Faisal Farooq, Priya Kaith, Ashok Bera
Designing a stable hybrid perovskite memristor that exhibits synaptic plasticity under both optical and electrical stimuli is a compelling direction for exploiting the unprecedented optoelectronic properties of hybrid perovskites to their full potential towards neuromorphic computing. Here, we report that introducing a thin Polymethyl Methacrylate (PMMA) barrier layer between CH3NH3PbBr3 (MAPbBr3) and the Ag electrode enables synaptic plasticity under optical stimuli in the Ag/PMMA/MAPbBr3/TiO2/FTO memristor. A controlled experimental design with and without PMMA, in combination with electrochemical impedance studies and transient photocurrent measurements, suggests that photo-generated hole trapping at the perovskite/PMMA interface likely leads to such synaptic behavior under pulsed illumination. Moreover, the device exhibits synaptic plasticity under pulsed electrical stimuli, confirming its optoelectrical synaptic properties. Additionally, a simulated artificial neural network (ANN) utilizing experimentally measured synaptic conductance states achieves 96.14% recognition accuracy on the MNIST dataset, highlighting its potential for neuromorphic image recognition applications. The device shows no significant degradation under prolonged illumination, vacuum, elevated temperatures up to 80 °C, or long-term storage in ambient conditions, demonstrating the stability of the device with a PMMA coating and the effectiveness of the hybrid perovskite for optoelectronic neuromorphic computing.