Crystalline nucleation in undercooled liquids: A Bayesian data-analysis approach for a nonhomogeneous Poisson process

摘要

A Bayesian data-analysis approach to data sets of maximum undercooling temperatures recorded in repeatednmelting-cooling cycles of high-purity samples is proposed. The crystallization phenomenon is described in termsnof a nonhomogeneous Poisson process driven by a temperature-dependent sample nucleation rate J (T ). Thenmethod was extensively tested by computer simulations and applied to real data for undercooled liquid Ge. Itnproved to be particularly useful in the case of scarce data sets where the usage of binned data would degrade thenavailable experimental information.