March 21, 2024

Adsorption modeling in nanoporous materials using machine learning

MOFs are nanoporous structures, capable of being tuned in their structure and chemistry to accommodate a wide range of applications, including gas storage, drug delivery, catalysis and sensing. At the core of these applications, modeling and understanding the adsorption properties of MOFs is a fundamental step. However, computational modeling via high-throughput molecular simulations are too expensive both in time and computational resources. Hence, novel approaches to modeling are needed to efficiently find the appropriate MOF for a given application. 

In this talk, ML models for the adsorption of nitrogen and carbon dioxide in MOFs will be presented. These models use the XGBoost algorithm to estimate the adsorption capacity of a material, given a temperature and pressure, using as input both structural and energetic features.

About the speaker

Juan received his bachelor's degree in pure mathematics from Universidad Nacional de Colombia, the top university in his home country. He came to UTRGV and received his master's degree in mathematics, and along the way he fell in love with statistics and machine learning, and their applications to big data. Currently he is a doctoral student at UTRGV in the Ph.D in Mathematics and Statistics with Interdisciplinary Applications, looking to find applications and solve real-world problems with the many tools learned in his journey.