Wagner receives DOE INCITE grant to generate fine-grained models of electrons’ motion

The goal: to simulate the motion of electrons with unprecedented detail

The ability to model materials’ properties is a crucial part of developing new materials with desired capabilities, but today’s modeling tools are sharply limited. A new grant from the U.S. Department of Energy’s INCITE program will support development of a new approach that includes more fine-grained information on the movement of electrons within a material, resulting in models that are significantly more accurate, quantitative, and predictive. The project will also generate valuable data on a diverse set of three classes of material systems that will be studied as examples.

Illinois Physics Professor Lucas Wagner, the principal investigator of the project, explains that existing modeling methods can handle only relatively coarse-grained information about materials under consideration.

The resulting predictions might be a bit on the rough side—or they might be flat-out wrong.

“What’s interesting about this is that the behavior you get at the larger length scale actually depends quite sensitively on what’s going on at the very fine length scales,” Wagner says. For example, “if you don’t correctly include very, very short-range details about how the electrons interact with each other, you would predict that silicon is transparent and an insulator, which is very much not the case!”

He explains that people use the approximate methods because a rigorous computation, based on first principles of quantum mechanics, would require solution of the Schrödinger equation—which “is very difficult to solve, as it turns out, for many particles.” The best available approaches are widely used, but their limitations have motivated a search for more powerful alternatives.

“You get into more complicated systems and you’d like to do better, right? You’d like to be able to model them better and understand them in a more detailed way,” Wagner notes.

Illinois Physics Professor David Ceperley, who is a Founder Professor, and Mechanical Science & Engineering Professor Elif Ertekin, who is an Andersen Faculty Scholar, will be co-principal investigators of the effort.

Wagner says that Ceperley was one of the developers of quantum Monte Carlo (QMC) techniques, whereby random numbers are used to solve quantum systems. The new project will use QMC to solve the Schrödinger equation for the considered materials. “What that allows us to do is to include the interaction between electrons natively in our calculations,” he explains.

Because QMC techniques are computation-heavy, the project will rely on Argonne National Laboratory’s Aurora supercomputer, which, as of this writing, is the world’s third most powerful supercomputer.

Wagner says that they need such a large computer “because we’re running these QMC simulations... that are like rolling dice, but we’re rolling dice probably 30 million times or so in the computer.”

As case studies, the team will examine three very different types of material systems.

The first is high-pressure hydrogen, which must be better understood, for example, to help planetary scientists study the interiors of Jupiter and other gaseous planets.

The second is atomic defects that emit and receive light. Such defects can determine many of a material’s observable properties, such as color. But they are also considered a candidate for quantum information storage, and might be useful for detection and sensing applications.

The third is ion conductors. Better understanding of ion transport would support future identification of materials that could enable better batteries.

The outcomes of the project are expected to include a suite of multiple software tools that others can use to perform similar analyses, along with the insights gained on the three example material systems.

“This project is part of the national Materials Genome Initiative, which aims to use simulations to design new materials for our future,” Wagner adds. “We are pushing forward our ability to do simulations that accurately reflect reality, which we hope will further our understanding of materials and help design new technologies.”