When browsing through theoretical lists of possible new supplies for particular purposes, such as batteries or different energy-associated units, there are often thousands of potential materials that might be considered and several criteria that should be met and optimized at once. Now, researchers at MIT have found a way to dramatically streamline the discovery procedure, utilizing a machine learning technology.
As a demonstration, the group arrived at a set of the eight most promising materials, out of almost 3 million candidates, for an energy storage system known as a flow battery. This culling procedure would have taken five decades by standard analytical methods, they say; however, they accomplished it in five weeks.
The discoveries are reported in the journal ACS Science, in a paper by MIT Prof. of chemical engineering Heather Kulik, Jon Paul Janet Ph.D. ’19, Sahasrajit Ramesh, and graduate pupil Chenru Duan.
To predict the properties of any one of hundreds of thousands of these supplies would require either time-consuming and resource-intensive spectroscopy and other lab work, or time-consuming, extremely complex physics-based computer modeling for 6possible candidate material or combination of materials. Each such research could take hours to days of work.