The new battery has lower conductivity than similar prototypes that use more lithium, and further work is needed to optimize it. However, the use of AI significantly accelerated the process of finding and testing new materials. The approach could face future challenges, as the data needed to train the AI is often sparse and other materials may require a more complex method of combining elements.
Key takeaways:
- Researchers at Microsoft used AI to design a new material for a working battery that requires up to 70 percent less lithium than some competing designs.
- The AI started with 23.6 million candidate materials and eliminated those that would be unstable or have weak chemical reactions, narrowing the list down to a few hundred candidates in just a few days.
- The chosen material, suggested by the AI, replaced half of the expected lithium atoms with sodium, a novel recipe for an electrolyte.
- The process of making the battery, from initial discussions to a functional product, took about nine months, demonstrating the potential of AI to accelerate the development of new materials.