Hylleraas Seminar, Lucía Morán

Speaker: Lucía Morán

Title: Integrating in silico design and data-driven strategies for advanced homogeneous bifunctional catalysts in the CO₂ to methanol transformation

Abstract: The rise in CO2 emission is driving the exploration of innovative strategies to convert CO2 into functional chemicals, such as methanol. Hydrogenation of CO2 to produce methanol can be catalyzed by various types of catalysts. In the realm of homogeneous catalysis, the development of bifunctional catalysts (BFCs) has demonstrated potential for efficient conversion processes under relatively mild conditions. However, improving conversion rates and reusability remains a challenge. In our pursuit of discovering new efficient catalysts, we employ a data-driven strategy utilizing the tmQMg-L dataset and the Ru-MACHO framework as scaffold. The design of catalysts is achieved by varying the metal centre, the pincer ligand, the trans-ligand, and the axial ligand resulting in the generation of over one million potential catalysts. Subsequently, active machine learning is employed to simultaneously optimize two primary target properties in the reaction: catalyst hydrogenation and poisoning. The objective of this computational screening is to identify the most promising candidates, thereby accelerating the catalyst discovery process.