Regensburg 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
KFM: Fachverband Kristalline Festkörper und deren Mikrostruktur
KFM 17: Functional Materials: Performance, Reliability and Degradation; and Complex Materials (joint session MM/KFM)
KFM 17.2: Talk
Thursday, March 20, 2025, 12:00–12:15, H23
Searching for ferroelectric porous metal organic frameworks using machine-learning and Monte-Carlo-simulations — •Thomas Bergler1,2, Harald Oberhofer1,2, and Dirk Volkmer3 — 1University of Bayreuth, Germany — 2Bavarian Center for Battery Technologies — 3University of Augsburg, Germany
Metal organic frameworks (MOFs) have so far found a number of successful applications, among them as storage for gasses and filter for gas mixtures. So far these mostly incorporated them as passive materials, but recent research points the way towards a more active role, possibly through the external manipulation of the materials' internal properties. One recent example for such a property is the susceptibility of the lattice parameters of a number of MOFs towards electric fields. Inspired by this, the aim of our project is to further investigate this behavior and potentially design ferroelectric MOFs. Using a hierarchy of Monte-Carlo-simulations aided by Machine-Learning (ML) we sample the design space MOFs augmented by rotatable polar groups. In succession, we first sample a huge space of rotors in a simplified point-dipole model. A selection of thus uncovered MOF geometries is then investigated with a specially parameterized atomistic model to confirm earlier predictions. Using this data, an ML model is trained to predict the dielectric properties of such polar rotor-augmented MOFs. The best candidates extracted with this procedure are finally evaluated with density functional theory. MOF geometries surviving this funnel-like approach can finally be checked experimentally for a variety of applications, ranging from data-storage to gas nano-funnelling.
Keywords: metal organic framework; machine-learning; Monte-Carlo; material design; ferroelectricity