Neuromorphic weighted sums with magnetic skyrmions

We explored the properties of magnetic skyrmions—nanoscale, stable isolated magnetic textures—and used them to mimic the role of biological neurotransmitters in a neural network. Taking advantage of the particle-like behavior of skyrmions, which makes them countable and hence summable, enabled us to perform a basic neuromorphic computing operation: the weighted summation of synaptic signals. We demonstrate the precise control of the number of skyrmions, determined by electrical pulse inputs multiplied by the track synaptic weights. Magneto-ionic effects enable non-volatile and reversible tuning of magnetic properties, allowing gate voltage manipulation of synaptic weights. Detection of the skyrmion number is accomplished through non-perturbative anomalous Hall voltage measurements. We experimentally validate the weighted sum operation using two electrical inputs in a crossbar array configuration with two tracks. This ensures efficient execution of the fundamental weighted sum operation. Our experimental demonstration is scalable to accommodate multiple inputs and outputs using a crossbar array design, potentially approaching the energy efficiency observed in biological systems.