Crystallization of glass-forming melts: New answers to old questions

In classical nucleation theory, the thermodynamic driving force of crystallization is computed by comparing the bulk properties of the liquid and crystal phases for the respective macroscopic samples. Such approach represents a reasonable approximation as far as the bulk properties of the critical clusters do not deviate significantly from the properties of the newly evolving macroscopic phases. However, in general, this assumption is not true and may lead to incorrect results. As one consequence, the classical approach utilizing, in addition, the capillarity approximation overestimates as a rule the work of critical cluster formation and underestimates the steady-state nucleation rate. Based on a generalization of the classical Gibbs' approach, the basic equations are formulated here allowing one a correct determination of the properties of the critical clusters and of the work of critical cluster formation. In this way, a new tool for the quantitative description of crystal nucleation has been advanced retaining the advantages of the classical treatment but avoiding its shortcomings. Several general consequences are discussed. In particular, it is demonstrated that this method opens a new perspective in the interpretation of possible effects of atomic-scale structure of the melts on the rate of crystal nucleation.