Overcoming sampling limitations using machine-learned interatomic potentials: the case of water-in-salt electrolytes

TL;DR

This paper demonstrates that machine-learned interatomic potentials can effectively model highly concentrated water-in-salt electrolytes, overcoming sampling limitations of ab initio methods and accurately reproducing experimental data.

Contribution

It evaluates the performance of MACE potentials for water-in-salt electrolytes, highlighting the benefits of fine tuning foundation models and analyzing the impact of dispersion corrections.

Findings

Surrogate models match experimental structure factors.

Fine tuning improves data efficiency and sampling of rare configurations.

Dispersion correction schemes can be detrimental depending on the functional.

Abstract

Machine-learned interatomic potentials hold the promise to enable the modeling of highly concentrated liquids over meaningful timescales, far from reach for current ab initio electronic structure methods. Here we evaluate the performances of various MACE potentials in modeling a $21 m$ water-in-salt electrolyte based on lithium bis(trifluoromethanesulfonyl)imide. We test out-of-the-box foundation models, as well as both fine tuning and from scratch training strategies. Our simulations demonstrate that surrogate models allow to overcome sampling limitations of ab initio molecular dynamics, reaching an excellent agreement with experimental observables such as the structure factor. We also demonstrate the benefit of fine tuning a foundation model over training from scratch: in terms of data efficiency, but most importantly as a means to provide information regarding configurations hard to sample, such as short Li$^+$--Li$^+$ distances. Finally, we show that depending on the reference exchange-correlation functional, empirical dispersion correction schemes can be detrimental. All in all, our work shows that machine-learned interatomic potentials are a good fit for the modeling of highly concentrated electrolytes over long timescales.