Nanowire Nanocomposite Networks with Controlled Resistive Switching and Learning Abilities

ABSTRACT

In‐materia processing offers a new paradigm for computing by enabling learning capacity and pattern recognition directly within a physical substrate. We report the development of a 3D nanowire network formed by a polyvinylpyrrolidone matrix embedded with silver nanowires, tailored for neuromorphic applications. Compared to conventional 2D networks, these nanocomposite films lower the switching voltage by effectively increasing the population of filamentary junctions. The polymer's intrinsic hygroscopic properties enhance hydration, acting as a dynamic humidity‐responsive reservoir that facilitates silver ion mobility and modulates resistive switching. Under elevated relative humidity (RH), the film exhibits reversible swelling and enhanced ion mobility, reducing the threshold voltage required to activate conductance. Once activated at high RH, the conductive state remains highly sensitive when returning to lower RH, providing an encoding strategy for retaining information. Multi‐electrode devices were fabricated to demonstrate spatial switching dynamics, short‐term memory, and associative learning tasks. Finally, an epoxy‐resin encapsulation strategy is proposed to preserve hydration and stabilize the neuromorphic performance under diverse environmental conditions. The synergistic combination of a 3D nanocomposite network with hydration‐assisted ion transport offers a promising route toward adaptive, energy‐efficient neuromorphic devices.