Engineered Inclined Energy Landscapes Enable Free Flow of Magnetic Microstructures for Spintronic Neuron Applications

The quest to build brain-inspired computing systems has focused increasingly on spintronics, a technology leveraging the unique properties of magnetic materials, but creating reliable and efficient devices remains a significant hurdle. Anmol Sharma, Ranjeet Kumar Brajpuriya, and Vivek K. Malik, from UPES Dehradun and the Indian Institute of Technology Roorkee, alongside Vishakha Kaushik and Sachin Pathak, now demonstrate a breakthrough in controlling the movement of magnetic microstructures. Their work engineers a specifically shaped energy landscape for these structures, enabling remarkably free and predictable motion, and successfully mimics the fundamental ‘integrate and fire’ function of biological neurons. This innovative design requires minimal energy, just 23. 66 femtojoules per spike, and represents a crucial step towards building low-power, skyrmion-based devices for future computing applications.

Skyrmions for Neuromorphic Computing and Memory

Magnetic skyrmions, nanoscale swirling magnetic textures, are attracting significant attention for their potential in next-generation technologies. These structures possess topological protection, meaning they are remarkably stable and resistant to deformation, making them ideal candidates for advanced applications. Researchers are actively investigating skyrmions for use in high-density, low-power data storage and, crucially, for building more efficient artificial intelligence systems through neuromorphic computing, an emerging field aiming to mimic the structure and function of the human brain to create faster and more energy-efficient computers. A key focus of current research lies in understanding how skyrmions move and respond to external stimuli, such as electric currents and magnetic fields.

The Dzyaloshinskii-Moriya interaction, a fundamental property of certain materials, plays a critical role in both the formation and stability of these structures. Scientists are also exploring how gradients in magnetic anisotropy can be used to precisely control skyrmion motion, with techniques like spin-orbit torque and spin transfer torque offering further methods for manipulation. The inherent stability of skyrmions, stemming from their topological nature, provides a significant advantage over other potential building blocks for future devices. Research in this area utilizes a variety of materials, including layered structures of platinum, cobalt, and palladium, as well as combinations of cobalt, iron, and magnesium oxide.

These materials are chosen for their ability to support the formation and manipulation of skyrmions. Researchers employ advanced techniques like Lorentz transmission electron microscopy to directly visualize these nanoscale structures and magnetometry to characterize their magnetic properties. They also utilize simulations to model skyrmion behaviour and predict their response to different stimuli, driving progress in the field through this combination of experimental and theoretical approaches.

Engineered Anisotropy Creates Sawtooth Energy Landscape

Scientists have developed a novel method for controlling magnetic microstructures within nanotracks, achieving exceptionally low energy consumption for potential neuromorphic computing applications. By meticulously tailoring the magnetic anisotropy within a cobalt-platinum nanotrack using simulations, researchers created a sawtooth-type energy landscape featuring strategically placed pinning locations to trap microstructures and inclined paths to enable their free-flow motion, effectively mimicking the behaviour of synapses in the brain. The precision of this design demonstrates its experimental feasibility. The core of this work involves solving the Landau-Lifshitz-Gilbert equation to determine stable configurations of the magnetic material under external fields, accounting for the influence of the effective magnetic field and spin transfer torque arising from electric current within the nanotrack.

To achieve controlled depinning, scientists employed carefully calibrated current pulses, triggering motion from the pinning locations along the inclined paths, allowing for spontaneous propagation of the microstructures and reducing the need for continuous external forces. The resulting system demonstrates remarkably low energy consumption, achieving 23. 66 fJ per spike, a significant improvement over existing approaches. Researchers validated the design through extensive micromagnetic simulations, meticulously modelling the magnetic interactions within the cobalt-platinum nanostructure. By precisely controlling the anisotropy gradient, the study successfully emulated the integrate-and-fire function of biological neurons, paving the way for the development of energy-efficient skyrmion-based devices for future computing applications, with materials compatible with focused ion beam fabrication and nanoscale resolution ensuring its potential for practical implementation and scalability.

Skyrmion Motion Emulates Neuron Function

Researchers have made significant progress in emulating biological neuron function using spintronic devices, specifically by manipulating magnetic microstructures known as skyrmions. This work demonstrates a design that requires less energy than existing approaches and allows for controlled movement of these microstructures, paving the way for advanced computing architectures. By engineering a sawtooth-type energy landscape, scientists successfully emulated the integrate-and-fire function of biological neurons, a crucial step towards neuromorphic computing. The team meticulously investigated skyrmion stability within nanoscale tracks, identifying optimal conditions for sustained motion and defining a phase diagram mapping regions of stability based on the interplay between the anisotropy constant and the Dzyaloshinskii-Moriya interaction strength.

The red region of this diagram denotes conditions where skyrmions remain stable during motion, achieved through a precise balance of these parameters. Further analysis focused on the influence of inclined energy landscapes on skyrmion dynamics. Using the Thiele formalism, researchers modeled skyrmion motion as a rigid particle subject to various forces, deriving an equation demonstrating the direct relationship between the velocity of the skyrmion and the anisotropy gradient. Measurements confirm a low energy consumption of 23. 66 fJ per spike, demonstrating the potential for highly efficient computing and establishing a foundation for developing skyrmion-based devices with significantly reduced energy demands and enhanced functionality.

Sawtooth Landscape Drives Neuronal Magnetic Computation

This research demonstrates a novel approach to controlling magnetic microstructures, such as skyrmions and domain walls, for potential use in brain-inspired computing. Scientists successfully engineered an energy landscape with a sawtooth profile, enabling both the free flow and controlled pinning of these microstructures, crucial for emulating neuronal function. By modifying the magnetic anisotropy, the team achieved an energy-efficient method for driving these structures, requiring only 23. 66 fJ per spike, and effectively replicated the integrate-and-fire behaviour of biological neurons. The work establishes four distinct resistive states based on the engineered anisotropy, providing a foundation for building more complex computational devices.

Simulations reveal that increasing the anisotropy modification also increases the current required to release the pinned microstructures, demonstrating precise control over their movement. This method offers a pathway towards low-power, energy-efficient spintronic devices applicable to logic, memory, and, importantly, brain-inspired computing architectures. The authors acknowledge that further research is needed to validate these findings experimentally using advanced fabrication techniques and to explore the scalability of this approach for creating more complex neural networks. Despite these limitations, this research represents a significant step forward in harnessing the potential of magnetic microstructures for future computing technologies.