Berlin 2024 – wissenschaftliches Programm

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

DY: Fachverband Dynamik und Statistische Physik

DY 34: Poster: Machine Learning, Data Science, and Reservoir Computing

DY 34.11: Poster

Mittwoch, 20. März 2024, 15:00–18:00, Poster C

Ab-initio-based interatomic potential for laser-excited Bismuth — •Jimiben Patel, Bernd Bauerhenne, and Martin Garcia — Institute for Physics, University of Kassel, Kassel, Germany

The intricate processes involving atomic motions, occurring on a sub-picosecond timescale, influence phenomena like chemical reactions, bond formation, and breaking. To get deep insights into these processes, femtosecond laser pulses have proven indispensable. The ultrashort interaction time in this context ensures that the laser field strongly influences electrons. A precise description of the structural relaxation of materials after femtosecond laser excitation is achieved through Te dependent Density Functional Theory (DFT). However, employing Te dependent DFT for simulations involving a large number of atoms is computationally expensive or even impossible. To address this challenge, our work introduces a polynomial Te dependent interatomic potential (Φ(Bi)(Te)) for Bismuth, which is trained using a database constructed from DFT simulations. Bismuth is an experimentally widely used material, for which so far no theory to describe ultrafast processes has been developed. In our analysis, we compared the physical properties of our polynomial potential Φ(Bi)(Te), with those obtained from ab initio simulations. Additionally, we conducted an examination of the thermophysical properties of the polynomial potential, including the determination of the melting temperature. This innovative approach allows for efficient and accurate exploration of complex material behaviors, offering a valuable alternative to computationally demanding Te dependent DFT simulations.

Keywords: Interatomic potential; laser-excited; Ab-initio-based; Bismuth