Insights into synaptic functionality and resistive switching in lead iodide flexible memristor devices

Abstract

Neuromorphic platforms are gaining popularity due to their superior efficiency, low power consumption, and adaptable parallel signal processing capabilities, overcoming the limitations of traditional von Neumann architecture. We conduct an in-depth investigation into the factors influencing the resistive switching mechanism in memristor devices utilizing lead iodide (PbI₂). We establish correlations between device performance and morphological features, unveiling synaptic like behaviour of device making it suitable for range of flexible neuromorphic applications. Notably, a highly reliable unipolar switching mechanism is identified, exhibiting stability even under mechanical strain (with a bending radius of approximately 4 mm) and in high humidity environment (at 75% relative humidity) without the need for encapsulation. The investigation delves into the complex interplay of charge transport, ion migration and the active interface, elucidating the factors contributing to the remarkable resistive switching observed in PbI₂-based memristors. The detailed findings highlight synaptic behaviors akin to the modulation of synaptic strengths, with an impressive potentiation and depression of 2 × 10⁴ cycles, emphasizing the role of spike time-dependent plasticity (STDP). The flexible platform demonstrates exceptional performance, achieving a simulated accuracy rate of 95.06% in recognizing modified patterns from the National Institute of Standards and Technology (MNIST) dataset with just 30 training epochs. Ultimately, this research underscores the potential of PbI₂-based flexible memristor devices as versatile component for neuromorphic computing. Moreover, it demonstrate the robustness of PbI₂ memristors in terms of their resistive switching capabilities, showcasing resilience both mechanically and electrically. This underscores their potential in replicating synaptic functions for advanced information processing systems.