Low Power FA<sub>2</sub>PbI<sub>4</sub>/SiO<sub>2</sub> Bilayer Memristors with Pt Nanoparticles Exhibiting Reconfigurable Synaptic and Neuron Properties for Compact Optoelectronic Neuromorphic Systems.

Bousoulas, Panagiotis; Orfanoudakis, Spyros; Spathi, Danai; Pagonis, Victoras; Tsetseris, Leonidas; Tsioustas, Charalampos; Tsipas, Polychronis; Kontos, Athanassios G et al. · Nano Lett · 2025

basic_science · Level V

Where this comes from

Abstract

The development of artificial neural networks with biorealistic computing properties represents a frontier in the neuromorphic computing era. However, achieving compact and energy-efficient integration of silicon-based synapses and neurons remains challenging due to complexities in their electrical circuits. Herein, we fabricated a low power Ag/SiO<sub>2</sub>/FA<sub>2</sub>PbI<sub>4</sub>/Pt nanoparticles/ITO bilayer memristor with reconfigurable properties, exhibiting dual switching modes and neuromorphic functionalities. These effects were experimentally investigated through transient response and endurance measurements, while valuable insights were provided using a comprehensive numerical model. The SiO<sub>2</sub>/FA<sub>2</sub>PbI<sub>4</sub> and FA<sub>2</sub>PbI<sub>4</sub>/Pt nanoparticle interfaces played a critical role in regulating ion migration, stabilizing filament dynamics and enhancing device reliability. A compact optoelectronic neuromorphic system was demonstrated by integrating synaptic and neuronal elements, enabling precise control of the firing activity. An ultralow power consumption (∼10 fJ/spike) was achieved, comparable to that of the human brain and state-of-the-art memristive technologies, thereby paving the way for energy-efficient optoelectronic computing platforms.