Improving Neuromorphic Memory in Organic Electrochemical Transistors through Electrostatic Ion Trapping

Organic electrochemical transistors (OECTs) offer unique opportunities for neuromorphic electronics owing to the versatile molecular design of polymer semiconductors. Nevertheless, their application as memory elements has been hindered by rapid ion relaxation and limited charge retention. Although previous approaches have utilized physical confinement of ions within crystalline polymer domains, the achievable memory window has typically remained below 5.3 V. In this work, we introduce a molecular strategy for regulating hysteresis through a zwitterionic crosslinker that simultaneously controls ion transport and retention via electrostatic interactions. The immobilized sulfonate moiety acts as an electrostatic barrier that suppresses ion injection, whereas the quaternary ammonium center serves as a strong ion-capturing site, generating an energy barrier of 2.03 eV that stabilizes the doped state. The synergistic action of these functionalities results in substantially enhanced hysteresis behavior. Operando GIWAXS measurements under sequential gate biasing reveal reversible structural evolution within the lamellar domains, arising from ionic interactions between side-chain-anchored zwitterionic units. As a consequence, the resulting OECTs exhibit a hysteresis strength of 96.4 V and a memory window of 8.65 V, surpassing previously reported values while preserving a high on/off ratio and retaining 86.4% of the conductance after 200,000 stimulation cycles. Importantly, the zwitterionic functional polymer additive (Z-FPA) strategy is broadly applicable across different polymer semiconductor systems, demonstrating its generality beyond a specific host material. Devices incorporating Z-FPA display improved synaptic symmetry, enabling a recognition accuracy of 92.87% on the MNIST dataset in device-aware neural network simulations. Furthermore, reservoir computing based on electrocardiogram signals verifies the suitability of these devices for real-world biosignal processing. These results establish electrostatic ion trapping as an effective molecular design principle for achieving durable memory and high-performance neuromorphic functionality in next-generation OECT platforms.