Stanford Team Effectively Integrate AI and Atomic-Scale Images to Facilitate Better Batteries

Although present-day rechargeable batteries are amazing, they are far from perfect. Sooner or later, they tend to wear out, leading to expensive substitutions and recycling.

Artist’s rendition of a particle analyzed by a combination of machine learning, X-Ray, and electron microscopy. (Image credit: Ella Maru Studio).

“But what if batteries were indestructible?” probes William Chueh, an associate professor of materials science and engineering at Stanford University and senior author of a new study covering a first-of-its-kind analytical methodology to constructing better batteries that could help speed things up. The study has been published in the journal Nature Materials.

Chueh, lead author Haitao “Dean” Deng, Ph.D. ’21, and collaborators at Lawrence Berkeley National Laboratory, MIT, and other research organizations used artificial intelligence (AI) to examine new types of atomic-scale microscopic images to comprehend precisely why batteries wear out. Ultimately, they say, the findings could pave the way to batteries that last a lot longer than present-day models.

In particular, they sought a specific type of lithium-ion batteries based on so-called LFP materials, which could result in mass-market electric vehicles as they do not utilize chemicals with constrained supply chains.

Nanofractures

Think of a battery as a ceramic coffee cup that expands and contracts when it heats up and cools off. Those changes eventually lead to flaws in the ceramic. The materials in a rechargeable battery do the same each time you recharge it and then use up that electricity, leading to failure.

William Chueh, Senior Study Author and Associate Professor of Materials Science and Engineering, Stanford University

In the battery, Chueh observed, the temperature is not what causes the fissures, but the mechanical strain one material has against the other with each charge cycle.

Unfortunately, we don’t know much about what’s happening at the nanoscale where atoms bond. These new high-resolution microscopy techniques allow us to see it and AI helps us understand what is happening. For the first time, we can visualize and measure these forces at the single nanometer scale.

William Chueh, Senior Study Author and Associate Professor of Materials Science and Engineering, Stanford University

Chueh said that the performance of any particular material is the work of both its physical interaction and chemistry in the material at the atomistic scale, what he calls “chemo-mechanics.” Furthermore, miniaturization and the more diverse the atoms in the material are, the tougher it is to estimate how the material will act. At this juncture, AI is the solution.

A Transformative Tool

Employing AI for image analysis has been done before, but this is the first time it is used to explore atomic interactions at the smallest of scales. In medicine, AI has become a revolutionary tool in examining images of all issues from faulty knees to terminal cancers.

Meanwhile, in materials science, new approaches of high-resolution electron, X-Ray and neutron microscopy are facilitating direct visualization at the nanoscale.

For their study, the researchers selected lithium iron phosphate or “LFP,” a common material used in positive electrodes, which is gaining acceptance with electric car makers and other battery-intensive companies. This electrode does not contain nickel and cobalt, which are used in a number of batteries available in the market. LFP batteries are also safer; however, they store less electricity per pound.

Though LFP has been explored for about 20 years, two important outstanding technical questions could only be predicted thus far. The first includes comprehending a material’s elasticity and deformation as it charges and discharges. The second relates to how it expands and contracts in a particular regime where the LFP is partly stable, or “metastable.”

Deng described both for the first time using his image-learning methods, which he applied to a series of two-dimensional images created by a scanning transmission electron microscope, and to advanced (spectro-ptychography) X-Ray images.

The findings, he said, are significant to a battery’s capacity, rate and energy retention. Better yet, he believes it is generalizable to the majority of crystalline materials that might also be used to create good electrodes.

AI can help us understand these physical relationships that are key to predicting how a new battery will perform, how dependable it will be in real-world use and how the material degrades over time.

Haitao “Dean” Deng, PhD ’21, Study Lead Author and Department of Materials Science and Engineering, Stanford University

New Directions

Chueh refers to Deng as an “academic entrepreneur.” His background is chemistry but he taught himself the intricacies of AI to handle this challenge. Deng said the method is a form of “inverse learning” wherein the outcome is known – high-resolution still images of degraded LFP – and AI helps rebuild the physics to illuminate how it got that way. That new knowledge, in turn, becomes the foundation for enhancing the materials.

Deng observed that earlier non-AI studies have described correlations in how mechanical stresses influence electrode durability, but this new approach offers both a stimulating way and the motivation to formulate a more basic understanding of the mechanics at play.

The scientists are now already at work to bring their methods to elucidate promising new battery designs at the atomic scale. One result might be new battery control software that accomplishes charging and discharging in ways that can enhance battery life.

Another exciting possibility is the development of more precise computational models that enable battery engineers to search for alternative electrode materials on a computer rather than in a lab.

That work is already underway. AI can help us look at old materials in new ways and maybe identify some promising alternatives from some as-yet-unknown materials.

William Chueh, Senior Study Author and Associate Professor of Materials Science and Engineering, Stanford University

Chueh is also a senior fellow at Stanford’s Precourt Institute for Energy and a principal investigator at the Stanford Institute for Materials and Energy Sciences.

Other Stanford co-authors not mentioned are Wei Cai, professor of mechanical engineering; Norman Jin, Ph.D. ’21; Ph.D. students Eder Giovanni Lomeli and Rui Yan; and Jueyi Liu, MS ’21.

Other co-authors of this research are scientists at Massachusetts Institute of Technology, Lawrence Berkeley National Laboratory, University of Lyon, Chungbuk National University, and the University of California-Berkeley.

This study received support from the Toyota Research Institute. Additional support was provided by the U.S. Department of Energy, Lawrence Berkeley National Laboratory, SLAC National Accelerator Laboratory, and the National Science Foundation.

Journal Reference:

Deng, H. D., et al. (2022) Correlative image learning of chemo-mechanics in phase-transforming solids. Nature Materials. doi.org/10.1038/s41563-021-01191-0.

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
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.