Realization of Multi‐Level State and Artificial Synapses Function in Stacked (Ta/CoFeB/MgO)N Structures

Abstract

Abstract Spintronic devices can realize multi‐state storage and be used to simulate artificial synapses or artificial neurons, which makes them have promising application prospect in the field of artificial neural networks (ANN). This work investigates the current‐induced magnetization reversal in stacked (Ta/CoFeB/MgO)N structures and their application in ANN. It is demonstrated that the complete current‐induced magnetization reversal with large intermediate transition region can be achieved in the sample with N = 2. The magneto‐optical Kerr microscope imaging shows that the large transition region for the sample is ascribed to the “layer‐by‐layer” reversal, owing to the difference of the coercivity of two CoFeB layers. In addition, the simulation of artificial synapses and artificial neurons function based on current‐induced magnetization reversal in the sample is also demonstrated. These results substantiate the stacked (Ta/CoFeB/MgO)N structures as a promising platform for realizing the multi‐level state and artificial synapses function, and its potential application in the field of ANN.

Keywords

• artificial synapses
• memristors
• multi‐state storage
• spin orbit torque