Li-ion batteries with polymer-electrolytes still have several disadvantages related to costs, safety, and materials’ availability. It would be desirable to develop alternative storage devices comprising solid-state electrolytes and sodium ions, which may replace current Li-ion batteries for certain applications. The Na2Ge-GeSe2 chalcogenide glass-ceramic was proposed as a promising material to be used as a solid electrolyte for Na-ion batteries. The complex chemistry and structural heterogeneity of this material are hardly described by empirical potentials, calling for first-principles modeling. However, first-principles simulations involve high computational costs and cannot scale to the size and time scales necessary to characterize the functional behavior of this material, thus requiring a multiscale approach. <br/>Initially, we show that parameter-free ab initio molecular dynamics (AIMD) simulations reproduce the experimental properties of Ge-Se melts at different stoichiometry, and predict the structural variations induced by the inclusion of Na. In this study, we overcome the size and time scale limitations of AIMD by constructing suitable machine-learning potentials - based on the Chebyshev Polynomials paradigm (ChIMES) - with which we can simulate quenching from the melt and Na diffusion. This study will allow us to finally characterize amorphous Na-intercalated GeSe2 and predict its ionic conductivity at various conditions.

Modeling the amorphous Na2Ge-GeSe2 chalcogenide system

ACS Fall 2020 Virtual Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.