Defect-induced melting and solid-state amorphization

Abstract

DESPITE the strong interest in melting over the past 100 years, a general theory for the crystal–liquid transition has not been established1. Lattice-instability models, which are either vibrational2, elastic3, isochoric4, defective5 or entropic6 in nature, all predict a melting point somewhat above the experimentally observed thermodynamic melting temperature, with the ultimate stability limit of a superheated crystal being determined by the equality of crystal and liquid entropies4,6; this forces regular melting to be a first-order transition. Here I present a model of melting that is driven by the incorporation into the lattice of randomly frozenin defects. An isentropic condition limits the stability of the crystal as a function of defect concentration; above the glass transition temperature the crystal melts to a liquid, whereas below it 'melting' produces an amorphous solid. This model yields a generic melting diagram with a tunable parameter (defect concentration) that can characterize the static disorder present in solid-state amorphization7–9, the thermodynamic stability of small clusters10 and nanocrystalline materials11, and the frustration present in spin glasses12. The model is also relevant to glacial13, geological14 and stellar-atmospheric15 melting processes.