Researchers at the California Institute of Technology and Yale University have, for the first time, calculated the Kondo effect in an actual material using fundamental physics rather than simplified approximations, according to ScienceAlert. The findings were published in Science.
The Kondo effect describes how magnetic atoms embedded in a metal disrupt the flow of electricity. Ordinarily, cooling a metal lowers its electrical resistance, because electrons move more freely at lower energy. Introduce a magnetic impurity, and below a threshold known as the Kondo temperature, resistance begins climbing again instead. ScienceAlert reports that the team's approach reproduced laboratory measurements of this behavior with high fidelity, and in some cases outperformed existing simplified models by a factor of 100.
"It is now possible to predict the properties of some complicated materials purely through computation without referring to experiments," physicist Garnet Chan of Caltech told ScienceAlert.
Why it matters
The Kondo effect is a foundational phenomenon in quantum physics, and it has been studied for decades. What has been missing is a way to predict its precise behavior in specific metal–impurity combinations without first running an experiment. ScienceAlert notes that until now, researchers relied on lab work and stripped-down models, which never delivered predictions at the new method's level of precision.
That gap matters because the materials with the greatest technological promise — including high-temperature superconductors and quantum magnets — are also the ones whose electron behavior is hardest to model. A computational tool that can describe those interactions reliably could let scientists evaluate candidate materials before committing to costly fabrication and testing.
A molecule-style treatment of a single atom
The obstacle is what physicists call a many-body problem: temperature, electron spin and magnetism all interact on extraordinarily small scales once the Kondo effect sets in. ScienceAlert reports that the complexity exceeds what even the most powerful supercomputers can handle directly.
To get around this, the Caltech and Yale team took an ab initio, or first-principles, route. They borrowed techniques from chemistry to model a single magnetic atom in full detail, treating it as though it were a larger molecule rather than simplifying it. Applying mathematical shortcuts and substantial computing time, they then ran the basic laws of quantum physics against that model repeatedly, narrowing their predictions with each pass.
The underlying mechanism itself was already understood. Electrons gather around a magnetic impurity, neutralizing the atom's magnetism while simultaneously obstructing current. As Chan put it to ScienceAlert, that signature arises from electrons in the impurity interacting with electrons moving through the surrounding metal.
What the new work adds is quantitative command over that interaction. By checking their computed results against lab measurements of the Kondo effect, the researchers demonstrated that their predictions hold up. ScienceAlert describes the materials examined as the most basic metal–magnetic atom pairings available.
Baby steps toward harder problems
Chan characterized those first materials as an early, prototype stage on the road to more complex phenomena, naming high-temperature superconductors and quantum magnets as eventual targets.
Chemist Linqing Peng of Caltech framed the result as part of a broader shift. "We are in an exciting era in which faithful predictive quantum descriptions of the full chemical complexity of real materials are coming within reach," Peng told ScienceAlert, adding that predicting material-specific behavior of correlated electrons from first principles is becoming realistic even for some of the most challenging classes of quantum materials.
Better prediction methods carry practical weight. Superconductors are widely pursued for power grids and large-scale computing, and the ability to model a material's behavior in advance — without waiting on lab data — would change how such materials are screened and developed.
The approach has now been shown to work on simple cases. The next question is how far it can scale toward the complex materials where it would matter most.
