Analytical Approach to Engineer Strain-Gradient for Magnetic Skyrmion-based LeakyIntegrate and Fire Neuronal Dynamics

Magnetic skyrmions, with their chiral spin textures and topological stability, hold great promise for advancing energy-efficient spintronic technologies. However, their practical integration into devices is hindered by conventional driving methods, such as spintransfer torque (STT), spin-orbit torque (SOT), and current-induced magnetic fields, which introduce Joule heating and electrical noise, compromising the intrinsic non-volatility of magnetic systems. To address this limitation, we propose a novel strain-gradient-based mechanism to drive skyrmion motion without relying on electrical currents. We found that directional strain gradients can reliably manipulate skyrmion dynamics, enabling controlled propagation with an opposite skyrmion Hall effect depending on the type of strain. Specifically, tensile and compressive strain gradients on a nanotrack generate skyrmion velocities of up to 80 m/s and 60 m/s, respectively, at low damping, with motion directed from high to low strain regions (and vice versa). Leveraging this mechanism, we further demonstrate a current-free leaky integrate-and-fire (LIF) neuron, where tensile strain induces integration and compressive strain triggers leak behavior, both controlled via voltage polarity. This approach achieves ultra-low energy consumption of 24.06 fJ per neuron spike, which is ~3.8× lower than previously reported DMI-gradient-driven skyrmion LIF neurons. These findings offer a significant step toward low-power, current-free neuromorphic computing using strain-engineered skyrmionics.