Stanford University scientists have turned to artificial intelligence (AI) and machine learning to help create a safer lithium-ion battery.

Lithium-ion batteries that power smartphones, laptops and other electronic devices have been known to explode and catch fire.

Researcher Austin Sendek, Stanford University Dept. of Applied Physics. Credit: Stanford

Researcher Austin Sendek, Stanford University Dept. of Applied Physics. Credit: Stanford

In August 2016, Samsung announced the recall of all its newly issued Galaxy Note 7 smartphones after several of the devices exploded on their users. The situation became so serious that in October, the U.S. Department of Transportation along with the Federal Aviation Administration issued an emergency order banning all Galaxy Note7 smartphone devices from air transportation in the United States.

The cause of the problem, researchers suggest, is that the liquid electrolytes that shuttle lithium ions back and forth between the battery’s positive and negative electrodes can catch fire if the battery overheats or is short-circuited.

Now, Stanford University researchers have identified nearly two-dozen solid electrolytes that could someday replace these volatile liquids based on techniques adapted from AI and machine learning.

“The main advantage of solid electrolytes is stability,” says lead researcher Austin Sendek, a doctoral candidate in applied physics. “Solids are far less likely to blow up or vaporize than organic solvents. They’re also much more rigid and would make the battery structurally stronger.”

Despite years of investigation, researchers haven’t found an inexpensive solid material that performs as well as liquid electrolytes at room temperature. But instead of continuing to test individual compounds, Stanford researchers used AI to build predictive models from existing experimental data.

“The number of known lithium-containing compounds is in the tens of thousands, the vast majority of which are untested,” Sendek says. “Some of them may be excellent conductors. We developed a computational model that learns from the limited data we already have, and then allows us to screen potential candidates from a massive database of materials about a million times faster than current screening methods.”

Still, the team spent more than two years gathering all known scientific data about solid compounds containing lithium. Their model used several criteria to screen promising materials, including stability, cost, abundance and their ability to conduct lithium ions and re-route electrons through the battery’s circuit.

“We screened more than 12,000 lithium-containing compounds and ended up with 21 promising solid electrolytes,” Sendek says. “It only took a few minutes to do the screening. The vast majority of my time was actually spent gathering and curating all the data.”