On classical and non-classical views on nucleation

Classical nucleation theory (CNT) is based on the notion of critical nuclei serving as transition states between supersaturated solutions and growing particles. Their excess standard free energy depends on supersaturation, and determines the height of the barrier for phase separation. However, predictions of CNT nucleation rates can deviate from experimental observations by many orders of magnitude. We argue that this is due to oversimplifications within CNT, rendering the critical nucleus essentially a conceptual notion, rather than a truly existing physical entity. Still, given adequate parametrization, CNT is useful for predicting and explaining nucleation phenomena, since it is currently the only quantitative framework at hand. In the recent years, we have been introducing an alternative theory, the so-called pre-nucleation cluster (PNC) pathway. The truly "non-classical" aspect of the PNC pathway is the realization that critical nuclei, as defined within CNT, are not the key to nucleation, but that the transition state relevant for phase separation is based on a change in dynamics of PNCs rather than their size. We provide a summary of CNT and the PNC pathway, thereby highlighting this major difference. The discussion of recent works claiming to provide scientific evidence against the existence of PNCs reveals that such claims are indeed void. Moreover, we illustrate that an erroneous interpretation of the concentration dependence of the free energy has led to a postulated rationalization of the standard free energy of ion pairs and stable ion associates within CNT, which is not sustainable. In fact, stable ion associates are stuck in a free energy trap from the viewpoint of CNT and cannot be considered in a straightforward manner. On the other hand, the notions of the PNC pathway, by dismissing the idea of the CNT-type critical nucleus as a required transition state, overcome this issue. While a quantitative theory of the PNC pathway is eagerly anticipated, the rationalization of experimental observations that are inconsistent with CNT proves its qualitative explanatory power, underpinning great promise towards a better understanding of, for instance, polymorph selection and crystallization control by additives.