Opus 5.5 agents discover two room-temperature magnetic semiconductor candidates

This is a very bizarre introduction. People encounter diamagnets (e.g., copper) and paramagnets (e.g., aluminum) way more than they encounter antiferromagnets. I don't know why you'd ever cast magnetism as a false binary between ferromagnets and antiferromagnets, without acknowledging any other types of magnetic order.

(I did a PhD in magnetic materials)

Edit: I'll add that whether an antiferromagnet is useful, say, for exchange biasing a ferromagnetic thin film, depends on many factors. Just looking at antiferromagnetism alone you've got collinear vs non-collinear, G-type vs A-type vs C-type, commensurate vs incommensurate, and isotropic vs anisotropic; and all of that interacts with the interface structure, yada yada yada. It would be helpful if the authors elaborated on the expected properties of these materials. I personally don't know what people want room-temperature magnetic semiconductors for, but I'd be curious to learn what set of properties they think would be useful.

Amazing times.

Actual title:

Two Room-Temperature Antiferromagnetic Semiconductor Candidates

There's nothing unusual about finding room temperature semiconductors. I assume whoever posted it misread this as room temperature superconductors, but it has nothing to do with that.

What's interesting here is the antiferromagnetic part of the title, which was removed. I think this makes it relevant for e.g. RAM, but not superconducting. Someone can correct me if I'm wrong.

Thus, you can build ai/ml models+agents to explore those spaces, at a speed and scope much larger than what any human can do

I can imagine findings like these are going to keep increasing in frequency to a point in which the bar for novelty goes a lot higher

    The agents ran quantum-mechanical simulations of each crystal with the standard method for this, density functional theory, at two levels of approximation: a faster one (PBE+U) and a slower, usually more accurate one (HSE06). The band gaps and spin windows below come from the more accurate one.

Except they didn't discover new superconductors. Their AI came up with a novel idea to discover new superconductors, and they didn't even bother to check whether the results it hallucinated checked out.

The distinction between "having an idea" and publishing a paper where you demonstrate that the idea has merit, where you validate it, is one you shouldn't have to really explain to a high school student, much less a grad student publishing their first paper.

I don't see any claims that this is better than the current silicon and gallium arsenide semiconductors that we use. And the use of "room temperature" seems a deliberate attempt to misconstrue this with superconductors

If ai becomes so prolific that we humans all stop doing those things then will they still work?

Are they a promoter / influencer for Anthropic?

Many more things will be like this. The massive amounts of 'genius' buried under corporate management and obscurity in the past 500 years will be a treasure trove.

Not sure why we're calling it a discovery, when they've literally been made before, by a human.

I can think of worse uses of VC AI funding.

So, while this is great and I think LLMs will accelerate materials science the main bottleneck isn't having enough promising candidate materials, I think we have a backlog of at least a few hundred candidates that are worth pursuing. Maybe some money that goes to data centers would better go into CVD machines.

It didn't discover anything. This is how cooked people are.

PS: edit - I was thinking about room-temperature superconductors. My mistake.

What percentage of these comments did any form of validation of the findings before diminishing the author for AI use and discarding the findings as though they had invalidated the results?

Your normal heuristics like word choice cannot help you validate or invalidate a superconductor.

"In the spirit of transparency I invite you to ask your own LLM to verify what my LLM did"

Traditional superconductivity is limited by the masses of atoms vibrating in the lattice, which sets an upper temperature threshold because thermal noise acts like a phase change that returns the superconductor to being an ordinary conductor or insulator above its critical temperature.

Electrons are fermions with half-integer spin, meaning that their total quantum wave function must be antisymmetric. Swapping any two electrons flips the sign of their wave function, meaning that they can't overlap with the same spin, by the Pauli exclusion principle. But two electrons with opposite spin form a spin-singlet state where their spatial wave functions can overlap. Which is why two electrons with opposite spin can be in the same atomic orbital forever.

For larger quantum wave functions in molecules, electrons occupy larger molecular orbitals forming covalent bonds, and in conductors form what's known as a Fermi sea. If many electrons spontaneously align their spins via the exchange force which arises from exclusion and electrostatic repulsion (the same effect which causes Fermi holes, conceptually analogous to electron vacancies in semiconductors), their magnetic moments become large enough to be observed at the macro-scale as ferromagnetism.

At low temperatures, electrons have restricted momentum states at the sharp edge of the Fermi sea, so by the Heisenberg uncertainty principle, their spatial wave function uncertainty (coherence length) expands, causing them to overlap. Below the critical superconducting temperature, they pair up to become composite bosons. Bosons have integer spin and symmetric wave functions, so they aren't bound by the exclusion principle and undergo Bose-Einstein condensation.

In Bardeen-Cooper-Schrieffer (BCS) theory, these composite bosons are known as Cooper pairs. They share a wide coherence length and act as a unified macroscopic quantum state that flows around atomic nuclei without scattering, even in the presence of atomic vibrations (phonons) or imperfections in the lattice. In effect, this makes the superconductor a macro-scale orbital, similarly to how a permanent magnet is macro-scale spin.

BCS theory explains superconductivity from the bottom up like an engineering approximation. More foundational frameworks like Ginzburg-Landau (GL) theory use spontaneous symmetry breaking to explain it from the top down.

Using GL theory and going back to first principles, instead of avoiding phonons/vibrations to raise the critical temperature, we can manipulate lightweight electron clouds. To do this, we use electron-hole pairs (excitons) in semiconductors to dynamically polarize the electron clouds. This electronic framework leads to orbital fluctuation theory and excitonic superconductivity.

It was recently discovered that graphene layers twisted at a 1.1-degree magic angle create superconductivity. The overlapping atomic lattices form a Moiré superlattice that slows electrons down in places to restrict their momentum and expand their spatial uncertainty. This mimics the effects of low temperatures spatially instead of thermally, and is known as twistronic superconductivity.

Note that I wrote this summary by hand after doing a deep dive with AI, which also helped me edit it. While it's notable that AI can explore large problem spaces and find solutions independently, I'm at least as fascinated that it can now help us tackle complex topics in an approachable manner, which should lead to new insights and hopefully inspire future researchers.

> A designed web version with the same text and diagrams is in docs/index.html; turn on GitHub Pages for the /docs folder to serve it.

Yeah, the author did not even reads the slop Claude produces.

> Candidate 1: Designed a Luttinger Compensated Magnet, YBaMnFeO₅

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

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

An actual validation would be complex; so let's try and see what parts of this argument are grounded and feasible?

/? Prussian blue antiferromagnetic: https://scholar.google.com/scholar?q=Prussian+blue+antiferro... :

- a number of articles confirming antiferromagnetic effects

/? Luttinger-compensated Prussian blue:

- "Luttinger-compensated bipolarized magnetic semiconductor" (2025) https://journals.aps.org/prb/abstract/10.1103/9syc-71w8 .. "[2502.18136] Luttinger compensated bipolarized magnetic semiconductor" https://arxiv.org/abs/2502.18136 :

> The Luttinger compensated magnetism not only has the zero total magnetic moment as the antiferromagnetism, but also has the -wave spin splitting as the ferromagnetism, thus our work not only provides theoretical guidance for searching Luttinger compensated magnetic materials with distinctive properties, but also provides a material basis for the application in spintronic devices.

/? Luttinger-compensated : https://www.google.com/search?q=Luttinger-compensated

We could model this as a logical proof that's checkable also in lieu of doing actual work to confirm or reject the (AI) hypothesis, but first let's reason about the feasibility:

Are the described effects real?

Are there reported, reputable similar findings in similar materials?

So, at least the blue one could really work. Like it's 1999.

---

Without even validating the argument, what else do we think we know about this problem?

Did the authors know that there is a laser way to laser program the normally random domains of a magnet or antiferromagnet?

But is this going to be lower-cost and more sustainable than carbon-based room-temperature semiconductor computing, and does anyone know whether that will work yet (with ABC stacking in trilayer and pentalayer rhombohedral graphene) either?

I guess we can follow up later by searching for citations that reference this article that does not have DOI (which are free from Zenodo and FigShare).

Have we sufficiently reasoned or inferred whether the study is repeatable and reproducible?

At least we didn't inappropriately reject the hypothesis without experimentation or evidence