Deterministic domain selection of antiferromagnets via magnetic fields

In the world of digital technology, information is stored and processed using tiny magnetic switches. For decades, engineers have relied on ferromagnets, materials like iron where all the atomic magnets point in the same direction, to act as these switches. However, these materials are sensitive to stray magnetic fields, which can accidentally flip the data, and they generate their own magnetic noise that interferes with neighboring components. Scientists have long looked to a different class of materials called antiferromagnets as a superior alternative. In these materials, the atomic magnets are arranged in a strict pattern where neighbors point in opposite directions, canceling each other out. This makes them invisible to external magnetic fields and incredibly stable, while also allowing them to switch states at speeds far faster than current technology. The major hurdle has been control: because the magnetic fields cancel out inside the material, applying an external magnetic field usually does nothing to change the internal arrangement. It is as if trying to push a door that is locked from the inside; the force simply passes through without effect.

A team of researchers, Sophie F. Weber and Veronika Sunko, has now identified a way to bypass this lock, revealing that the key lies not in the bulk of the material, but in its very edges. They focused on a specific antiferromagnetic material known as MnBi2Te4, which is made of stacked layers of atoms. While the interior of this material behaves as expected, the surfaces where the material meets the air or a substrate behave differently. The researchers discovered that the magnetic connections between the atoms at the very top and bottom surfaces are often weaker or stronger than those in the middle. By exploiting this subtle difference, they found a way to use a magnetic field to deterministically select one specific magnetic state over another, a feat previously thought impossible for materials with no net magnetic moment.

The researchers approached this problem by building a detailed computer model of the material, simulating how the atomic spins behave when a magnetic field is applied and then slowly removed. They started with the material in a high-field state where the atomic spins were tilted, a condition known as the spin-flop phase. As they gradually reduced the magnetic field strength, they observed a critical intermediate stage. In a perfectly uniform material, the system would remain in a state of indecision, with equal probability of settling into either of the two possible magnetic arrangements. However, in their simulations, they introduced a slight asymmetry: they made the magnetic connections at the top surface different from those at the bottom.

This small difference changed everything. As the magnetic field decreased, the atoms at the surface with the weaker magnetic connections were the first to react. They became pinned to the direction of the field, effectively acting as an anchor. This anchoring forced the rest of the material to organize itself in a specific way to accommodate the pinned surface. The result was a "bilayer" state, where a pair of atomic layers at the surface became aligned with the field while the rest of the material remained in its alternating pattern. This intermediate state was not random; it was dictated by which surface had the weaker connection. Once this specific state was formed, the path to the final magnetic arrangement became clear. When the field was reduced to zero, the material would inevitably settle into the magnetic domain where the surface spin pointed opposite to the original field direction.

Crucially, the researchers found that the direction in which the magnetic field was ramped down determined the outcome. If the field was decreased from a positive value, the material selected one domain; if it was ramped up from a negative value, it selected the opposite domain. This allowed for a deterministic selection, meaning the outcome was predictable and controllable, rather than random. The mechanism relies entirely on the surface properties, not on any net magnetic moment within the material. This distinguishes it from other methods that require the material to have a slight imbalance in magnetism, which is often difficult to achieve or control.

To verify that their theoretical model matched reality, the researchers compared their simulation results with recent experimental data on MnBi2Te4. They simulated the signal that would be detected by a magneto-optical Kerr effect experiment, a technique that uses light to measure magnetic states. The simulation reproduced the key features of the experimental data, including a hysteresis loop where the signal depends on the history of the magnetic field. The model successfully predicted two distinct critical field values observed in experiments: one where the material transitions from a tilted state to a multidomain state, and another where it settles into a single magnetic domain. The agreement between the simulation and the experiment was strong, suggesting that the mechanism of surface exchange asymmetry is indeed the driver behind the observed behavior.

The study also explored how robust this mechanism is across different conditions. They found that the effect works best when the surface with the weaker magnetic connection is the one exposed to the light in the experiment, which aligns with how the material is typically prepared. They also determined that the thickness of the material matters. In very thin films, the effect is strong even with small differences in surface properties. As the material gets thicker, approaching a bulk-like state, a larger difference in surface properties is required to maintain the deterministic control. However, the researchers calculated that the necessary difference remains within the range of what is physically realistic, even for materials that are hundreds of layers thick. This suggests that the phenomenon is not limited to ultra-thin films but could be a general feature of many layered antiferromagnets.

The implications of this finding extend beyond just understanding MnBi2Te4. The researchers argue that this mechanism is likely common in many experimental systems because surfaces are rarely identical. Whether due to the way a crystal is grown on a substrate, the presence of defects, or exposure to air and oxidation, the top and bottom surfaces of a material almost always have different chemical environments. These differences naturally lead to variations in magnetic exchange, providing the necessary asymmetry for this control method. This means that magnetic field-based control of antiferromagnetic domains might be achievable in a wide variety of materials without the need for complex engineering or the introduction of extra magnetic layers.

The work provides a clear path forward for the development of antiferromagnetic spintronics. By understanding that the surface plays a decisive role, scientists can now design experiments and devices that intentionally manipulate surface properties to achieve reliable switching. The researchers suggest that future work should focus on identifying the specific materials and conditions where this mechanism is most effective, potentially using controlled oxidation or specific substrate choices to tune the surface exchange. This approach offers a route to controlling the magnetic state of materials that were previously considered too stable or too insensitive to be useful for logic and storage applications. The discovery turns a perceived weakness—the sensitivity of surfaces to their environment—into a powerful tool for controlling the magnetic heart of next-generation computing devices.