Abstract: Satellites, defense systems, and other advanced technologies depend on critical materials like samarium and dysprosium. But separating these nearly identical elements is one of chemistry’s most persistent challenges. These materials are primarily imported, with limited domestic supply. Traditional separation methods are slow, toxic, and costly. My research takes a different approach: combining safe kitchen chemistry with artificial intelligence to explore the molecular forces that drive separation. With machine learning, we can accelerate discovery, reduce trial and error, and develop a faster, smarter path to separation. This reduces reliance on imports and helps secure materials our energy systems and national infrastructure depend on.

Bio: Karen is a postdoctoral research associate at INL, specializing in machine learning approaches for critical materials separation. Her research focuses on developing generative artificial intelligence tools to support human-centered and robust nuclear safety applications. She received her PhD in computer science from Purdue University.