Uncertainties in permittivities computed from molecular dynamics simulations and temperature correction of dielectric properties of condensed polar systems

TL;DR

This paper introduces a statistical method to estimate uncertainties in dielectric permittivity from molecular dynamics simulations and proposes a temperature correction approach for better prediction of dielectric properties in polar systems.

Contribution

It presents a novel, efficient procedure for uncertainty quantification in dielectric permittivity calculations and offers a temperature correction method using minimal experimental data.

Findings

Uncertainty estimation for dielectric permittivity using statistical methods.

A temperature correction approach for dielectric properties.

Validation of the method with molecular dynamics data.

Abstract

A robust, simple and fast procedure for the calculation of uncertainties in relative static dielectric permittivity ($\varepsilon_s$) computed via molecular dynamics (MD) is proposed. It arises as a direct application of well founded statistical methods for auto-correlated variables. Also, in order to deal with the lack of experimental data about $\varepsilon_s$ and relaxation times ($\tau$) at different temperatures, a method for their prediction is suggested. It requires one experimental value and at least two MD simulations. In the case of relaxation times, a theoretical justification is provided.