TWO-DIMENSIONAL MATERIALS-BASED SYNAPIC DEVICES FOR NEUROMORPHIC/IN-MEMORY COMPUTING APPLICATIONS

Abstract

Two-dimensional (2D) materials-based synaptic devices are showing great potential to build neuromorphic/in-memory computing hardware. Their unique switching mechanisms and novel heterostructures provide new perspectives for synaptic device applications. This thesis aims to realize the 2D materials-based synaptic devices and their crossbar arrays (CBAs). Firstly, several experimental techniques are developed to achieve ReS2-based synaptic memristors with low variation. Two novel 2D material post-treatment techniques, including electron beam irradiation and molybdenum irradiation, are investigated to stabilize the switching variation (6.3%/5.3%) and enhance the learning accuracy. Secondly, a wafer-scale HfSe2 molecular beam epitaxy (MBE) growth method and a wafer-scale metal-assisted van der Waals (vdW) transfer technique are developed for large-scale production of CBAs. Finally, the as-fabricated large-scale HfSe2-based CBA is integrated with peripheral circuits to achieve a full-hardware accelerator, exhibiting high-accuracy hardware multiply-and-accumulate (MAC) operation (93.34%) and convolution image processing with tight current distribution (0.25%).