Pitfalls and successes of the classical nucleation theory. A 50-year retrospective (PLENARY TALK)

Reference Presenter Authors
Edgar Dutra Zanotto Zanotto, E.D.(Universidade Federal de São Carlos); Crystallization is a vital phenomenon that controls many natural and technological processes, such as ice formation, metal solidification, the shelf-life of certain pharmaceutical drugs, and the development and thermal stability of glasses and nanostructured glass-ceramics. The Classical Nucleation Theory (CNT) is a very good descriptor of the temperature dependence of homogeneous nucleation rates in supercooled liquids. However, in tests and applications, due to the scarcity of experimental data, one normally uses a fitted value of the nucleus/liquid interfacial energy, and viscosity data as a proxy for the diffusivity. In that case, however, the CNT fails in describing the actual magnitudes of the nucleation rates, leading to discrepancies between experimental and theoretical values of up to 50 orders of magnitude! (some researchers claimed that this is the worst theory ever). This apparent inability of the theory to describe the magnitudes of nucleation rates been associated with assumptions related to the thermodynamic and diffusional parameters. In this work, we revise most of the possible explanations for this failure (the breakdown of the Stokes-Einstein relation, formation of metastable phases, and negligence of elastic stresses, among others). Finally, we eliminate most of these uncertainties through molecular dynamics (MD) simulations and re-test the theory. We will leave the main results as a surprise to the attendees.
<< Back