Enabling double mott transitions at high temperatures in Ti–VOx formed by proximity-oxidation for neuromorphic image recognition based on CNN-LeNet

Abstract

Mott transition is one of the most dramatic physical phenomena involving abrupt insulator-to-metal transition (IMT) in some metal oxides. Recently, it has become one of the highlighted research topics as various Mott devices have been investigated for emerging memory and steep switching transistor applications. That has inspired us to create a proximity-oxidation-produced Ti–VOx thin film sputtered on n-Si substrate. The key finding here is the mixed phase of the Ti–VOx film and the interface of Ti–VOx/n-Si enabling a double Mott switching characteristic – in other words, a Ti–VOx device can switch from two HRS regions to two LRS regions. Particularly, the Ti–VOx/n-Si interface plays two roles: supporting the double Mott switching and facilitating a controlled Mott transition behavior up to 120 °C. Our research demonstrated by a deep learning-based modeling technique, yielding a staggering recognition accuracy of 97.78 %. The result is within the theoretically ideal recognition rate of 99.37 %. Due to the remarkably endurant synaptic weight behavior up to 120 °C, the usability of our Ti–VOx devices for synaptic applications is demonstrated and discussed in detail to pave a way for uses in advanced neuromorphic applications.