Optoelectronic Synapses Based on Confined Growth of Ruddlesden–Popper Perovskite Nanocrystals in Carbon Nanotubes

Two-dimensional Ruddlesden–Popper perovskites (RPPs) are emerging materials for neuromorphic optoelectronics owing to their tunable band gap, strong light–matter interaction, and environmental stability. Here, we demonstrate the in situ growth of RPP nanocrystals inside multiwalled carbon nanotubes (MWCNTs), forming a hybrid nanocomposite with enhanced charge transport and moisture resistance. This confined growth strategy provides intimate interfacial contact between the perovskite and CNT walls, facilitating efficient photoinduced charge transfer. The resulting two-terminal RPP/MWCNT device exhibits key synaptic functionalities like paired-pulse facilitation, long-term potentiation/depression, and memory retention over 1000 s under low optical power density. Using experimentally derived synaptic weight dynamics in a convolutional neural network model, we achieved 92% and 85% recognition accuracy on MNIST and Fashion-MNIST datasets, respectively. Our results introduce a stable and energy-efficient perovskite–carbon hybrid platform that bridges fundamental material design and functional neuromorphic applications in artificial vision systems.