The research of Louis Vanduyfhuys focusses at developing thermodynamic and kinetic models for a characterization of adsorption and diffusion in nanoporous materials starting from molecular simulations. The main goal is not only to reproduce experimental results, but to understand the causal relations between molecular structure and macroscopic behavior and predict the behavior of new materials. As such, it fits in the “Model and Software Development” and “Nanoporous materials” research lines within the Center for Molecular Modeling. Such computational research further requires three important ingredients described below, which are briefly summarized below, and each of them hosts a variety of master thesis proposals.
First, computer simulations of molecular structures require a model for the intra- and intermolecular interactions at play at the atomic level. Such models can be ab initio (e.g. electronic density functional theory), force fields (parametrized expressions for the potential energy) or machine-learning potentials (neural networks that take the atomic configuration as input and are trained to predict the ab initio energy as output). Second, to make the bridge between microscopic simulations and macroscopic behaviour, advanced thermodynamic models based on the concepts of statistical physics are being developed and implemented in new software packages. These can range from models to construct, transform and interpret free energy profiles from simulation data and extract observable properties, to the simulation protocols themselves that allow for efficient phase space sampling (such as classical density functional theory). Finally, all these methodologies are used to characterize new materials such as MOF membranes for energy efficient gas separation or MOF/zeolite sensor arrays for detection of Volatile Organic Compounds (VOCs). As such, we do not only aim to understand and predict their behaviour, but also to allow for application-oriented design of new materials.