(2bo) A Data-Driven Approach to Materials Development for Emerging Separations Challenges

The pursuit of chemical separation units that are competitive with current best practices while also having lower energy requirements has spurred a flurry of research into materials development. Thousands of new materials have been developed across classes such as polymers and metal-organic frameworks, however, matching these materials with interesting applications has seen less progress. One of the primary reasons for this lack of progress is that scientists primarily attempt to navigate the huge material search space with heuristic-driven hypotheses based on a few material properties. While several success stories have emerged following this methodology, it follows a “solution in search of a problem” approach to research, ultimately reducing the rapidity at which important separation challenges can be addressed via these low-energy technologies. Machine learning strategies are ideally suited to address such a problem as they are designed to efficiently navigate large search spaces. With the increasing accessibility of machine learning and data science techniques, I believe a more data-driven approach to research will accelerate advances in the separations field. I propose a data-driven methodology following a “Top-Down” approach to separations technology development: 1) A target separation is selected, either because of industrial relevance or its potential to expand the fundamental understanding of a class of separations. 2) Using information available in open-source databases and the literature, a robust dataset of possible materials and their chemical and physical properties will be constructed. Machine learning will be used to assist in determining the necessary physical properties for effective separation. 3) Machine learning methods will be combined with experimental intuition to recommend materials with the target properties. By applying a rigorous materials informatics framework to separations research, I envision advancing both fields to more readily meet industry challenges.

Teaching Interests

Teaching is often not straightforward, therefore it is important to have clearly defined goals by which to determine effectiveness. My teaching philosophy incorporates three primary goals to guide course development and instruction: 1) Students develop a firm understanding of first principle concepts and how they are related 2) Students can use first principle concepts to explain macro-scale phenomena 3) Students can apply material learned in class to real-world examples. For engineers, everything starts with first principle equations. Without rate laws or thermodynamic equations, the field of chemical engineering would not have progressed as far as it has today. Hence, no matter the course, students must leave with a firm understanding of the basic equations. This does not mean to merely memorize the equations, but to understand conceptually what they mean. Visualizing the implications of first principle concepts helps with achieving my second teaching goal of ensuring that students can explain macro-scale phenomena through an underlying conceptual framework. While first principle understanding is important, it is of little direct use. Students must be able to see how theory relates to practice. Ultimately, students should be able to take concepts learned in the classroom and apply them to other scenarios once they join the workforce, especially engineers. This is the most important goal of my teaching philosophy since it is the most important indicator of how much students actually learned. Since I am a separations scientists, I am eager to teach any separation-focused courses. Thermodynamics is also a foundational subject for separations and indeed all of chemical engineering, so I would enjoy helping students navigate this important topic as well. With clear goals and teaching strategies, I hope that I can give every student an opportunity to achieve their ambitions.