Researchers at Rice University, together with the University of Minnesota and Switzerland's Paul Scherrer Institute, working with contributing researchers from the Gwangju Institute of Science and Technology, Kyung Hee University, and Myongji University in South Korea, the University of Illinois Urbana-Champaign, the University of West Bohemia in the Czech Republic, and beamline scientists at Lawrence Berkeley National Laboratory's Advanced Light Source and Brookhaven National Laboratory's National Synchrotron Light Source II, have found that ultrathin, epitaxially strained films of ruthenium dioxide (RuO2) display a spin texture consistent with altermagnetism, a form of magnetic order the material's bulk and thick-film forms have consistently failed to show despite RuO2 standing for years as one of altermagnetism's leading candidate materials.
Illustration of strain-induced emergent magnetism in ultrathin RuO2/TiO2. Possible theoretic altermagnetic spin density is rendered with orange and blue visual effects in the top RuO2 layers using AI. Credit: Rice University/Yichen Zhang.
Using spin-resolved angle-resolved photoemission spectroscopy (ARPES) on a 2.7-nanometer-thick, substrate-strained RuO2 film, the team observed a spin texture combining mirror-odd and mirror-even components that a comprehensive symmetry analysis rules out explaining through nonmagnetic effects, pointing instead to intrinsically broken time-reversal symmetry.
Altermagnetism is a recently proposed third class of collinear magnetic order, distinct from ferromagnetism and antiferromagnetism, in which magnetic sublattices with opposite spins are related by rotational rather than translational symmetry. That produces momentum-dependent, nonrelativistic spin splitting even though the material carries no net magnetization, a combination researchers say could enable smaller, more efficient RAM architectures and other next-generation spintronic devices. RuO2's metallic conductivity, room-temperature magnetic ordering predicted by ab initio calculations, and large predicted spin-splitting energy scale made it one of the first and most closely watched altermagnetic candidates, and early experiments including field-induced nonlinear Hall effect, planar Hall effect, and spin-splitter torque measurements were read as supporting evidence. But a growing body of work using x-ray diffraction, polarized and unpolarized neutron diffraction, muon spin rotation spectroscopy, infrared spectroscopy, quantum oscillations, and Mössbauer spectroscopy has since converged on the opposite conclusion for bulk and strain-relaxed thick-film RuO2: no detectable magnetism at all, leaving the material's magnetic status one of the more contested open questions in the field.
The team grew their RuO2 films by hybrid molecular beam epitaxy on Nb-doped TiO2 (110) substrates, producing a 2.7 nm RuO2 layer on a 2 nm TiO2 buffer layer that x-ray reflectivity, x-ray diffraction, reflection high-energy electron diffraction, and atomic force microscopy confirmed was atomically smooth, well-crystallized, and fully strained by the substrate along both in-plane directions. Rotational-anisotropy second-harmonic generation measurements confirmed the resulting structure adopts the polar mm2 point group. The researchers then carried out spin-resolved ARPES at the Advanced Light Source, probing the film's electronic structure and photoelectron spin polarization at 15 K under an applied in-plane or out-of-plane magnetic field, and compared the results against fully relativistic one-step-model photoemission calculations and both bulk and slab density functional theory calculations to distinguish intrinsic magnetic effects from measurement-geometry artifacts.
The measurements revealed narrow surface-derived bands near the Fermi level that are unique to the ultrathin, fully strained films and not seen in thicker or strain-relaxed RuO2. More strikingly, the spin-resolved data showed photoelectron spin polarization components that violate the selection rules expected for a nonmagnetic, time-reversal-symmetric mm2 crystal, including a finite spin polarization along the [001] direction at normal emission that should be strictly zero absent intrinsic magnetism. A group-theory analysis of the observed spin-splitting terms points to the magnetic point group m′m2′, with in-plane magnetic moments, consistent with either weak ferromagnetism or in-plane d-wave altermagnetism - a different magnetic point group than the out-of-plane m′m′2 symmetry inferred from separate second-harmonic-generation measurements taken near room temperature on similar films, suggesting the material's magnetic symmetry may itself be temperature- or strain-magnitude-dependent.
"Ruthenium dioxide was one of the first materials to be proposed as an altermagnetic candidate, but studies on its bulk form didn't return evidence of magnetism," said Ming Yi, associate professor of physics and astronomy at Rice University and a corresponding author on the study. "Our research shows that its ultrathin form, on the other hand, may be the key in making it magnetic." First author Yichen Zhang, a recent Rice graduate, said the team's spin-resolved measurements, combined with theoretical calculations, "found that, in our experimental conditions, the ruthenium dioxide shows spin textures consistent with unconventional magnetism," suggesting bulk and ultrathin RuO2 "may have distinctly different magnetic properties" under the right conditions. Those conditions center on lattice strain: without it, as in RuO2's natural bulk form, the electron spins showed no sign of altermagnetism. "The strain-dependent nature suggests that we may be able to use lattice strain as a tuning knob to induce or control altermagnetism," Zhang said, adding that this could prove useful for next-generation spintronics and RAM architectures.
The researchers frame the result as evidence that RuO2's magnetic state depends sensitively on dimensionality and strain rather than being a fixed material property, adding a new, thickness-dependent data point to a long-running debate rather than settling it outright. "This work shows just how complex these questions can be," Yi said. "The high quality material prep and the careful measurement protocol were critical to our observation of the correct electron spin properties. Through this, we were able to determine not only the magnetic state symmetries but a potential way to manipulate it in next-generation quantum materials."