Home
The Entropy for Energy (S4E) Laboratory develops computational and AI methods to design chemically disordered materials for nuclear energy, hydrogen production, chemical manufacturing, and extreme environments. There are far too many possible compositions and atomic arrangements to calculate or test one at a time. We develop models that learn from this space and propose new materials, then make and test them in our laboratory and with collaborators.
Our work at Johns Hopkins University centers on four questions. How can AI learn the chemistry and atomic structure of disordered materials well enough to design new ones? Which materials will let fission reactors run safely for decades and immobilize their waste, and which will survive the conditions inside a fusion reactor? Which materials discoveries can quantum computers make possible that classical computation cannot? And how can hydrogen be produced from water without precious metals? Recent results include the first demonstration that iodine can be incorporated directly into a pyrochlore crystal lattice, platinum-free fuel-cell catalyst candidates, and new computational methods for high-entropy materials. The program is supported by ARPA-E, the Seaver Institute, ROSEI, and the Data Science and AI Institute. The group has filed two provisional patents.
The S4E lab is recruiting PhD students and postdoctoral researchers with backgrounds in materials science, physics, chemistry, and computer science. See the Jobs page for how to apply.
