Electric field doubles the speed of ultrafast magnetic processes

Modern electronic devices work by controlling the flow of electric current, that is the movement of electrons. However, electrons, in addition to their negative charge, also possess a magnetic property called spin, which is like a tiny magnet. In fact, these tiny electronic magnets together make a material magnetic. Combining electronics with magnetism has given rise to the field of spintronics, which uses the spin of electrons and has enabled groundbreaking advancements in data storage capacity in hard disk drives and got much smaller and much more effecient, from refrigerator-sized disk drives with a few MB memory storage capacity to coin-sized disk drives with TB memory storage capacity.

Increasing the speed of data storage can be transformational for reducing power consumption in data centres, such as Google Drive or BOX, and artificial intelligence infrastructure, like ChatGPT. This requires magnets to switch faster, and the challenge is to achieve this without altering the magnetic material or shape in devices.

For the first time, researchers have developed a new device called a spin field effect junction, SFEJ, that can double the speed of demagnetization of a magnetic thin film by an electric field^[1].

When a femtosecond laser pulse, lasting a millionth of a billionth of a second, hits a ferromagnetic cobalt layer, the magnet rapidly loses its magnetization in a process called ultrafast demagnetization. In the SFEJ, graphene, an atomically thin material composed of carbon atoms, is placed under an ultrathin layer of cobalt (as depicted in the top figure). Graphene is renowned for its exceptional charge and spin transport properties^[2] with high spin diffusion^[3]. As a result, in SFEJ, spins from the cobalt layer flow into the graphene layer as ultrafast spin currents. By applying an electric field, the team controlled these ultrafast spin currents, reducing cobalt’s demagnetization time from 200 to less than 100 femtoseconds, more than doubling the speed.

“Until now, speeding up magnetic processes meant changing the material itself or its structure. This is the first demonstration of electrically controlled ultrafast magnetism without any alteration,” says David Muradas-Belinchón, PhD student at the Department of Physics and Astronomy.

The spin-field effect junction device operates similarly to a transistor: the gate electrode applies an electric field to graphene, thereby shifting its electron density. Since graphene is in direct contact with cobalt, this tuning regulates how efficiently spins flow across the interface, providing a direct handle on ultrafast magnetism. The speed of demagnetization was measured using a femtosecond pump–probe laser system based on the time-resolved magneto-optical Kerr effect (TR-MOKE). A pump pulse excites cobalt and triggers demagnetization, while a time-delayed probe pulse reflects from the sample tracking its magnetization (the Kerr effect). By varying the time delay, the researchers were able to track how the speed of demagnetization changes as they applied the electric field.

Theoretical simulations, conducted by Francesco Foggetti, Postdoc at the Department of Physics and Astronomy, and Peter Oppeneer, Professor at the Department of Physics and Astronomy, confirmed these experimental results. Using superdiffusive spin transport theory^[4], they demonstrated that tuning the graphene–cobalt interface directly modulates the flow of spin angular momentum, consistent with the experimental results.

“This is a new paradigm. Magnetic devices are central for both non-volatile memory and logic, and speeding up these processes with an electric field opens exciting prospects for low-power AI hardware and quantum sensing,” says Venkata Kamalakar Mutta, Associate Professor at Uppsala University and senior author of the study.

By demonstrating electrical control of spin currents on femtosecond timescales, the study merges two previously separate areas: ultrafast magnetism and field-effect electronics. The ability to control magnets more quickly and with lower energy is critical for future devices that aim to combine speed, efficiency, and intelligence. This could pave the way for field-programmable magnetic memory, reconfigurable spin-logic, neuromorphic devices that mimic the brain, and precision quantum sensors.

Camilla Thulin