Parylene-MoO_x nanocomposite memristor crossbar for neuromorphic computing applications

Memristors, devices demonstrating the resistive switching (RS) effect, have been the subject of active investigation in recent years. It is currently considered that memristors are promising candidates for creating neuromorphic computing systems based on them. In this study, we examined the key characteristics of crossbar array of parylene (PPX)-MoO memristors with a view to subsequent neuromorphic applications. Memristive devices based on PPX$\mathbf{M o O}_{x}$ nanocomposite were found to exhibit cycle-to-cycle (C2C) uniformity and device-to-device (D2D) reproducibility. The coexistence of volatile and non-volatile modes of RS was demonstrated. The first results of a reservoir computing system simulation with PPX-MoO ${ }_{x}$ memristor-based reservoir for the MNIST dataset classification are presented