Defect-unbinding transitions and inherent structures in two dimensions

TL;DR

This large-scale computational study of two-dimensional Lennard-Jones systems supports the inherent-structures theory and KTHNY theory by identifying three phases and observing defect-unbinding transitions in the inherent structures.

Contribution

It provides the first direct computational evidence linking inherent structures to the defect-unbinding transitions in two-dimensional melting.

Findings

Identification of three distinct inherent structure phases: crystal, hexatic glass, liquid glass.

Observation of defect-unbinding transitions in inherent structures.

Bond-orientational correlations match equilibrium phase behavior.

Abstract

We present a large-scale (36000-particle) computational study of the "inherent structures" (IS) associated with equilibrium, two-dimensional, one-component Lennard-Jones systems. Our results provide strong support both for the inherent-structures theory of classical fluids, and for the KTHNY theory of two-stage melting in two dimensions. This support comes from the observation of three qualitatively distinct "phases" of inherent structures: a crystal, a "hexatic glass", and a "liquid glass". We also directly observe, in the IS, analogs of the two defect-unbinding transitions (respectively, of dislocations, and disclinations) believed to mediate the two equilibrium phase transitions. Each transition shows up in the inherent structures---although the free disclinations in the "liquid glass" are embedded in a percolating network of grain boundaries. The bond-orientational correlation functions of the inherent structures show the same progressive loss of order as do the three equilibrium phases: long-range to quasi-long-range to short-range.