Oxygen Incorporated MoS2 for Rectification‐Mediated Resistive Switching and Artificial Neural Network

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Molybdenum disulfide (MoS2) based memristor presents intriguing chance for the implementation bio‐inspired artificial synapse and neural interactions. However, the electrical properties of intrinsic MoS2 single‐crystal film suffer from inevitable lattice defects or structure vacancies as well as restraining reliable resistive switching mechanism. Here, a fundamental study on tunable p‐type doping in MoS2 film by controllable oxygen passivation is reported. In situ KPFM measurements reveal a near‐linear increase in Fermi‐level energy. The TiN/O‐MoS2 memristor exhibits surprising bipolar switching feature, revealing an interesting transition between rectification‐mediated and conduction‐mediated characteristics by means of controllable surficial oxygen kinetics. The resistive window demonstrates nearly symmetric SET and RESET domain and capability to analog programming conductance. Moreover, based on the MoS2 memristor, the accuracy up to 94% for MINST recognition ensures the implementation of neural network and LTP/LTD behaviors. This as‐prepared memristor is capable of mimicking synaptic feature through regulated resistive mechanics and electrode contact interaction, which can be an up‐and‐coming strategy for bio‐realistic electronics.
Funding Information
  • Natural Science Foundation of Beijing Municipality (M22020)
  • National Natural Science Foundation of China (62022047, 52272160, 52002254)
  • Institute of Microelectronics
  • Chinese Academy of Sciences

This publication has 41 references indexed in Scilit: