Octupole-driven spin torque switching of antiferromagnetic tunnel junctions

Antiferromagnets have long been described as the quiet members of the magnetic materials family. They produce no stray fields, they are robust against external magnetic disturbances, and their spins can in principle oscillate at terahertz frequencies. Yet that same quietness has made them difficult to read and write. A new report in Nature on octupole-driven spin torque switching of antiferromagnetic tunnel junctions suggests that this barrier may finally be giving way.

The work shows that all-antiferromagnetic tunnel junctions can display both a large tunnel magnetoresistance (TMR) and a reciprocal switching effect, octupole-driven spin-transfer torque. Together, these two properties point toward nanoscale magnetic memory and potentially room-temperature terahertz spintronic devices.

Why Antiferromagnets Matter for Future Spintronics

Conventional spintronic memory, such as magnetic random-access memory, relies on ferromagnets. Information is stored in the magnetization direction of a thin layer, and it is read through the resistance of a magnetic tunnel junction. This technology is mature, but ferromagnets have limits in speed and in how densely devices can be packed.

Antiferromagnets offer a different path. Their neighboring spins cancel one another, so there is almost no net magnetization. Bits stored in such materials would not interfere with neighbors, which could allow much tighter integration at the nanoscale.

Their intrinsic dynamics are also far faster. While ferromagnets typically operate in the gigahertz range, antiferromagnetic resonances reach into the terahertz regime, opening a route to ultrafast memory and signal-processing components.

The Challenge of Reading Zero Net Magnetization

The very feature that makes antiferromagnets attractive also makes them hard to use. With no net magnetization, standard methods of detecting the magnetic state do not work well. Researchers have therefore struggled to obtain strong electrical readout signals from antiferromagnetic order.

Earlier attempts relied on effects such as anisotropic magnetoresistance, which produce only small signal changes. These signals are usually too weak for practical memory devices, where a clear difference between the 0 and 1 states is Essential.

A large TMR, the kind routinely achieved in ferromagnetic junctions, has been a long-sought goal for antiferromagnetic systems. Achieving it would put electrical readout on a comparable footing with established technology.

Tunnel Junctions Built Entirely from Antiferromagnets

The new study focuses on tunnel junctions in which both electrodes are antiferromagnetic, separated by a thin insulating barrier. In this configuration, the relative orientation of the antiferromagnetic order in the two electrodes determines how easily electrons tunnel through, producing a measurable resistance difference.

The key finding is that these all-antiferromagnetic junctions can display a large TMR. This demonstrates that electrical readout of antiferromagnetic states does not have to be weak, provided the materials and symmetry conditions are chosen appropriately.

Such a result matters because it removes one of the main objections to antiferromagnetic memory. A device with no stray field and strong readout is a considerably more convincing candidate for real applications.

The Role of Magnetic Octupoles in Switching

Central to the work is the concept of the magnetic octupole. In certain noncollinear antiferromagnets, the arrangement of spins gives rise to a higher-order magnetic multipole that acts as an effective order parameter, behaving somewhat like magnetization in a ferromagnet even though the net moment is nearly zero.

This octupole order can influence the electronic structure and spin-dependent transport. As a result, it can generate spin-polarized currents, which in turn can be used to exert torque on a second antiferromagnetic layer.

That mechanism underlies the octupole-driven spin-transfer torque described in the paper. It provides a way to manipulate antiferromagnetic order electrically, instead of relying on external magnetic fields.

A Reciprocal Relationship Between Readout and Writing

One striking aspect of the results is reciprocity. The same physics that yields a large TMR for reading the state also allows spin-transfer torque to switch it. In other words, the coupling between spin-polarized charge transport and antiferromagnetic order works in both directions.

This echoes the well-known relationship in ferromagnetic junctions, where TMR and spin-transfer torque are linked. Seeing a similar pairing in antiferromagnets suggests that the established design logic of spintronics can be carried over to this new material class.

For device engineers, such a link is valuable. A single junction structure could serve as both the writing and reading element, simplifying memory cell design.

Implications for Nanoscale Magnetic Memory Devices

If antiferromagnetic tunnel junctions can be switched by current and read with a large resistance contrast, they could form the basis of a new type of nonvolatile memory. The absence of stray fields would let cells be placed very close together without crosstalk.

Scaling is an important advantage. Because antiferromagnetic bits are insensitive to external magnetic fields, they may be more stable against disturbance, a useful property for dense and reliable storage.

Switching speed is another potential benefit. Antiferromagnetic dynamics are intrinsically fast, so writing operations could eventually be far quicker than in present ferromagnetic devices, although practical speeds will depend on engineering details.

Toward Room-Temperature Terahertz Spintronic Technologies

The study also highlights the potential for terahertz operation at room temperature. The terahertz range is hard to access with conventional electronics, and sources and detectors in this band remain bulky or inefficient.

Antiferromagnetic tunnel junctions might serve as compact oscillators, detectors, or modulators in this regime. Their natural resonance frequencies and electrical controllability make them appealing building blocks for such devices.

Applications could include high-speed communications, sensing, and imaging. While these remain prospects rather than demonstrations, the new results provide a concrete foundation on which to build.

Open Questions and the Road Ahead

Despite the excitement, significant work remains. Researchers will need to optimize materials, interfaces, and barrier quality, and they must confirm that the effects persist robustly across many devices and operating conditions.

Energy efficiency is another question. Switching currents, endurance, and reproducibility will need to meet industrial requirements before antiferromagnetic memory can compete with existing technologies.

Even so, the demonstration that large TMR and octupole-driven spin-transfer torque can coexist in all-antiferromagnetic junctions marks a significant step. It moves antiferromagnetic spintronics from a promising concept closer to a practical technology, and the field will be watching closely to see how quickly the idea develops.

If you have any questions, please don't hesitate to Contact Us