We’ve discovered a new kind of magnetism. What can we do with it?

Libor Šmejkal has a fondness for the artwork of M. C. Escher, whose work was often inspired by mathematics. One of Šmejkal’s favourite pieces is Horseman, a striking picture that features an elaborate, tessellating series of mounted figures. Strangely enough, it was this piece that inspired him to predict the existence of an entirely new kind of magnetism.

We have known of magnets for millennia. Today, they are at the heart of a raft of modern technologies, from electric generators and smartphones to loudspeakers and hospital scanners. And yet for 100 years, we have been missing something about them. We always assumed there were only two types. It was Šmejkal’s art-inspired insights that finally gave the lie to that in 2022.

Fast-forward to today, and we know that what Šmejkal called “altermagnets” aren’t just an idea. We have discovered real examples and are working out how to make this new kind of material in practical and useful ways. There is even a possibility that these magnets could help us build a completely new kind of computer. “Altermagnets could actually have all the functionalities of current devices, but much faster, with less energy consumption, and smaller,” says Šmejkal.

To understand magnetism and why it is so important, we need to start with the electrons that whizz around in atoms. Each of these particles has an intrinsic quantum property called spin. This isn’t quite like anything in the everyday world, but you can imagine it as a tiny spinning top that can rotate in two directions, which scientists label “up” or “down”.

Electrons like to orbit their atoms in pairs, one spin-up, one spin-down, so the spins cancel out. But that isn’t always possible, as an electron sometimes ends up alone. With nothing to cancel it out, the unpaired electron gives the atom what is known as a magnetic moment: like spin, it can either be up or down, and it governs the atom’s magnetism. If you have enough atoms with magnetic moments pointing in the same direction, they create a strong, directional magnetic field. This is called ferromagnetism. It is a phenomenon that even the ancient Greeks knew about – they found rocks that attracted iron nails or filings to them – and it is also what is going on inside the magnets many of us have stuck to our fridges.

Ferromagnets are easy to spot because they attract or repel other magnetic materials, like nickel or cobalt. But there is another, more subtle kind of magnetism that wasn’t discovered until the 1930s. Antiferromagnets also have magnetic arrows, but this time they point in alternating directions – picture a line of arrows going up, down, up, down and so on. The result is a magnetic stalemate, a solid with magnetic order on the atomic level, but no unified, detectable magnetism on the scale of ordinary objects.

The mental model of imagining tiny arrows pointing up and down inside magnets was invented by physicist Louis Néel, who theorised the first antiferromagnets – which were experimentally confirmed in the decades following – and who won a share of the 1970 Nobel prize in physics.

Magnetic symmetry

That’s how things have stood for nearly 100 years: two types of magnetism, nice and neat. It was 2018 when Šmejkal, who is now based at Johannes Gutenberg University Mainz in Germany, began to suspect there might be more to the picture. At the time, he was a young PhD student in Prague, Czech Republic, studying a strange phenomenon sometimes seen in antiferromagnets called the anomalous Hall effect. Šmejkal’s breakthrough was to realise that this effect and similar arcane magnetism puzzles couldn’t be explained with the model that Néel had developed – he needed to go beyond it.

This is where Escher’s 1946 horsemen artwork came in. The riders in the image slot together in alternating colours with an elaborate, beautiful symmetry. Take one of the lighter figures, flip it, shift it sideways a jot and change its colour, and you match one of the darker riders. As he mused on this, Šmejkal realised there was an alternative mathematical method for describing this symmetry operation. “I realised that you can actually define this operation, this changing of colour or orientation, in another way,” he says.

And here’s the thing: understanding symmetry has always been crucial in physics, and this is particularly true in materials science, where the intricate relationships between different kinds of atoms are best described in that language. Indeed, Néel’s way of thinking about atoms’ magnetic moments has symmetry at its core. But by using his new mathematics as a framework, Šmejkal began to extend Néel’s model, firstly by thinking in three dimensions instead of two and secondly by including atoms with no magnetic moment in the picture.

As he did so, a new possibility began to emerge. You could still have neighbouring atoms with magnetic moments that point in opposite directions: up, down, up, down, as in antiferromagnets. But every alternate atom would be rotated by 90 degrees, hence the name altermagnet (see diagram, below). Šmejkal says this rotation can happen as a result of magnetic atoms existing in a sea of non-magnetic atoms. Though the arrows still alternate in pointing up and down, the rotated atoms give rise to a subtle effect that enables some magnetism to leak through.

This had all started as an attempt to solve a particular set of puzzles in magnetism, but Šmejkal says it amounted to something much grander: it predicted that a whole new kind of magnetism was possible. Altermagnets would have no net magnetism, like antiferromagnets, but they would have some of the quantum properties that make ferromagnets so useful in technology. In 2022, Šmejkal and his colleagues published what he calls a “complete mathematical framework” of altermagnetism. “The whole community was quite excited because these systems seem to combine the prized advantages of ferromagnets and antiferromagnets,” he says.

Confirming that altermagnets exist

The world only had to wait two years before the prediction was confirmed. In 2024, Juraj Krempaský at the Paul Scherrer Institute in Villigen, Switzerland, and his colleagues studied manganese telluride, a compound thought to have the right structure to produce altermagnetism. To check if it did, they used light beams to track the precise movements of electrons inside the material – and these turned out to closely match simulations of what would be expected for an altermagnet.

The discovery of a third kind of magnetism is huge in its own right, but what makes it even more exciting is that it could solve a long-standing technological problem. To see why, we need to know a little about how computers store information. Today, they tend to do so in chips, essentially through the presence or absence of electric charge to signify a digital 0 or 1. But researchers have long been interested in the idea of using magnetism to store information, too – floppy disks, which were used in the 1990s, worked on magnetic principles. A more recent concept called spintronics takes things a step further: the idea would be to use not just the presence or absence of electric charge, but also the spin of the electrons too.

In theory, spintronics would enable us to cram much more information into computer memory, making it more efficient. But there has always been one big problem. For it to work, we need materials in which the up and down spins can be split into separate strands. Anna Hellenes, who works in Šmejkal’s university research group, likens it to a ballroom full of dancers. In a non-magnetic material, all the couples waltzing clockwise or counterclockwise – the electrons spinning up or down – remain mixed on the dance floor. “But if we now have spin-splitting, these dancers spinning in one direction can separate from the others spinning in the other, and dance separately,” she says.

The problem is that this spin-splitting effect, the bedrock of any spintronic device, was only found in ferromagnetic materials. This made sense because all the arrows in a ferromagnet point the same way, so electrons whose spin points in the direction of all those cumulative arrows are in a slightly different environment than those with spins pointing the other way. But if you try to cram lots of ferromagnets onto a tiny chip, they do exactly what you might expect: attract or repel each other. As a result, says Hellenes, spintronics has hit a ceiling.

Making spintronics

Could altermagnets step into the breach? “This unique combination of features from altermagnets — no net magnetisation, but still spin-split bands — could be very advantageous for potential spintronic devices,” said Igor Mazin, a physicist at George Mason University in Virginia.

Since it was confirmed that manganese telluride was altermagnetic in 2024, researchers have been busy trying to create new materials that have this curious property. One trick is to take a known antiferromagnet and apply mechanical strain to it in the hope of deforming the internal magnetic symmetry and coaxing altermagnetism into being. In 2024, researchers led by Atasi Chakraborty, a member of Šmejkal’s research group, demonstrated that applying compressive strain to rhenium dioxide – long known to be an antiferromagnet – triggers a transition into an altermagnetic state.

What’s more, a trio of researchers at the Beijing Institute of Technology in China realised that you can also create the right internal magnetic disturbances by stacking an antiferromagnet between layers of a different material, like a sandwich. The top and bottom layers induce internal electric fields that mimic the crystal environment of naturally occurring altermagnets.

However, researchers tend to feel that these clever tricks may not lead to scalable altermagnets anytime soon, as the methods are difficult to pull off. Instead, it seems more likely that we can find practical altermagnets by looking at naturally occurring ones. “For the vision for altermagnetism over the next 10 years, I could quite easily see these materials becoming commercially viable,” says Oliver Amin, a researcher at the University of Nottingham, UK, who created the first experimental image of manganese telluride after it was confirmed as an altermagnetic material. In a paper published in December, his team demonstrated that researchers could not only see the structures that gave this material its magnetic properties, but also control the direction and layout of them by heating and cooling the material in a magnetic field. “This is the first step towards realising these materials as practical materials for devices,” says co-author Alfred Dal Din at the University of Nottingham.

We have good computational models of the kinds of atomic structures that are likely to exhibit this new magnetism, and Šmejkal and his colleagues used them to digitally comb through possible materials. They have identified at least 200 candidates, published shortly after their landmark altermagnetism paper. Confirming all those candidates experimentally will take time, but we already know that, other than manganese telluride, there is also strong evidence that ruthenium dioxide is an altermagnet.

Other than being the only certified, bona fide altermagnet, manganese telluride is an established material that scientists know how to grow in the lab at high qualities – the primary hurdle for many experimentalists. “The form of manganese telluride we’re working on now has been studied in the form we’re looking at for at least 20 years, probably more,” said Amin.

A fourth kind of magnetism?

Just as researchers rush to get to grips with altermagnets, Šmejkal has another surprise up his sleeve. In a paper that hasn’t yet been peer-reviewed, he and his colleagues predict the existence of yet another kind of magnetism, which he calls antialtermagnetism.

In materials with this strange property, neighbouring spins don’t just alternate up and down like in an antiferromagnet, they also form zigzags. Picture tiny arrows lying next to each other, the first pointing north-west, then north-east, then south-east, then south-west – tracing out a zigzag. The neighbouring arrows are mirror images of each other, so that adding up the directions across all the mirrored pairs will cause them to cancel out, as happens in antiferromagnetism. But the mirrored pattern subtly reshapes how electrons move through the material in such a way that also causes spin-splitting, says Šmejkal.

The idea of antialtermagnetism builds on the complex and beautiful symmetries that Šmejkal was so taken by early on in his work.

Perhaps we can say magnets are like one of those Escher artworks that he likes so much – the more you look, the more delightful details you notice.