Artificial intelligence was used to discover a brand new substance that could make safer, more energy-dense and less resource-intensive batteries. Microsoft and PNNL on Wednesday announced that they used advanced AI to screen more than 32 million potential candidates to discover and synthesise the new materials that could help create better batteries.
Researchers used new AI methods to screen the 32 million materials and found over 500,000 stable candidates. But identifying half a million candidates is just the beginning. Finding a material with the right properties among them wouldn’t be dissimilar to looking for a needle in a haystack. Typically, that would involve high-performance computing and lab experimentation that could take “multiple lifetimes” to complete, according to Microsoft.
Microsoft’s Azure Quantum team worked with PNNL, which is part of the United States Department of Energy, to identify the new material which was unknown and not present in nature. After that, PNNL scientists synthesised and tested the material candidate to develop a working prototype. This demonstrated the material’s unique properties and its potential for use as energy-storage material—it used significantly less lithium and other materials.
That is how they developed the new solid-state electrolyte, a material that is also less likely to burst into flames than today’s lithium-ion batteries that use liquid electrolytes. Not only are solid-state batteries considered safer than conventional batteries but they could also provide more energy density.
Further, lithium is quite expensive and relatively scarce; mining it is problematic both environmentally and socially. This new battery material could reduce the use of lithium by as much as 70 per cent, with massive economic, social, and environmental benefits.
Of course, there is still a lot of work left to be done to see whether the new material is actually viable for use at a commercial scale. But the most exciting thing about the discovery is how AI was used to speed up a material science process that would have been so much longer and more expensive otherwise.