September 24, 2025
Dhruv Menon
Doctoral Candidate in Chemical Engineering & Biotechnology
EPSRC CDT in Nanoscience & Nanotechnology
Cavendish Laboratory
University of Cambridge
Winton-Kavli Exchange Fellow, Yaghi Group
"De novo design of metal-organic frameworks using generative deep learning"
The emergence of ‘reticular chemistry’ has facilitated the geometry-guided design of periodically extended crystalline structures by linking molecular building blocks through strong bonds. The pioneering class of reticular materials, metal-organic frameworks (MOFs) are of interest across diverse technologies due to their exceptional surface areas and ability to interact with other chemical species, such as gases and liquids, not only at their external surface but also throughout their internal porosity. Yet, their precursor space – spanning inorganic and organic building blocks – is effectively unbounded with over 100,000 MOFs reported and millions more theorized. Traditionally, discovery has relied on experimental expertise and chemical intuition, yielding standout materials but limited coverage and weak translation to industry. In this context, computational methods can disrupt experimental timescales. This seminar overviews computational strategies for de novo MOF design using generative deep learning. We will discuss the development of chemical language models based on molecular transformers for theory-guided discovery of MOF building blocks, and pair them with diffusion or conditional flow-matching generators to optimize candidates for targets such as water harvesting and gas storage. We show how these models focus the search, prioritize high-value candidates, and reveal design trends for high-performance materials. Finally, we argue for integrating computation into experimental pipelines to streamline – and inspire – new experiments.
Biosketch
Dhruv Menon completed his Bachelor of Technology in Materials Science and Engineering at the Indian Institute of Technology (IIT) Gandhinagar. He subsequently completed his Masters degree in Physics at the Cavendish Laboratory, University of Cambridge. He is currently a second year PhD student in Chemical Engineering at Cambridge, supervised by Professor David Fairen-Jimenez. His work focuses on the development of computational approaches for the discovery, design and optimization of MOFs – primarily for drug delivery applications. His work places a strong focus on clinical and industrial translation. He recently joined the group of Professor Omar Yaghi as a Winton-Kavli Exchange Fellow to develop generative deep learning models for the discovery of MOF water harvesters.