Ferroionic‐Assisted, Charge‐Driven Polymorphic Phase Engineering for Multilevel Memristive Devices

ABSTRACT Memristive device operation can arise from diverse microscopic mechanisms, including charge polarization and trapping, magnetic ordering, ionic migration, and structural phase transitions. Among these, crystalline–crystalline phase transitions in polymorphic materials offer a unique platform for nonvolatile resistance modulation but are often hindered by large energy barriers and limited endurance. Here, we demonstrate a multimodal, charge‐driven phase modulation in an atomically thin polymorphic material, Mo 0.92 W 0.08 Te 2 , enabled by electrostatic coupling to a ferroionic substrate, CuInP 2 S 6 . The polymorphic channel exhibits reversible switching between semiconducting and metallic phases via electrostatically accumulated charges, reflecting the intrinsically low energy barrier between its competing polymorphs. Integration with the ferroionic substrate enables spatially localized, nonvolatile charge accumulation in the channel, which provides distinct, programmable phase‐control pathways under a two‐terminal bias configuration. The phase‐induced resistance modulation through field‐driven ionic processes is governed by the evolution of a phase‐dependent Schottky junction at the electrode–channel interface, yielding multiple, stable conductance states under pulsed electrical excitation. Leveraging these experimentally derived synaptic characteristics, we conducted an MNIST inference simulation with a convolutional neural network, achieving 98.45% accuracy. This ferroionic‐assisted phase‐engineering strategy offers new opportunities beyond traditional phase‐change devices for adaptive electronic and neuromorphic material platforms.