Event

Machine learning driven discovery of 2D materials for sustainable energy solutions

Thursday, January 16, 2025
11:00 a.m.-12:00 p.m.
1202 Engineering Lab Building
Catherine Stephens
301 405 9378
csteph5@umd.edu
http://energy.umd.edu

Energy Matters Lecture: Machine learning driven discovery of 2D materials for sustainable energy solution

Moses Abraham, A.J. Drexel Nanomaterials Institute, and Department of Materials Science and Engineering Drexel University, Philadelphia 19104, USA.

Email: mb4522@drexel.edu

Abstract:

To address the global sustainability challenges and ensure a better quality of life, it is essential to develop cutting-edge energy generation solutions such as water-splitting devices, fuel cells and rechargeable batteries. The success of these technologies is highly dependent upon materials that are both environmentally friendly and economically viable, but finding them is limited by the property trade-offs and multidimensional complexities. Particularly, the field of catalysis experiences hurdles from the random nature of screening numerous materials under operating conditions, a process that heavily depends on traditional trial-and-error strategies and restricts practical applications. Artificial intelligence, employing data mining and machine learning (ML) tools revolutionizes material discovery by accelerating the screening process through efficient exploration of vast chemical space¹. MXenes, a class of two-dimensional (2D) materials derived from the selective etching of layered carbides, have emerged as promising candidates for development of innovative energy materials. This presentation highlights the pivotal role of ML in extracting knowledge from MXenes by using implicit data patterns and complex correlations, providing intelligent guidance for the development of advanced catalysts. I will talk about the methods to enhance catalyst performance and explore their impact on the future of sustainable energy technologies. Following this, I will present our recent research on a multistep procedure using supervised ML algorithms to build data-driven models that predict HER and CO₂ activation across thousands of MXene configurations². I will also discuss how computational and experimental integration can accelerate the discovery of efficient catalysts. To conclude, I will present future directions where ML can accelerate autonomous experimentation, quantum-based catalytic innovations and facilitate the development of multifunctional catalysts for future energy solutions.

References:

1. M. Abraham et al., "Catalysis in the Digital Age: Unlocking the Power of Data with Machine Learning" WIREs Computational Molecular Science 14 (5), e1730 (2024).
2. M. Abraham et al., "Fusing machine learning strategy with density functional theory to hasten the discovery of 2D MXene based catalysts for hydrogen generation" J. Mater. Chem. A, 11 (15), 8091 (2023).

Audience: Graduate  Faculty