ニューロモルフィック応用向けデジタル・アナログスイッチングを可能にする有機エンジニアリングMXene

MXene-based devices have emerged as a promising platform for high-speed and multifunctional resistive switching electronics. However, their intrinsic tendency toward oxidation severely restricts practical application. Here, we report a robust strategy to address this challenge by hybridizing MXene nanosheets with a soluble organic semiconductor of BTCN. The BTCN molecules interact with MXene through π-π stacking and electrostatic coupling, enabling stable passivation and tailored interfacial electronic states. Memristors based on inert Au/MXene-BTCN/ITO structure exhibit reproducible digital-type resistive switching with narrow distributions of set/reset voltages, highly stable endurance over 3 × 10⁴ cycles, and ultrafast switching dynamics down to 5 ns. By replacing the top electrode with active Ag, the devices succeed to trigger digital-to-analog synaptic emulators, demonstrating polarity-dependent conductance modulation. Furthermore, a convolutional neural network (CNN) employing MXene-BTCN hybrid memristive arrays achieves a high digit recognition accuracy over 97 %. This work establishes a versatile design principle that couples solution-processable MXene-organic hybrids with electrode engineering, bridging digital memory and analog synaptic functionalities for brain-inspired computing applications.