Tackling Electronic Correlations
Materials with strongly correlated electrons exhibit many phenomena with great potential for applications, from high-temperature superconductivity to colossal magnetoresistance. But describing a correlated material theoretically is extremely hard to do because each of its electrons has a strong influence on the sea of electrons that surround it. Now, Fredrik Nilsson at Lund University, in Sweden, Lewin Boehnke at Fribourg University, in Switzerland, and co-workers have developed a new simulation approach that can tackle a broad range of correlated materials. The result breaks ground by describing both local and long-range correlations from first principles. As such, it may be more general and have greater predictive power than previous approaches based on adjustable parameters that have to be fit to experiments.
Existing theoretical approaches provide good approximations only for two specific situations. So-called GW methods account for long-range correlations (between electrons at different atomic sites of a lattice) but break down for systems in which local correlations (between electrons at the same site) are strong. The dynamical-mean-field method, on the other hand, can capture strong local correlations but ignores long-range ones. Following a theoretical framework proposed in 2003, Nilsson and colleagues combined the GW and dynamical-mean-field approaches into one that encompasses both short- and long-range correlations.
The researchers tested their new approach on several materials. Studying stretched sodium—in which the strength of correlations can be tuned by expanding the lattice—they proved that their method can describe materials with both moderately and strongly correlated electrons. They also investigated strontium vanadate, a material considered to be representative of many metals with short-range correlations. The team, however, surprisingly found that the electronic properties of the material could be accurately described only by including long-range correlations in their simulations.
This research is published in Physical Review Materials.
–Matteo Rini
Matteo Rini is the Deputy Editor of Physics.