The molecular-level insights gathered through “in silico” studies have become an essential asset for the elucidation of complex reaction mechanisms. Indeed, the applicability of computational chemistry has strongly widened due to the vast increase in computational power happening along the last decades. In this sense, not only the accuracy of the applied methods or the size of the target systems have increased, but also the level of detail attained for the mechanistic description. The ultimate goal of these approaches is to gain a better understanding of the underlying processes that may then lead to improve their performance: such a rational optimization approach is a cornerstone for the eventual development of cleaner, more efficient chemical transformations. However, performing this kind of deeper descriptions of chemical systems, most often resorting to automation techniques that allow to easily explore larger parts of the chemical space, comes at the cost of also augmenting their complexity, rendering the results much harder to interpret. Throughout this Thesis, we have proposed, developed and tested a collection of tools aiming to process this kind of complex chemical reaction networks (CRNs), in order to provide new insights on reactive and catalytic processes. All of these tools employ graphs to model the target CRNs, in order to be able to use the methods of Graph Theory (e.g. path searches, isomorphisms…) in a chemical context. The tools that are discussed include amk-tools, a framework for the interactive visualization of automatically discovered reaction networks, gTOFfee, for the application of the energy span model to compute the turnover frequency of computationally characterized catalytic cycles, and OntoRXN, an ontology for the description of CRNs in a semantic manner integrating network topology and calculation information in a single, highly-structured entity.
If you would like to follow the ceremony on Zoom, please register here.