How a few extra rhodium atoms reshape a magnetic transition
09 Sep 2026Polarised neutron imaging at the ILL reveals how adding just 1–2% rhodium changes the way a magnetoelastic phase transition unfolds through bulk iron-rhodium.
-
Environnement
Iron-rhodium (Fe–Rh) is a promising material for future solid-state cooling. Instead of using refrigerant gases, this solid material can absorb or release heat when exposed to magnetic fields, pressure or mechanical stress.
This behaviour is linked to a first-order phase transition. As Fe–Rh is heated, its magnetic order changes while its crystal lattice expands slightly. But the transformation does not happen everywhere at once: the new magnetic phase may appear in some regions first and then spread through the material.
Using polarised neutron imaging on MoTo at the ILL, researchers have now watched this process unfold through bulk samples of three almost identical Fe–Rh alloys. The difference was striking. In Fe50Rh50, containing equal parts of iron and rhodium, different regions transformed at different temperatures, producing a broad and uneven transition. Adding just 1–2% more rhodium made the transformation much sharper and more uniform in the bulk of the samples, while also shifting it to lower temperatures.
The results show how a very small change in composition can reshape the way an entire material transforms and how polarised neutron imaging can reveal behaviour hidden beneath the surface and behind an average measurement.
Iron-rhodium: one transition, several routes to cooling
Cooling does not necessarily have to rely on compressing and expanding refrigerant gases. Researchers are also exploring caloric materials: solids that can reversibly heat up or cool down when exposed to an external stimulus, offering possible routes towards future more environment friendly solid-state refrigeration.
Iron-rhodium (Fe–Rh) is one of the standout materials in this field. It exhibits world-leading multicaloric responses, meaning that its thermal behaviour can be influenced in more than one way: by applying a magnetic field, pressure or mechanical stress. These responses are known respectively as the magnetocaloric, barocaloric and elastocaloric effects.
What makes this possible is a phase transition in which the magnetic order and the crystal lattice change together. At lower temperatures, the iron magnetic moments are arranged in opposite directions. This antiferromagnetic order means that their magnetic contributions largely cancel and the material has almost no net magnetisation. On heating, Fe–Rh switches to a ferromagnetic state, in which there is an induced magnetic moment in the Rh sites and the magnetic moments of both Fe and Rh atoms become aligned in the same direction so that a net magnetisation develops.
The transition also affects the structure of the material: the volume of the crystallographic unit cell expands byabout 1%, while retaining the same symmetry. This coupling between magnetic order and the lattice makes the phase transition a magnetoelastic first-order phase transition and helps explain why the same transition can be influenced through magnetic fields as well as pressure or mechanical stress.
But knowing that the transition takes place is only part of the picture. A first-order phase transition does not necessarily sweep through a material all at once. The new phase first appears in certain regions (a process called nucleation) and then grows, meaning that antiferromagnetic and ferromagnetic regions can coexist during the transition.
For a functional material, this matters. Where the new phase first appears, how it spreads and over what temperature range the transformation takes place can all influence the overall performance. And although Fe–Rh is known to be highly sensitive to composition and defects, seeing how those factors affect the transition through the bulk of the material has remained difficult.
Polarised neutron imaging reveals how the magnetic transition unfolds through the bulk
To investigate this, the researchers compared three alloys prepared under the same conditions: Fe50Rh50, Fe49Rh51 and Fe48Rh52. The difference between them amounts to changing the rhodium content from 50 to 51 or 52 atoms in every hundred.
Neutron diffraction measurements on the ILL’s D1B diffractometer first confirmed the expected ordered crystal structure and antiferromagnetic state at room temperature. The researchers then used polarised neutron imaging on MoTo at the ILL while slowly heating and cooling the three samples through the transition.
Polarised neutron imaging is particularly well suited to this problem because neutron spin is sensitive to magnetism. Before reaching the sample, the neutron beam is polarized, meaning that the neutron spins are oriented in the same direction. This polarization is largely preserved while the neutrons pass through antiferromagnetic Fe–Rh. In the ferromagnetic phase, however, internal magnetic fields of each domain rotate the neutron spins by different amounts, progressively depolarizing the beam.
By calculating this depolarization across the sample as the temperature changed, the researchers could follow the development of the ferromagnetic phase.
The strength of the method is that it can follow the transition through a bulk piece of the material, rather than only at its surface. As the samples were heated and cooled, the neutron images revealed how different regions evolved following different pathways from the antiferromagnetic to the ferromagnetic state and from the ferromagnetic to antiferromagnetic state, respectively.
MoTo (Monochromatic Tomography) is an instrument that provides monochromatic and advanced imaging modes.
A small change in composition, a very different transition
The neutron images revealed two very different transformation pathways.
In Fe50Rh50, the ferromagnetic phase first appeared in selected regions rather than throughout the sample at once. As the temperature increased, these regions expanded while other parts remained antiferromagnetic. Different areas therefore transformed at different temperatures, producing an extended coexistence of the two phases.
The two Rh-richer alloys behaved very differently. In Fe49Rh51 and Fe48Rh52, the transition was much more uniform across the samples and occurred over a considerably narrower temperature range. Increasing the rhodium content also shifted the transition towards lower temperatures.
The effect of composition therefore goes beyond changing the transition temperature. It changes how the new phase nucleates and grows through the material.
The authors suggest that the additional rhodium may provide more favourable locations for nucleation, allowing the new phase to form more readily throughout the sample. But the microscopic origin is not yet settled.
More broadly, the study shows the value of polarised neutron imaging for following magnetic phase transitions in bulk materials. In Fe–Rh, it exposed strikingly different transformation pathways in alloys with only slightly different compositions, showing just how much information can remain hidden beyond the surface and behind a material’s average response.
Watching a magnetic transition unfold. Polarised neutron imaging reveals how the magnetic state changes inside three bulk Fe–Rh alloys as they are heated from the antiferromagnetic to the ferromagnetic state and cooled back to the antiferromagnetic state. The Fe₅₀Rh₅₀ sample (left) transforms gradually and unevenly over a broad temperature range, whereas the Rh-richer Fe₄₉Rh₅₁ (top) and Fe₄₈Rh₅₂ (bottom right) alloys transform much more uniformly and over a narrower temperature range during both heating and cooling. Credit: Commun Mater (2026). DOI: 10.1038/s43246-026-01251-5
Reference
Padrón-Alemán, K., Shen, J., Tengattini, A. et al. Direct imaging of composition-driven magnetoelastic phase transformations in bulk Fe–Rh. Commun Mater (2026). https://doi.org/10.1038/s43246-026-01251-5
ILL Contact Persons: Kenny Padron Aleman, Alessandro Tengattini
Institutions involved in the research: Universidad de Oviedo, University Grenoble Alpes,Instituto Potosino de Investigación Científica y Tecnológica