Two Room-Temperature Antiferromagnetic Semiconductor Candidates

We’re all used to 2 kinds of magnet. The frequent one, the fridge magnet, is ferromagnetic — its atomic magnets all level the identical method (up or down), including their magnetic results. The much less well-known one, the antiferromagnet (AF), has neighbouring atomic magnets that time reverse methods and precisely cancel out magnetically.

For a very long time, there’s been an intense drive in pc reminiscence analysis to create supplies in between these two extremes. For this function, it helps to have a transparent image of what these extremes imply.

Today, I’ll share what we discovered. A crew of AI brokers and I designed one candidate magnet and located one other, first made in 1999, that our calculations predict has the properties we have been after.

But earlier than diving into the main points, let me first lay out a magnet primer that takes all of 90 seconds, assuming you aren’t an undergraduate in physics or chemistry.

A 90 second primer in magnets

Spin

Each electron has a quantum mechanical property known as ‘spin’, which is accountable for its magnetic second. We can mannequin the spin path for every electron as both pointing up or down.

Spintronics

We use spin for storage. A magnetised materials shops info primarily based on the spin up/spin down orientation of its electrons, in the identical method classical magnets retailer info primarily based on pointing up/down. The most outstanding instance of spintronics is the onerous drive learn head, the gadget that reads the magnetic bits on the disk. MRAM is one other kind of spintronics that makes use of the identical precept to retailer binary info in a non-volatile method.

In the globe of spintronics, we need to type electrons by their spin orientation so we will learn/retailer their info. Ferromagnets do that naturally: the electrons that carry present are largely of 1 spin, up or down. Ordinary antiferromagnets, nevertheless, can not distinguish up/down electrons.

This leads us into the subsequent part.

Three sorts of magnets

Ferro

Ferromagnetic supplies are characterised by having a macroscopic magnetic discipline, or a discipline that leaks out from the floor. This is why a fridge magnet sticks to your fridge door.

The drawback, nevertheless, is that the magnetic discipline interferes with close by supplies and is tough to regulate for storage functions. In addition, switching magnets forwards and backwards is comparatively gradual and consumes lots of energy.

Another characteristic of ferromagnetic supplies is that the spins are sorted (by up/down orientation) based on power stage. We can see this when wanting on the power spectrum of the electrons: close to the sides of the spectrum, the electrons all have the identical spin.

Antiferro

If we take the above and flip the logic, that means, the spins are unsorted, we find yourself with an abnormal antiferromagnetic materials.

Here the electrons of the identical power stage will as an alternative have blended spins. This leads us to 2 properties of curiosity:

• Because spins usually are not sorted based on power stage, it’s onerous to learn/retailer info utilizing spintronics methods.
• However, the shortage of macroscopic magnetic discipline permits us to pack these supplies nearer collectively, enabling greater efficiency for storage gadgets. In phrases of velocity, AF supplies are additionally a couple of thousand instances sooner to modify.

The solely drawback is, if the spins are blended at every power stage, how would we ever type them by power? How will we get a option to learn/retailer info utilizing spintronics with this lack of sorting?

Luttinger compensated

This brings us to a 3rd kind, Luttinger compensated (LC).

LC supplies are antiferromagnets the place the spin-up atoms and spin-down atoms have the identical magnitude of magnetism, making the online spin second zero (i.e. they cancel out). However, not like in abnormal antiferromagnets, the up and down atoms sit in inequivalent environments: they are often completely different components (e.g. one factor factors up and one other factors down), or the identical factor in two completely different sorts of web site. The identify comes from Luttinger’s theorem: in an insulator, the online spin second of every repeating unit of the crystal have to be a complete quantity, so as soon as it’s zero it stays locked at zero. Strictly, that holds for an ideal crystal close to absolute zero; smaller results equivalent to spin–orbit coupling, and warmth, can go away a slight imbalance.

Since the up and down atoms usually are not equal, up/down spins can now be separated (sorted) by power, similar to in ferro supplies.

Let’s take the earlier sections and have a look at how spin is distributed throughout the power panorama for ferro, antiferro and LC magnets.

What issues for storage is the “spin window”: the slice of power on the fringe of the band hole the place each accessible electron state has the identical spin. The bigger this window in contrast with the thermal jiggling at room temperature (about 26 meV), the higher the electrons keep sorted.

This means we’d like to have a semiconductor with a band hole, with out shedding the flexibility to separate spins by their power ranges, and with zero web magnetism.

This is the place our AI brokers (and me) enter the scene. Let’s dive into how these brokers discovered two promising supplies for subsequent technology pc reminiscence.

The brokers ran quantum-mechanical simulations of every crystal with the usual technique for this, density purposeful idea, at two ranges of approximation: a sooner one (PBE+U) and a slower, often extra correct one (HSE06). The band gaps and spin home windows under come from the extra correct one.

Candidate 1: Designed a Luttinger Compensated Magnet, YBaMnFeO₅

First, let’s see what our AI brokers designed:

• A brand new compound manufactured from solely 5 components (yttrium, barium, Mn, Fe, O). As far as we may discover, it has by no means been made, nor proposed as this type of magnet
• The compound was predicted to be a semiconductor
• This Luttinger-compensated magnet is predicted to have a 2.35 eV band hole, the place spin sorting happens at either side of the band hole: a window of 1.0 eV for holes and 1.4 eV for electrons. Note that thermal agitation at room temperature solely causes round 26 meV of fluctuation
• Finally, our agent predicted the fabric to retain magnetism as much as an elevated temperature: about 420 Okay within the uncooked simulation, or about 490 Okay after calibrating the simulation towards a recognized magnet

While this looks like a reasonably promising candidate, the design wants the Mn and Fe atoms to sit down in an ideal checkerboard. When our brokers simulated how the atoms prepare themselves at completely different temperatures, that checkerboard fell aside right into a random combine at round 950 Okay. Making this type of oxide takes about 900–1300 °C, and at decrease temperatures the atoms barely transfer, so normal synthesis would doubtless give a scrambled crystal.

A scrambled crystal loses the spin sorting, so this design could also be onerous to make in its helpful kind.

Candidate 2: Identified a Luttinger Compensated Magnet in KV[Cr(CN)₆] from 1999

Shortly after, my agent discovered this attention-grabbing materials, beforehand reported solely in 1999, that matches most of the necessities above. Its zero web magnetism isn’t new: the chemists who made it designed the 2 metals’ magnetism to cancel. Even its spin sorting was already on paper. A 2008 examine utilizing hybrid functionals like ours plotted its electron states spin by spin, and each band edges in that plot carry the identical spin. But that examine was about magnetic coupling beneath stress and by no means remarked on it. As far as we discovered, no one had identified that this makes KV[Cr(CN)₆] a Luttinger-compensated semiconductor, put numbers on its spin-sorted home windows, or examined how sturdy they’re. It had been hiding in plain sight.

KV[Cr(CN)₆] belongs to the identical household as Prussian blue, the 300-year-old pigment.

Our agent recognized it as a Luttinger-compensated materials:

• It predicts the fabric can have a band hole of about 2.1 eV, with each band edges sorted into the identical spin: home windows of two.6 eV for holes and 1.6 eV for electrons.
• The 1999 pattern stayed magnetically ordered as much as 376 Okay (103 °C), above room temperature, as measured by the chemists who made it (365 Okay after the pattern had been heated).
• And most significantly, its construction locks every metallic into its personal web site: chromium bonds to the carbon finish of every cyanide and vanadium to the nitrogen finish, which is precisely what YBaMnFeO₅ lacked.

These predictions are for an ideal, dry crystal. The solely pattern up to now, from 1999, is a powder with water in its pores, and it confirmed a small leftover magnetic second (0.125 Bohr magnetons per components unit, the place the right crystal would have zero). Our two simulation strategies disagree on how a lot the water weakens the impact: the extra correct one (HSE06) says the spin sorting survives, whereas the sooner one (PBE+U) says the opening window shrinks by greater than half. Neither the band hole nor the spin sorting has been measured but.

The backside line

We recognized a fabric that mixes properties of the 2 acquainted sorts of magnet and could possibly be helpful in spintronic gadgets. We imagine the 2 supplies mentioned on this article are promising examples of a category of supplies which have each antiferromagnetic and ferromagnetic properties.

The key options for each supplies: in our calculations for excellent crystals they’ve zero web spin second, but they’re semiconductors whose electrons are sorted by spin over home windows far bigger than the thermal jiggling at room temperature. With little stray magnetic discipline, they shouldn’t disturb their neighbours a lot.

Here is the rating of our two supplies:

• YBaMnFeO₅: designed candidate; could also be onerous to make.
• KV[Cr(CN)₆]: first made in 1999. Near-zero web magnetic second within the 1999 pattern (zero predicted for an ideal crystal), magnetic as much as 376 Okay.

This means that supplies for the subsequent step in spintronics might exist already, ready to be acknowledged. A 2025 examine that predicted two different Luttinger-compensated semiconductors discovered that each lose their magnetic order under room temperature, and named a room-temperature one because the open purpose. KV[Cr(CN)₆] has already been made, and its 1999 pattern stayed ordered as much as 376 Okay. Identifying room-temperature Luttinger-compensated semiconductors is a step towards sensible spin-based applied sciences, significantly if their magnetic properties could be managed by way of chemistry. The subsequent step is to make KV[Cr(CN)₆] once more and measure its spin sorting straight.

---

In the spirit of transparency, I invite the reader to undergo all of the computations that produced the above predictions, together with the calculations behind the candidate designs. The enter recordsdata, uncooked outputs and evaluation code behind the numbers on this submit, together with a one-command checker, unbiased re-runs and a listing of recognized caveats, have been shared on GitHub in a public repository I created to doc my analysis journey: github.com/spicylemonade/compensated-magnet-ledger.