In 1897, following a number of crystallization studies, Ostwald suggested that if a system has a number of metastable states open to it then it proceeds sequentially by means of the nearest thermodynamic state and so on. This conjecture, which is known as Ostwald’s Rule of Stages (ORS), has gained traction with many researchers in many different fields. However, there has been no general proof of its validity. In this paper the specific case of batch solution crystallization of a polymorphic compound using classical nucleation and growth theory is examined in some detail and the ORS conjecture is found wanting. Classical nucleation theory presents not a sequential process, as in ORS, but a parallel process of multiphase nucleation and growth development of all metastable phases. What is observed is not the appearance of metastable phases sequentially but the disappearance of phases sequentially in the order of their metastability. If the crystallization kinetics of the metastable phases are suitably differentiated and advantageous the system might appear to be sequential under suitable conditions. This analysis does not exclude ORS from being applicable in other transformation cases. For instance, it is possible that systems where the subsequent phase changes are governed by solid-state transformations then the sequential model proposed by Ostwald might be appropriate but whether this covers all available metastable states is an open question.
Read full abstract