ZURUECK HOCH VOR INHALT SUCHEN

» Back to overview
Proposing Institution

Institut für Physikalische Chemie, Georg-August-Universität Göttingen,
Project Manager

Marco Eckhoff
Tammannstraße 6
37077 Göttingen
Abstract
The lithium manganese oxide spinel LixMn2O4, with 0<x<2, is a prominent example of cathode materials in lithium ion batteries which offers advantages such as low costs and non-toxicity.However, a theoretical treatment using density functional theory is far from trivial due to its complex electronic structure, with a variety of close-lying electronic and magnetic states. The local density approximation as well as the generalised gradient approximation are unable to describe LixMn2O4 accurately. Hybrid functionals yield energetic, structural, electronic, and magnetic properties which are in good agreement with experimental measurements.Addressing the lithium diffusion process, for example, by ab initio molecular dynamics simulations is computationally very demanding, as large systems are required for realistic structural models. This limitation can be overcome by employing high-dimensional neural network potentials, which combine the accuracy of first-principles calculations with the efficiency of empirical potentials.In this study, a high-dimensional neural network potential is developed for LixMn2O4 in water.

Impressum, Conny Wendler