Caltech and Yale researchers calculate the Kondo effect from real materials' atomic structure

The Kondo effect is hard to compute because it is a many-body problem. It involves a huge number of interacting electrons in the metal, and their behaviour cannot be understood by treating them one at a time ^[9]. As the metal cools, electrons moving through it interact strongly with the impurity's unpaired spin. They flip their own spins in ways that partly cancel its magnetic moment ^[10]. As more of them join in, they surround the impurity and hide its magnetism, and the extra scattering produces the flattening and rise in resistance ^[10]. "That is the signature of the Kondo effect, and it's a property of the electrons in the impurity interacting with the electrons traveling through the bulk metal," Chan said ^[7].

Physicists had the general theoretical picture in the 1970s, and Kenneth Wilson, a Caltech alumnus, was among those who built it ^[8]. Caltech says the step that had not been possible was a precise calculation for a specific real material ^[1]. The new method takes the material's true atomic and electronic structure as its starting point ^[3].

Ordinary electronics explains why this matters. In silicon and other semiconductors, interactions among electrons are weak enough to ignore when describing how the material behaves ^[12]. Strongly correlated materials, which many proposed quantum technologies depend on, cannot be treated that way ^[12]. In high-temperature superconductors, what any one electron does depends strongly on what nearby electrons are doing, and conventional approximations struggle to describe them ^[11]. A single iron or manganese atom in copper is one of the simplest systems of this kind ^[5]. Caltech says the advance could help pave the way for realistic simulations of the more complicated ones ^[14].

The release does not say which impurity-metal pairs were computed, how closely the calculated Kondo temperatures matched measured ones, or how much computing time each calculation took. The last figure decides practical use. A method that has to be rerun for every candidate impurity and host metal is a screening tool only if each run is affordable, and a research result that reproduces known systems can still be too slow for that.

Chan, the paper's senior author and director of the Rudolph A. Marcus Center for Theoretical Chemistry at Caltech, set the scope himself ^[13]. "These first materials that we have studied are like a baby step, or a prototype problem, along the way to more complex phenomena such as high-temperature superconductors and quantum magnets," he said ^[16]. We think that description is accurate. The impurity calculations are the benchmark the method had to pass first, and by Chan's account the superconductor work comes later ^[16].