Instabilities govern the low-frequency vibrational spectrum of amorphous solids

TL;DR

This paper investigates how boundary-condition-induced instabilities influence the low-frequency vibrational modes in amorphous solids, revealing different scaling behaviors and proposing limits based on stability and stress conditions.

Contribution

It identifies two types of elastic branches affecting vibrational spectra and demonstrates how boundary conditions and residual stresses alter the low-frequency vibrational density of states.

Findings

Fictitious branches show a $\omega^3$ scaling.

True elastic branches exhibit a $\omega^{5.5}$ scaling.

Removing residual shear stress leads to a $\omega^{6.5}$ scaling.

Abstract

Amorphous solids exhibit an excess of low-frequency vibrational modes beyond the Debye prediction, contributing to their anomalous mechanical and thermal properties. Although a $\omega^4$ power-law scaling is often proposed for the distribution of these modes, the precise exponent remains a subject of debate. In this study, we demonstrate that boundary-condition-induced instabilities play a key role in this variability. We identify two distinct types of elastic branches that differ in the nature of their energy landscape: Fictitious branches, where shear minima cannot be reached through elastic deformation alone and require plastic instabilities, and True branches, where elastic deformation can access these minima. Configurations on Fictitious branches show a vibrational density of states (VDoS) scaling as $D(\omega) \sim \omega^3$, while those on True elastic branches under simple and pure shear deformations exhibit a scaling of $D(\omega) \sim \omega^{5.5}$. Ensemble averaging over both types of branches results in a VDoS scaling of $D(\omega) \sim \omega^4$. Additionally, solids relaxed to their shear minima, with no residual shear stress, display a steeper scaling of $D(\omega) \sim \omega^{6.5}$ in both two and three dimensions. We propose two limiting behaviors for amorphous solids: if the system size is increased without addressing instabilities, the low-frequency VDoS scales with an exponent close to $3$. Conversely, by removing residual shear stress before considering large system sizes, the VDoS scales as $D(\omega) \sim \omega^{6.5}$.