A cobalt-substituted layered crystal shows antiferromagnetic order near 250 K, and calculations point to an unusual arrangement in which all four cobalt atoms occupy a single iron layer. The preprint argues that this layer-selective ordering, together with cobalt-dominated magnetic interactions and correlated electronic states, may help explain the material’s unusually high ordering temperature.
The material, written as (Fe0.65Co0.35)4GeTe2, was compared with pristine Fe4GeTe2. In antiferromagnetism, neighboring magnetic moments are arranged in opposing directions. Here, the transition was identified through a peak in magnetic susceptibility and an anomaly in specific heat, with no splitting between measurements taken after cooling in zero field and in an applied field.
A layered structure with cobalt concentrated in one plane
The crystals were made by chemical vapor transport. X-ray diffraction found that the doped material retained the layered structure of Fe4GeTe2, while the calculated lattice spacing along the c axis was about 1.1% smaller. Density-functional theory was then used to compare possible atomic and magnetic arrangements.
Among the configurations tested, the lowest-energy arrangement put all four cobalt atoms in one iron layer. The result is a calculated preference, not a direct experimental map of every iron and cobalt site: the supplied analysis notes that the two elements have similar atomic numbers, making their occupancies difficult to distinguish precisely by experiment.
The calculations also describe antiferromagnetic order among cobalt atoms within a layer, with their moments slightly canted rather than perfectly opposed. Iron moments were assigned a different pattern: ferromagnetic within each layer but antiferromagnetic from one layer to the next. The preferred magnetization direction lay in the ab plane, which was calculated to be 0.89 meV per magnetic atom lower in energy than magnetization along the c axis.
Taken together, the energy calculations favor the antiferromagnetic configuration over the ferromagnetic one by about 0.9 meV per magnetic atom. That comparison describes modeled states rather than measuring a single microscopic interaction directly.
The electrons do not fit neatly into one category
The study’s proposed explanation also rests on the electronic structure. Specific-heat analysis produced a coefficient of 86.5 mJ/(mol·K2), while the electronic density of states at the Fermi level, a measure of how many electronic states are available near a material’s key energy boundary, was about 36 states per electron-volt per formula unit experimentally and 36.84 in the calculations. The authors associate a cobalt-related peak in that density with strong correlations and an itinerant-localized duality, meaning electrons that show both mobile and more confined behavior.
The magnetic story also extends above the main transition. Susceptibility measured in different directions crossed over near 320 K, and further magnetic and transport anomalies appeared around 360 K. The authors describe these features as evidence of persistent correlations above the 250 K transition, but the observations do not by themselves identify a single microscopic cause.
A fit to the inverse susceptibility gave an exponent of 0.66 in the power-law form used by the study. The authors describe that behavior as Griffiths-like, a term for a regime in which magnetic regions or fluctuations can behave unevenly rather than changing all at once at one sharp boundary.
Transport adds clues, with a cautious topological suggestion
The electrical resistivity increased by about a factor of 1.3 as the sample cooled from 300 K to 10 K and developed a broad hump near the antiferromagnetic transition. Compared with pristine Fe4GeTe2, the cobalt-substituted crystal also showed a smaller temperature dependence in its carrier density: the ratio between values at 10 K and 400 K was about 3.8, versus about 20.6 in the pristine material.
At low temperatures, the magnetoresistance became direction-dependent. With the field along the c direction, it stayed near 1.3% to 1.5% down to 10 K, while the in-plane value fell to about 0.8%. The authors interpret this anisotropy as a signature of low-temperature spin canting, although the supplied analysis treats that interpretation as an inference from the transport pattern.
The measurements also found a temperature-dependent anomalous Hall resistivity, and the calculations showed a pronounced peak in Berry curvature, a quantity that tracks how electronic bands bend through momentum space. The authors say the combination is consistent with a possible topologically nontrivial band structure and anomalous Hall conductivity. It is not a direct experimental demonstration of robust edge states or a topological phase.
Evidence still depends on calculations and replication
The central structural claim remains partly theoretical because the precise iron and cobalt occupancies were not experimentally resolved. The supplied analysis also says that the study did not report how many crystals, synthesis batches or independent specimens were examined, and it did not provide formal inferential statistics, confidence intervals or a broad sample-level uncertainty analysis.
Other open questions concern how the above-transition regime arises, whether the low-temperature canting can be confirmed by direct magnetic-structure measurements, and whether the proposed Berry-curvature effects can be verified through edge-transport experiments. Sensitivity analyses across different calculation settings were not reported, and independent replication was not supplied.
The document is a preprint on arXiv, identified as arXiv 2608.25209 in the supplied metadata. For now, the findings describe a proposed materials mechanism rather than a demonstrated device or application performance.
The acknowledgements report support from the Department of Defense, the Department of Energy and the National Science Foundation.
Paper data and sources
Original title: Origin of High-Temperature Antiferromagnetic Order in a van der Waals Material
Authors: Rabindra Basnet, Hari Paudyal, Gicela Saucedo Salas et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-25
DOI: Not available
Original paper · Full text