Intrinsically Nonvolatile n-Type Polymer Synapses for Neuromorphic Vision.

Zhang, Tian-Yu; Li, Yuan-Kai; Yu, Zi-Di; Kong, Hua-Kang; Zhang, Xiao-Yan; Long, Yao-Xi; Wang, Jie-Yu; Yao, Ze-Fan et al. · Adv Mater · 2026

basic_science · Level V

Where this comes from

Abstract

Neuromorphic vision requires photonic synapses capable of integrating optical sensing and memory. However, achieving synaptic plasticity in photonic synapses typically relies on auxiliary charge-trapping components, resulting in increased processing complexity and poor substrate compatibility that limit large-area integration. Although single-component photonic synapses represent an attractive alternative, their development has been hindered by the lack of rational material-design strategies. Here, we introduce an asymmetric carrier trapping (ACT) strategy that enables intrinsic nonvolatile photoresponse in n-type conjugated polymers. Through tailored frontier-orbital energies and distributions, ACT combines a deep-lying, delocalized lowest unoccupied molecular orbital (LUMO) for efficient electron transport with a localized highest occupied molecular orbital (HOMO) for stable hole trapping. Guided by computational screening, we fabricated single-polymer photonic synapses across diverse substrates. Enabled by intrinsic ACT behavior, these synapses operate without elaborate interfacial engineering while simultaneously performing optical signal reception, storage, and brain-like synaptic functions. Furthermore, an 8 × 8 flexible synaptic array exhibited highly uniform performance with paired-pulse facilitation (PPF) indices of 137 ± 1%. An artificial neural network (ANN) constructed using these polymer synapses achieved an image-recognition accuracy of 93.3% under ambient conditions. These findings establish ACT as a promising molecular strategy for intrinsically nonvolatile polymer synapses toward neuromorphic vision.