Skip to content Skip to navigation

Designing a Better Battery with Machine Learning

photo: Argonne National Laboratory cropped resized from Flickr

Evan Reed, Materials Science and Engineering

Solid-state lithium ion batteries hold promise as safer, longer-lasting alternatives to conventional batteries with the potential to drive significant improvements in the electrification of the transportation sector. The primary technological bottleneck to realizing a high-performance solid-state battery is identification of a suitable solid electrolyte material. Guess-and-check style efforts to find promising solid electrolytes have been made over the past several decades with limited success. The number of studied materials likely ranges only in the tens to low hundreds, but there are over 10,000 known materials that may be promising electrolytes. This project will leverage developments in machine learning techniques to efficiently and effectively learn from the data generated from past successes and failures. These algorithms allow the screening of tens of thousands of materials that have not been previously studied for battery applications—a process that would take several more decades with conventional methods. This approach has the potential to break the old paradigm of guess-and-check, accelerating progress towards production of a high-performance solid-state battery.

Read the spotlight article about this project>>

Publications and media:

"Holistic computational structure screening of more than 12,000 candidates for solid lithium-ion conductor materials" Energy & Environmental Science 10 (2017): 306-320.

"Bay Area researchers find ways to stop lithium ion batteries from exploding" ABC7 San Francisco News, February 2017.

"Stanford researchers work on solving lithium-ion battery explosiong" KCBS San Francisco News, December 2016.

"No more burning batteries? Stanford scientists turn to AI to create safer lithium-ion batteries" Stanford News, December 2016 (also posted on ScienceDirect).


Awarded 2016