Analysis of spatial correlations in a model 2D liquid through eigenvalues and eigenvectors of atomic level stress matrices

TL;DR

This study uses eigenvalues and eigenvectors of atomic stress matrices from molecular dynamics simulations to analyze structural correlations in a 2D liquid, revealing temperature-dependent stress correlations and clarifying the nature of angular dependencies.

Contribution

It introduces a novel eigenvalue/eigenvector-based method to characterize structural correlations in disordered materials, challenging previous interpretations of stress field anisotropies.

Findings

Stress correlation functions vary with temperature, especially at the first minimum of the pair density.

Angular dependencies are due to tensor rotational properties, not anisotropic stress fields.

The approach effectively characterizes structural correlations in disordered 2D liquids.

Abstract

Considerations of local atomic level stresses associated with each atom represent a particular approach to address structures of disordered materials at the atomic level. We studied structural correlations in a two-dimensional model liquid using molecular dynamics simulations in the following way. We diagonalized the atomic level stress tensors of every atom and investigated correlations between the eigenvalues and orientations of the eigenvectors of different atoms as a function of distance between them. It is demonstrated that the suggested approach can be used to characterize structural correlations in disordered materials. In particular, we found that changes in the stress correlation functions on decrease of temperature are the most pronounced for the pairs of atoms with separation distance that corresponds to the first minimum in the pair density function. We also show that the angular dependencies of the stress correlation functions previously reported in [Phys. Rev. E v.91, 032301 (2015)] related not to the alleged anisotropies of the Eshelby's stress fields, but to the rotational properties of the stress tensors.