to single-layer CuO devices. Moreover, the heterostructure device exhibited superior photosynaptic responses under combined optical and electrical stimuli, leading to improved plasticity and energy-efficient neural computation. Neural network simulations of single-layer CuO and CuO/MAPbI3 heterostructure memristors achieve 91 and 97% recognition accuracies with 8 × 8-pixel images, respectively. These findings establish that the heterostructure of the hybrid CuO/MAPbI3 memristors is a promising candidate for advanced neuromorphic hardware and next-generation brain-inspired computing technologies.