Dresden 2026 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

FM: Fachverband Funktionsmaterialien

FM 3: Focus Session: Novel mechanisms of ferroic switching (joint session MA/FM)

FM 3.1: Hauptvortrag

Montag, 9. März 2026, 09:30–10:00, POT/0151

From ML to Kinetics: Modeling the Switching in Ferroelectric Wurtzites — •Andrew Rappe, Drew Behrendt, Atanu Samanta, and Von Braun Nascimento — Department of Chemistry, University of Pennsylvania, Philadelphia, PA 19104-6323 USA

We present a series of works, from the development of a new force field (MLFF) for multi-scale simulations of bulk AlN, through application of MLFF to understand the atomistic switching mechanism, to the development of a new kinetic model to uncover how switching changes as a function of experimental conditions.

We train our MLFF to 1000s of DFT calculations, so the underlying calculations are as accurate as DFT with the flexibility in simulation size of classical MD. Powered by the MLFF, we can predict the energies, forces, and phonon dispersions of AlN at dramatically lower cost, thus enabling the study of emergent and long-range effects, such as the frequency-dependent dielectric functions and multiple FE domains.

Applying the AlN MLFF, we uncover the atomistic mechanism of domain wall (DW) migration and domain growth in wurtzites. We find that the critical nucleus is a single broken Al-N bond along the polar axis; this creates a cascade of bond breaking in a single column of atoms due to the stability of the 180° DW in wurtzites. We reveal the switching mechanism of 1D atomic columns propagating from a slow-moving 2D fractal-like DW in the basal plane.

Finally, we develop an analytical extension to the KAI model that accounts for fractal FE domains. To do this, we take the traditional model of circles that can nucleate and grow and add a budding term.

Keywords: AlN; wurtzite; ferroelectric; switching; machine learning