Many technologies are hindered by the properties of their component materials (their cost, performance, or lifetime). Today, it generally takes about 20 years to develop and commercialize new materials. In this talk, I will describe new computational and data mining techniques that are accelerating the pace at which materials with advanced properties can be discovered.
I will describe how it is now possible screen hundreds of thousands of chemical compositions virtually for various applications using a combination of high-throughput computing and the latest advancements in theoretical materials science and condensed matter physics.
How have such techniques resulted in new materials candidates across a wide variety of applications - including Li-ion batteries and thermoelectric materials for waste heat conversion? I will discuss our effort to crawl the scientific literature using natural language processing techniques, uncovering patterns that can used to predict the next breakthrough materials for an application. We have found, for example, that by analyzing over 3 million scientific abstracts we are able to suggest new thermoelectric compositions many years ahead of when they would likely have been found through business as usual. Finally, I will extrapolate where these advancements may lead in the future.
Speaker: Anubhav Jain, Lawrence Berkeley National Labs
Contact:Website: Click to Visit
Save this Event:iCalendar
Windows Live Calendar