Know what is the In thing now

Green Energy: AI-Designed Battery Reduces Lithium Use by 70%

AI designed battery with less lithium

Researchers have employed artificial intelligence (AI) to design a groundbreaking battery that uses up to 70% less lithium compared to existing designs. Lithium-ion batteries, prevalent in devices and electric vehicles, play a crucial role in storing renewable energy for green electric grids. However, the mining of lithium is costly and environmentally damaging. Addressing this, Nathan Baker and his team at Microsoft harnessed the power of AI to streamline the material discovery process, completing the task in months rather than years.

The research specifics

Focusing on batteries with only solid components, the researchers sought innovative materials for the electrolyte, the part through which electric charges move. Beginning with 23.6 million candidate materials, the AI algorithm rapidly narrowed down the options by identifying stability issues and weak chemical reactions. After a few days, a list of a few hundred promising candidates emerged, some of which were previously unstudied.

To validate their findings, the researchers consulted battery experts, including Vijay Murugesan at the Pacific Northwest National Laboratory. Further screening criteria were suggested, leading to the selection of an AI-recommended material for lab synthesis. Notably, this electrolyte featured a novel composition, with sodium replacing half of the expected lithium atoms, raising questions about the fundamental physics of its behavior in a battery.

Despite lower conductivity compared to prototypes using more lithium, Murugesan’s team successfully constructed a functional battery. Although Baker and Murugesan acknowledge the need for further optimization, the entire process – from initial discussions to a functional battery powering a light bulb – took approximately nine months.

The significance

Rafael Gómez-Bombarelli from MIT praised the project, highlighting the integration of AI with practical experimentation. While acknowledging the cutting-edge nature of the machine learning tools employed, he emphasized the significance of translating predictions into tangible experiments. However, potential challenges lie ahead, with sparse data for training AI in this domain and the possibility of more complex combinations for materials beyond battery components. This AI-driven breakthrough holds promise for reducing reliance on lithium in batteries, contributing to a more sustainable energy landscape.

You might also be interested in

Get the word out!