Posted in | News | Battery

Stanford Team Explores AI to Build Safer Lithium-Ion Batteries for Electronic Gadgets

Evan Reed, assistant professor of Materials Science & Engineering at Stanford, and graduate student Austin Sendek are using artificial intelligence to develop safer batteries. (Image credit: L.A. Cicero)

Scientists have spent decades searching for a safe substitute to the flammable liquid electrolytes used in lithium-ion batteries.

Recently, Stanford University researchers have identified about two-dozen solid electrolytes that have the potential of one day replacing the volatile liquids used in laptops, smartphones, and other electronic gadgets. The results, based on methods adapted from artificial intelligence (AI) and machine learning, have been published in the journal Energy & Environmental Science.

Electrolytes shuttle lithium ions back and forth between the battery’s positive and negative electrodes. Liquid electrolytes are cheap and conduct ions really well, but they can catch fire if the battery overheats or is short-circuited by puncturing.

Austin Sendek, Doctoral Candidate, Stanford University

The recent recall of almost two million Samsung Galaxy Note7 smartphones was due to battery fires, which is the most recent in a series of highly publicized lithium-ion battery failures.

“The main advantage of solid electrolytes is stability,” Sendek said. “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.”

Searching for Solids

Despite years of laboratory trial and error, researchers are still unable to find an economical solid material that performs as well as liquid electrolytes at room temperature.

Instead of haphazardly testing individual compounds, the team focused on AI and machine learning to construct predictive models based on experimental data. They trained a computer algorithm to learn how to detect good and bad compounds based on available data, similar to how a facial-recognition algorithm learns to detect faces after seeing many examples.

The number of known lithium-containing compounds is in the tens of thousands, the vast majority of which are untested. 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.

Austin Sendek, Doctoral Candidate, Stanford University

To construct the model, Sendek spent over two years collecting all known scientific data relating to solid compounds containing lithium.

“Austin collected all of humanity’s wisdom about these materials, and many of the measurements and experimental data going back decades,” said Evan Reed, an assistant professor of materials science and engineering and a senior author on the paper. “He used that knowledge to create a model that can predict whether a material will be a good electrolyte. This approach enables screening of the full spectrum of candidate materials to identify the most promising materials for further study.”

Screening Criteria

The model used several criteria to screen potential materials, including cost, stability, abundance, and their ability to conduct lithium ions and re-route electrons via the battery’s circuit.

Candidates were chosen from The Materials Project, a database that allows researchers to explore the chemical and physical properties of numerous materials.

We screened more than 12,000 lithium-containing compounds and ended up with 21 promising solid electrolytes. 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.

Austin Sendek, Doctoral Candidate, Stanford University

Eventually, the researchers plan to test the 21 materials in the laboratory to establish which are ideally suited for real-world conditions.

“Our approach has the potential to address many kinds of materials problems and increase the effectiveness of research investments in these areas,” Reed said. “As the amount of data in the world increases and as computers improve, our ability to innovate is going to increase exponentially. Whether it’s batteries, fuel cells or anything else, it’s a really exciting time to be in this field.”

Additional Stanford co-authors on the paper are doctoral candidate Qian Yang, postdoctoral scholar Ekin Dogus Cubuk, former doctoral student Karel-Alexander Duerloo and Yi Cui, a professor of materials science and engineering.

The research received funding from an Office of Technology Licensing Stanford Graduate Fellowship and a seed grant from the TomKat Center for Sustainable Energy at Stanford.

Building a better battery with machine learning

Reed Group Research/

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type