Azobenzene-introduced organic electrochemical transistors with enhanced synaptic properties for neuromorphic computing

The exponential growth of data-intensive applications has exposed the energy and latency bottlenecks of traditional von Neumann architectures, driving the demand for hardware-based neuromorphic systems. Organic electrochemical transistors (OECTs) have emerged as promising artificial synaptic devices. However, achieving robust long-term plasticity (LTP) remains a critical challenge due to the thermodynamically driven spontaneous de-doping of the active layer. Herein, we present a novel OECT-based synaptic device that overcomes this limitation by incorporating an azobenzene lithiation layer. This redox-active layer effectively compensates for charge imbalance during the electrochemical doping of the poly(3-hexylthiophene-2,5-diyl) (P3HT) channel, thereby maintaining the electrical neutrality of the electrolyte and significantly suppressing spontaneous de-doping. Consequently, the azobenzene-introduced Li-ion gel electrolyte transistor (AB-LGET) exhibits remarkably enhanced synaptic properties, including robust short-term plasticity (STP), prolonged LTP retention, and highly linear weight updates. Grazing-incidence wide-angle X-ray scattering (GIWAXS) analysis confirms that this enhanced plasticity originates from sustained molecular doping within the P3HT crystalline domains. Furthermore, the device achieves a high recognition accuracy of 93.02% in an artificial neural network simulation using the Modified National Institute of Standards and Technology (MNIST) dataset. This study provides a profound understanding of how managing interfacial electrical neutrality via redox reactions can fundamentally advance the synaptic performance of organic neuromorphic devices.