Blending With Insulating Polymers Enables Stable Anti‐Ambipolar Characteristics in n‐Type Organic Mixed‐Conductors for Heterogeneous Spiking Neurons

ABSTRACT Mammalian neurons exhibit diverse structural and physiological properties: a heterogeneity not present in current neuromorphic systems. Recent advancements have led to bio‐realistic spiking systems which are based on the integration and long‐term operation of organic electrochemical transistors (OECTs) with anti‐ambipolar characteristics. However, the library of materials enabling anti‐ambipolarity (AA) is restricted by the lack of stable, fast, and scalable organic mixed ionic and electronic conductors (OMIECs) exhibiting this specific OFF‐ON‐OFF behavior. This complicates mimicking specialized neuron types. Blending semiconducting with insulating polymers is a classic organic and OMIEC electronics strategy to improve device performance and enable tunability, yet unexplored for improving anti‐ambipolar stability in non‐ladder‐type anti‐ambipolar materials. Here, the n‐type OMIEC, P(gNDI‐gT2), is blended with the insulating polymer, PMMA, yielding stable anti‐ambipolar characteristics. The miscibility in solution forms a mechanically stabilizing and ion‐permeable framework in the solid state that considerably reduces swelling‐induced structural damage, thus preventing dramatic failure after the first operation cycles without compromising the conductive polymer's electronic pathways. Modulating the polymer blend ratio enables low‐power, fast‐spiking Hodgkin–Huxley neurons with tunable and programmable dynamics, without optimized device geometries, demonstrating frequency deconvolution and reservoir computing. These two computing primitive building blocks are essential for designing efficient spiking neural networks.