"Electrolytes shuttle lithium ions back and forth between the battery's positive and negative electrodes," said doctoral candidate Austin Sendek. "Liquid electrolytes are cheap and conduct ions well, but they can catch fire if the battery overheats or is short-circuited by puncturing.
"The main advantage of solid electrolytes is stability. Solids are far less likely to blow up or vaporise than organic solvents. They're also much more rigid and would make the battery structurally stronger."
Instead of testing individual compounds randomly, the team turned to AI and machine learning to build predictive models from experimental data. They trained a computer algorithm to learn how to identify good and bad compounds based on existing data. According to the researchers, it functions much like a facial recognition algorithm which learns to identify faces after seeing several examples.
"The number of known lithium-containing compounds is in the tens of thousands, the vast majority of which are untested," Sendek explained. "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."
To build the model, Sendek spent more than two years gathering all known scientific data about solid compounds containing lithium. The 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 concluded. "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, and developing metrics to define the confidence of model predictions."
