Scientists Discover New Spin Transfer Mechanism in Antiferromagnets

Researchers have discovered a physical effect that expands the fundamental understanding of how electric currents interact with magnetism, revealing an unexpected way to manipulate the magnetic state of antiferromagnetic materials.

A research team led by Northwestern Engineering’s Pedram Khalili has discovered what it calls octupole-driven spin-transfer torque (OTT). The effect allows an electrical current to manipulate the magnetic state of a class of materials called noncollinear antiferromagnets.

The researchers demonstrated the effect at room temperature in nanoscale devices constructed from two antiferromagnetic layers separated by a thin insulating barrier. The researchers call the structure an all-antiferromagnetic tunnel junction (AATJ).

“OTT could lead to higher-density, more reliable memories, which are very important for applications like AI,” Khalili said. “That by itself would be hugely impactful. At a more fundamental level, our findings provide a new understanding of how electric current interacts with magnetic order in a class of materials called spin-split antiferromagnets, which combine useful properties of ferromagnetic and antiferromagnetic materials within the same material system.”

Khalili is a professor of electrical and computer engineering at the McCormick School of Engineering. He is also Co-Director of the Applied Physics Program at Northwestern University. The findings were presented in the paper “Octupole-Driven Spin Torque Switching of Antiferromagnetic Tunnel Junctions,” published Oct. 21 in Nature.

Modern spintronics uses not only the electrical charge of electrons but also their spin to store, manipulate, and read information. Conventional spintronic devices typically rely on ferromagnets, materials such as iron, cobalt, and nickel in which electron spins align in the same direction.

That alignment creates a net magnetization with north and south magnetic poles. It also means an external magnetic field can interfere with information stored in a ferromagnetic material.

Antiferromagnets work differently. Neighboring spins point in different directions and cancel one another, leaving the material without a net magnetization even though its spins remain highly ordered at the microscopic level.

“These are magnets that have no magnetic dipole associated with them,” Khalili said. “If you store information in a ferromagnet, a large enough magnetic field nearby can destroy the information. None of that would happen with antiferromagnets.”

That resistance to magnetic fields could be useful in everyday electronics. Khalili pointed to magnetic memories in devices such as smartwatches, where magnetic wristbands, beads, or connectors can create nearby magnetic fields. An antiferromagnetic memory would not face the same vulnerability.

But exploiting antiferromagnets has presented another problem. For decades, researchers understood important spintronic effects as depending on electrical currents in which electron spins have a net alignment. Because the opposing spins in an antiferromagnet cancel each other, conventional thinking suggested those effects should not work in the same way.

“The conventional wisdom was that you should not have spin-transfer torque or giant or tunneling magnetoresistance with these sorts of materials,” Khalili said. “And this is basically what we show is not true.”

Khalili demonstrated two complementary capabilities in the antiferromagnetic devices.

One is a large tunneling magnetoresistance effect at room temperature, which provides an electrical way to distinguish between magnetic states and read information encoded in the material. 

The other is OTT, which enables a current passing through one antiferromagnet to exert torque on another and switch its state, effectively providing a way to write information.

“We show that there are antiferromagnetic materials that exhibit both of these effects,” Khalili said. “You can make a tunnel junction with very large tunneling magnetoresistance at room temperature, as large as with ferromagnets, and you can use a current to exert a torque on another antiferromagnet and switch it.”

Explaining why the switching occurred required the researchers to reconsider the underlying physics. Rather than relying on the net spin polarization associated with conventional spin-transfer torque, the effect is governed by the material’s magnetic octupole, a quantity that describes the more complex arrangement of spins in the noncollinear antiferromagnet.

“We report the first experimental observation, but this torque was really something that had not even been theoretically predicted before,” Khalili said. “Our paper also provides a new theoretical framework, which explains why this effect happens. The discovery of OTT is a step towards a broader understanding of how magnets interact with electric currents, which is no longer based only on magnetic dipoles.”

Mohammad Hamdi, a postdoctoral scholar and co-first author of the paper, agreed.

“When we first saw the switching, we still had to figure out what was happening at the microscopic level. That meant taking on the exciting challenge of developing a theoretical framework to explain the effect,” Hamdi said. “That understanding now gives us a foundation for designing devices that could take advantage of the ultrafast dynamics of antiferromagnets for memory and computing.”

“Our work now shows that we can both write and read information electrically in these junctions at room temperature, respectively by using octupole-driven spin-transfer torque and tunneling magnetoresistance,” added Jaimin Kang, a postdoctoral scholar who is also a co-first author of the paper. “It is exciting to be part of this progress, and I am looking forward to seeing where these discoveries take us in the next decades.”

Memory is only one possible application.

The discoveries of giant magnetoresistance, tunneling magnetoresistance, and spin-transfer torque opened a new field and set a whole community’s direction for 20 to 30 years. It was a long-standing question whether similar effects could be realized with antiferromagnets, and we show that the answer is yes. It is exciting to see where this new understanding will lead us. Pedram Khalili

Antiferromagnets can respond to electrical stimulation on much shorter timescales than conventional ferromagnets. While ferromagnetic dynamics typically occur at gigahertz frequencies, antiferromagnetic dynamics can extend into the terahertz range.

That makes them potentially useful for technologies ranging from high-frequency communications and computing to terahertz detectors and oscillators. The new work does not demonstrate terahertz operation, but Khalili sees that as one avenue for further research.

“With these effects that we have discovered, we may be able to, for the first time, electrically interact with these dynamics in antiferromagnets in an efficient manner,” Khalili said. “We knew the dynamics were there, but so far, you couldn’t really build a device with them because you didn’t have large effects that would allow you to excite those dynamics with a current or detect them with a voltage.”

The team’s next steps include testing other noncollinear antiferromagnetic materials to determine whether they produce stronger effects and investigating whether the devices can directly access terahertz dynamics.

Researchers also could explore integrating all-antiferromagnetic tunnel junctions with conventional semiconductor electronics. Khalili said such combinations could eventually be used to create devices that are more resistant to magnetic fields, operate faster, or increase computing throughput.

The study builds on nearly a decade of Khalili’s research at Northwestern into electrically controlling antiferromagnets. His lab, the Physical Electronics Research Laboratory, previously investigated different mechanisms for manipulating the materials and developed progressively more effective ways to detect their magnetic states, work that helped lay the groundwork for the new nanoscale device.

“I think this is really a foundational discovery,” Khalili said. “The discoveries of giant magnetoresistance, tunneling magnetoresistance, and spin-transfer torque opened a new field and set a whole community’s direction for 20 to 30 years. It was a long-standing question whether similar effects could be realized with antiferromagnets, and we show that the answer is yes. It is exciting to see where this new understanding will lead us.”