Highly Reliable and Uniform Synaptic Transistors Enabled by a Polymer-Engineered Semiconducting Nanotube Network

In this study, we demonstrate an 8 × 8 synaptic transistor array based on a poly(9,9-di-<i>n</i>-dodecylfluorenyl-2,7-diyl) (PFDD)-wrapped semiconducting single-walled carbon nanotube (s-SWCNT) network. The PFDD polymer wrapping of the SWCNTs offers a synergistic combination for synaptic devices: high selectivity of sorting s-SWCNTs and energetically stable charge-trapping sites along the nanotube surface, enabling controlled conductance modulation. In addition, the simple polymer-SWCNT hybrid structure contributes to highly uniform electrical properties in the array. A reproducible conductance-tuning capability is demonstrated in a PFDD-SWCNT synaptic transistor, resulting in robust synaptic plasticity. Diverse synaptic functions, such as excitatory post-synaptic currents, short- and long-term memory, long-term potentiation (LTP) and depression (LTD), and paired-pulse facilitation, are emulated. The stable dynamic modulation of LTP/LTD in the PFDD-SWCNT synaptic device is validated over 10 000 pulse endurance cycles. Importantly, low nonlinearity values of 2.08 and 2.95 for potentiation and depression, respectively, with an asymmetric ratio of 63.8%, lead to a high recognition accuracy of 90.26% in the handwritten image classification task based on an artificial neural network simulation. The PFDD-SWCNT hybrid structure establishes a robust platform for high-precision and reliable neuromorphic circuitry.