Research Overview
My research is situated at the interface of molecular simulation, statistical thermodynamics and materials modeling, with a strong focus on stimuli-responsive nanoporous materials, in particular metal–organic frameworks (MOFs). In collaboration with other staff members of CMM, I aim to understand and predict how microscopic interactions give rise to macroscopic thermodynamic and kinetic properties under realistic operating conditions, in particular for adsorption and diffusion. These insights enable the computationally guided design of advanced materials for applications such as gas separation, gas sensing and energy storage.
Figure taken from Van Speybroeck, V., Philos. Trans. R. Soc. A. 2023, 381 (2250), 20220239
A central theme of my work is the development of integrated simulation workflows (as illustrated in the figure above) that start from realistic representations of the molecular structure, accurate level of theory to compute the potential energy, advanced sampling techniques of the phase space, and rigorous analysis of the results to obtain measurable physical and chemical properties.
Methodology
- Potential Development. I develop high-fidelity interaction potentials for predictive molecular simulations of complex materials. While my earlier work focused on physically motivated classical force fields derived from ab initio reference data, my current research increasingly emphasizes machine-learning interatomic potentials. These data-driven models combine near–quantum-mechanical accuracy with the efficiency required to access large systems and long timescales. A key focus is the construction of reliable training datasets, validation of transferability across phases and thermodynamic conditions, and seamless integration of machine-learning potentials into molecular simulation and free-energy workflows.
- Advanced Molecular Simulation. The developed potentials are employed in Molecular Dynamics, Monte Carlo, and classical Density Functional Theory (cDFT) simulations, combined with enhanced sampling techniques such as umbrella sampling, thermodynamic integration, metadynamics, variationally enhanced sampling, replica exchange and transition matrix Monte Carlo. These approaches enable efficient exploration of complex free-energy landscapes associated with structural flexibility, phase transitions, adsorption, and diffusion.
- Advanced Statistical-Thermodynamical Modelling. From these simulation results, I extract thermodynamic properties (such as free energy profiles, henry constants, adsorption isotherms), kinetic properties (such chemical reaction rates and diffusivities) and dielectric properties (such as dielectric constant and refractive index). This is supported by development of advanced models in thermodynamics and statistical physics (as implemented in ThermoLIB, see below) that allow to (1) construct, transform and (de)project free energy surfaces (FESs) in terms of collective variables, (2) decompose these FESs into entropic and enthalpic contributions, (3) compute related thermodynamic and kinetic properties and (4) reliably estimate the sampling uncertainty resulting from the finite simulation time.
Software Development
An important component of my research is the development of open-source computational tools that support reproducible molecular modeling:
ThermoLIB - https://github.com/molmod/ThermoLIB: a software framework for constructing, analyzing, transforming and (de)projecting free-energy profiles in terms of collective variables including a reliable uncertainty estimation.
CmmDFT - https://github.com/molmod/CmmDFT: a Python package for classical density functional theory calculations, enabling efficient prediction of adsorption properties of nanoporous materials.
QuickFF - https://github.com/molmod/QuickFF: a Python package for the automated derivation of accurate classical force fields from ab initio reference data.
Together, these tools form a coherent software ecosystem that underpins my methodological research and its applications.
Applications
Most application-driven work focuses on MOFs, a class of hybrid crystalline materials with tunable nanoporosity and chemistry. Many MOFs have a pronounced structural flexibility and respond to temperature, pressure, or guest adsorption through small local structural changes (e.g. gate-opening in ZIF-8) or large collective phase transitions (e.g. breathing in MIL-53). Some of these materials even exhibit counterintuitive thermo-mechanical behavior, such as negative thermal expansion, negative linear compression, and negative gas adsorption. My research aims to quantitatively characterize the mechanical, thermal, adsorption and diffusion properties underlying these phenomena, providing design principles for high-performance materials in gas separation and sensing, catalysis, and nano-mechanical energy storage.
Teaching
I am the responsible lecturer for the following courses at Ghent University:
Modelling and Engineering of Nanoscale Materials
Quantum Mechanics I
Physics III (Thermodynamics and Statistical Physics)
Statistical Physics