Review of applications of 2D materials in memristive neuromorphic circuits

Neuromorphic systems with large-scale parallel computing capability and low power consumption have become important for the development of artificial intelligence technologies. Memristors have been designed to achieve various computing functions and have been further applied in neuromorphic circuits owing to their high storage density, fast switching speed, ultra-low power consumption, and long endurance cycles. Among the different types of functional materials, 2D materials, which are a novel class of functional materials, have shown great potential in memristive neuromorphic applications because of their atomic-scale thickness, excellent electronic properties, thermal stability, mechanical flexibility, and strength. In addition, by stacking different 2D materials together, van der Waals (vdW) heterostructures retain not only the properties of each 2D material but also exhibit more interesting properties than their respective counterparts; therefore, vdW heterostructures are promising for flexible neuromorphic applications. In this review, we discuss the applications of 2D materials and their vdW heterostructures in memristive neuromorphic circuits from the perspective of material systems, physical mechanisms, advantages, and future challenges.