Voltage-Controlled Skyrmion-Based Leaky Integrate and Fire Neurons for Spiking Neural Networks

Spintronics is an emerging technology for data storage and computation. Magnetic skyrmion-based devices are attractive in this field due to their small size and low energy consumption. However, controlling skyrmions' creation, deletion, and motion is challenging. In this article, we propose a novel energy-efficient skyrmion-based device structure and demonstrate its use as a leaky integrate and fire (LIF) neuron for neuromorphic computing. Using micromagnetic simulations, we show that skyrmions can be confined by patterning the geometry of the free layer (FL) in a magnetic tunnel junction (MTJ) and demonstrate that the size of the skyrmion can be adjusted by applying pulsed voltage stresses, and when tuned, the device acts as an LIF neuron. The input voltage spikes at the input terminal control the spike rate. The MTJ dissipates energy around 10.23 fJ/spike. A spiking neural network (SNN) of such skyrmion-based LIF neurons can classify images from the Modified National Institute of Standards and Technology (MNIST) dataset.