The success of a promising class of next-generation batteries may hinge on something almost impossibly thin: a coating just a nanometer thick - roughly 100,000 times thinner than a human hair.
In new research, scientists at the U.S. Department of Energy's (DOE) Argonne National Laboratory combined computation and experiment to find candidate protective coatings for sulfide-based solid electrolytes and uncover what makes those coatings work. The study points to magnesium oxide as a particularly promising new coating and sets up a faster way to find others.
Solid-state batteries could store more energy and improve safety compared with today's lithium-ion batteries. But some of the most promising solid electrolytes, especially sulfide-based ones, are chemically fragile. They can react at key battery interfaces, especially where the electrolyte touches lithium metal. Those reactions can hurt performance and shorten battery life.
To tackle that problem, the team studied a type of sulfide solid electrolyte called lithium phosphorus sulfur chloride, or LPSCl. They used an approach based on a computational technique called density functional theory to screen a wide range of oxide coatings made by atomic layer deposition (ALD) - a method that deposits ultrathin, uniform layers with near-atomic precision. They predicted how those coatings would behave at three important battery interfaces: where the coating meets the electrolyte, the lithium metal and the cathode materials.
"This work focused on using computation to guide that search," said Justin Connell, an Argonne materials scientist and University of Chicago Consortium for Advanced Science and Engineering (CASE) scientist. "We can't experimentally explore the full range of possible materials in any reasonable way. That would take forever, and it's just not possible."
The team found that the best coatings were not always the least reactive. Instead, the most important factor was what compounds formed when the coating reacted at the interface. The best coatings formed reaction products that still let lithium ions move while limiting electron flow.
"It turned out that the reaction products really dominate the behavior," Connell said. "Zirconium oxide was one of the most stable materials by itself, but it was one of the worst-performing coatings we investigated."
The researchers then tested several candidate coatings by applying them to LPSCl powder with ALD. Magnesium oxide in particular stood out. It made the electrolyte more stable when in contact with lithium metal, reduced resistance at the interface and improved performance. It also helped block electron flow while still allowing lithium ions to move efficiently.
"Atomic layer deposition gives us a unique way to apply uniform coatings that are only about a nanometer thick, even on complex powder surfaces," said senior chemist and Argonne Distinguished Fellow Jeffrey Elam. "That level of control lets us test new coating chemistries efficiently and connect computational predictions to real materials."
The team also used scanning transmission electron microscopy and energy dispersive X-ray spectroscopy at the Center for Nanoscale Materials, a DOE Office of Science user facility at Argonne, to confirm the coatings were uniformly distributed on the powder surfaces.
By contrast, zirconium oxide formed less favorable reaction products and performed poorly. Zinc oxide, despite being predicted to be more reactive overall, still yielded beneficial transport behavior because of the reaction products it formed.
In addition to evaluating candidate protective coatings, this work also provides a better way to search a much larger materials design space.
"Our calculations helped identify which interfacial reactions are most likely to occur and which reaction products will support or hinder battery performance," said Argonne physicist Peter Zapol, who led the study's calculations and computational screening. "That gives us a more predictive way to evaluate coating candidates, rather than relying on trial and error."
Connell said the approach should help speed up the search for future coatings beyond the oxide systems studied here.
"This work shows us a better way to ask the design question, and now we can use the same approach to study many other chemistries," he said. "That could mean sulfides, fluorides, other binary chemistries, ternary coating chemistries or combinations of materials. It lets us explore that large chemical space much faster, without having to test everything experimentally first."
The results of this research were published in Advanced Science.