<i>N</i>-Doped Graphene/MXene Nanocomposite as a Temperature-Adaptive Neuromorphic Memristor
Lijie Kou, Rad Sadri, Damilola Momodu, Edward P.L. Roberts, Muhammad Aniq Shazni Mohammad Haniff, Chaoxing Wu, Chang Fu Dee, Poh Choon Ooi
Abstract
Due to intensive integration and seamless continuous operation, the overheated artificially intelligent (AI) integrated circuit systems will affect the operation system’s effectiveness, stability, and lifetime. Therefore, we proposed a temperature adaptability memristor in the silver nanowires (AgNWs)/nanocomposite/indium–tin-oxide structure in this study. The nanocomposite is the nitrogen-doped graphene/Ti 3 CNT x MXene blend in the polyvinylidene fluoride matrix. The device has been prepared by using heterostructure nanocomposites with a low-cost and facile all-solution method. The device mimicked a series of trained behaviors inherent with biological synapses, including spike-timing-dependent plasticity, paired-pulse facilitation, short-term potentiation/depression, long-term potentiation/depression, and excitatory postsynaptic currents. In addition, the device demonstrated significant self-adaptability to temperature due to the involvement of the homogeneously distributed conductive heterostructure and the filament formation ascribed to the low melting point of AgNWs. The operation of the device shows temperature adaptability owing to the excellent thermal conductivity and small thermal expansion coefficient of nitrogen-doped graphene/Ti 3 CNT x . This finding provides viable strategies to address the critical challenges of deploying AI in different environments, paving the way for the development of more efficient and resilient neuromorphic computing systems.