This study explores the use of atomically thin two-dimensional (2D) transition-metal dichalcogenides (TMDCs), specifically WS₂, for synaptic transistors in neuromorphic computing. Niobium (Nb) atoms were substitutionally doped at the W sites in WS₂ to enhance the switch ratio. The resulting synaptic transistor devices exhibited a significantly improved switch ratio (10³), 100 times greater than undoped WS₂ devices. The Nb atoms facilitated electron trapping and detrapping, effectively simulating synaptic potentiation, inhibition, and learning processes. The Nb-WS₂ synaptic transistor achieved 92.30% recognition accuracy on the MNIST handwritten digit dataset.

Kejie Guan, Yinxiao Li, Lin Liu, Fuqin Sun, Yingyi Wang, Zhuo Zheng, Weifan Zhou, Cheng Zhang, Zhengyang Cai, Xiaowei Wang, Simin Feng, Ting Zhang