FINDSPINGROUP puts magnetic arrangement and relativistic spin-orbit effects in a single symmetry framework, so researchers can tell which of the two permits a given exotic behaviour.

Some magnetic materials do things that look impossible. An antiferromagnet, in which neighbouring magnetic moments point in opposite directions and cancel out, is not supposed to deflect electrons sideways — that is the anomalous Hall effect, and it belongs to materials with a net magnetisation. Yet some antiferromagnets do it anyway. Other materials show electronic bands that split as though a magnetic field were present when none is.

Explaining these behaviours has been difficult less because the physics is obscure than because two different causes produce similar signatures. One is the geometry of the magnetic order itself: the spatial arrangement of the moments can break symmetries all on its own. The other is spin-orbit coupling, a relativistic effect that links an electron’s spin to its motion and that can break the same symmetries by a different route.

A team at the Southern University of Science and Technology has built a tool to tell them apart. Writing in National Science Review, the group led by Qihang Liu and Xiaobing Chen presents FINDSPINGROUP, an open-source platform that places both contributions inside one material-specific symmetry framework.

Asking a sharper question

Liu frames the problem as one of attribution rather than classification. “The important question is not only what symmetry a magnetic material has, but where each of its allowed responses comes from,” he said.

The framework the platform uses is called the oriented spin space group, or OSSG. Conventional magnetic space groups describe the symmetry of a magnetic structure as a whole. Spin space groups go further by treating the spin arrangement and the crystal lattice as capable of being transformed somewhat independently — which is the right description when spin-orbit coupling is weak, because the spins are then only loosely tied to the lattice directions.

The oriented version adds the orientation information that matters once spin-orbit coupling is switched on. That is what allows a direct comparison. “By placing magnetic geometry and spin-orbit effects in the same material-specific frame, the OSSG allows the two contributions to be compared directly,” Liu said.

In practice the platform takes a magnetic structure file — CIF, MCIF or SCIF — and automates the analysis, deriving the corresponding magnetic space group and reporting which responses symmetry permits with spin-orbit coupling and which it permits without.

Three materials, three different answers

The team demonstrates the approach on cases where the attribution question has real content.

V₂Se₂O, a two-dimensional altermagnet, shows spin splitting of its electronic bands. Altermagnets are a recently recognised class in which the moments cancel, as in an antiferromagnet, but the symmetry of their arrangement nonetheless permits spin-split bands. The platform’s verdict is that the splitting arises from magnetic geometry alone — no spin-orbit coupling required.

MnSe₂ is analysed for ferroelectric behaviour: the magnetic order is found to allow electric polarisation, and the platform identifies candidate switching endpoints, the configurations between which the polarisation could in principle be flipped. A magnetic material whose polarisation can be switched is the object of a long search in multiferroics research.

CoNb₃S₆, a non-coplanar antiferromagnet whose moments do not all lie in a single plane, is shown to permit an anomalous Hall response without spin-orbit coupling. That is the attribution question in its starkest form: a Hall effect in a compensated magnet, arising from the geometry of the spin texture rather than from relativistic coupling.

What it tells you, and what it does not

The limitation is built into the method and the authors are clear about it: the outputs are symmetry constraints, not numerical predictions. The platform will tell a researcher that an anomalous Hall current is allowed in a given material and that spin-orbit coupling is not needed to allow it. It will not say how large that current will be, or whether it will be large enough to measure.

This is the standard division of labour in condensed-matter physics, and it is not a weakness so much as a scope. Symmetry analysis rules things out definitively — a response that symmetry forbids will be exactly zero — and rules things in permissively. Everything quantitative requires electronic-structure calculation or experiment. What a tool like this changes is the cost of the first step: deciding which calculations are worth running, and which candidate materials are worth synthesising.

A file format, and the slow business of standards

Alongside the platform the team introduces SCIF, a spin crystallographic information file intended as a portable container for spin-group information, which they are developing with the International Union of Crystallography’s Commission on Magnetic Structures.

That detail is easy to skip past and probably matters as much as the platform. Crystallography’s existing CIF standard is why structural data moves between software packages without translation; magnetic structures have had a patchier history. A format developed with the IUCr has a chance of becoming the common currency, which would make symmetry analysis of magnetic materials routine rather than bespoke.

The platform is available at findspingroup.com with source code on GitHub. The release provides no funding information and no conflict-of-interest statement, and it does not give the paper’s title, naming only the platform; the endnote below therefore cites the DOI without one.

The field this serves is moving quickly. Altermagnetism went from a proposed classification to an experimentally confirmed phenomenon within a few years, and the recognition that magnetic geometry alone can produce effects previously attributed to spin-orbit coupling has sent researchers back through a long list of known materials. A tool that answers the attribution question in a few minutes rather than a few months is useful precisely because that list is long.

---

Notes

1. “FINDSPINGROUP.” National Science Review, 2026. DOI: 10.1093/nsr/nwag614. The release does not give the paper’s title.
2. Science China Press. “How can scientists disentangle the origins of unconventional magnetic properties?” EurekAlert!, 6 October 2026. https://www.eurekalert.org/news-releases/1146642.