Speaker: Maren Podewitz, TU Wien

Computational chemistry has become an important tool for accelerating experimental discovery by exploring the reaction mechanisms of catalytic systems, guiding experimental design. However, achieving robust and transferable insights requires methodological advances in system representation and computational strategy. Reliable results depend on developing and selecting appropriate chemical and computational models, both of which are evolving continuously. In recent years, the field has moved beyond static, single-structure, implicit-environment studies towards ensemble-based, dynamic explorations with explicit solvation. This opens up new opportunities, but also raises challenges in terms of computational scalability, model reliability, and data analysis.

This talk will focus on methodological developments in modelling catalytic mechanisms across multiple levels of complexity. First, I will present advances in static modelling and highlight microsolvation strategies that identify strongly interacting solvent molecules, thereby improving the description of explicit environments [1, 2]. Then, I will discuss emerging tools for reaction dynamics and demonstrate how explainable machine learning (ML) approaches and time-resolved analyses provide mechanistic insights that are inaccessible to static models [3]. Together, these developments illustrate how new frameworks and computational tools can expand the accessible free energy landscape and enable a deeper, more predictive understanding of catalytic reactivity.

[1]         M. Steiner, T. Holzknecht, M. Schauperl, M. Podewitz, Molecules 26, 1793 (2021).
[2]         L. B. Magenheim, R. A. Talmazan, M. Podewitz. Manuscript in preparation (2025).
[3]         R. A. Talmazan, J. Gamper, I. Castillo, T. S. Hofer, M. Podewitz*. ChemRxiv. 2024; doi:10.26434/chemrxiv-2024-nd20j.