Voltage-Controlled Spintronic Stochastic Neuron Based on a Magnetic Tunnel Junction

Stochastic units based on magnetic tunnel junctions have shown a high energy-efficient pathway to perform neuromorphic computing. We propose a voltage-controlled spintronic stochastic device based on a magnetic tunnel junction by introducing perpendicular anisotropy into the free layer. We observe a stochastic behavior at low bias and demonstrate that this behavior can be used to mimic the artificial neurons of an artificial neural network to recognize the handwritten digits in the Mixed National Institute of Standards and Technology (MNIST) database. Furthermore, the stochastic behavior can be modulated by a bias voltage owing to the voltage-controlled magnetic anisotropy effect. The voltage-controlled stochastic behavior is theoretically and experimentally studied, which indicates that it can be used to perform as an adaptive neuron. These results provide a way for building energy-efficient spintronic neuromorphic computing systems.