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

O: Fachverband Oberflächenphysik

O 40: Ultrafast electron dynamics at surface and interfaces – Poster (joint session O/TT)

O 40.4: Poster

Dienstag, 10. März 2026, 14:00–16:00, P2

Capturing thermalization through electron-electron scattering with machine learning — •David L. Kaiser, Tobias Held, Christopher Seibel, Markus Uehlein, Sebastian T. Weber, and Baerbel Rethfeld — Department of Physics and Research Center OPTIMAS, RPTU University Kaiserslautern-Landau

Ultrafast excitation of metals by optical laser pulses induces nonequilibrium energy distributions in the electron system. This nonequilibrium gives rise to complex electron-electron scattering processes, which typically restore a Fermi-Dirac distribution on a femtosecond timescale. Accurately modelling the thermalization requires evaluating the full electron-electron collision integral, which is however computationally costly.

In this study, we explore the possibility to use machine learning to emulate the dynamics generated by the full collision integral. Our goal is to significantly accelerate these calculations, enabling efficient simulations of bulk and multilayer systems. This approach opens the door to uncovering new relaxation pathways and predicting the response of complex material systems to ultrafast excitation.

Keywords: electron-electron scattering; nonequilibrium; ultrafast dynamics; Boltzmann equation; machine learning