Influence of a Novel 2D Material MXene on the Behavior of Memristor and Its Crossbar Array

Stable and high-density electronic synapses are the most critical units for the hardware implementation of artificial neural network. However, the power consumption and stability of resistance states of electronic synapses still need to be further improved. In this work, we investigate synaptic resistance switching (RS) devices and its crossbar arrays with the addition of a novel two-dimensional material MXene (Ti
₃
C
₂
) using a silicon compatible process. As a result, stable OFF resistance states have been achieved as well as low operation voltage. In addition, data retention of TiN/Cu/SiO
₂
TiN devices has been measured for (2×10)
⁴
seconds. Furthermore, the conduction mechanism of TiN/Cu/SiO
₂
TiN devices has been analyzed in details revealing that ohmic conduction and space-charge-limited-conduction (SCLC) dominate at the low and high resistance state, respectively. It discloses that the reliability of OFF resistance states can be improved significantly by using MXene in the RS layer.