A model-based approach to unravel the driving forces in attrition-enhanced deracemization processess.

Molecular chirality is an important geometrical property with far-reaching implications in the pharmaceutical, food and agrochemical industries. Chiral molecules can exist in two non-superimposable mirror-image forms called enantiomers, which often possess substantially different biological effects. Since most chemical synthetic processes of chiral molecules result in an equimolar mixture of both enantiomers, referred to as a racemic mixture, there is high demand for developing downstream separation processes for enantiomers. Solid-state deracemization processes are a new class of crystallization-based methods that achieve such task by completely converting a racemic suspension of conglomerate crystals into a single chirality, simply by applying mechanical (Viedma ripening) or thermal treatments (temperature cycling-enhanced deracemization) to the crystals in contact with a racemizing solution. Compared to other methods applied for chiral resolution, they are advantageous, as they operate without the need for expensive chiral auxiliaries or extensive kinetic and thermodynamic information. However, they suffer from several limitations, most notably long processing times, limited applicability and challenging scalability. On top of that, the interplay between the mechanisms underlying the phenomena is not sufficiently understood, which hinders their effective application.

The aim of this doctoral project is to address these limitations by using various experimental and theoretical approaches. With regards to enhancing the rates of these processes and improving their productivity, ultrasound-enhanced grinding and microwave heating are used to intensify the particle breakage mechanism in Viedma ripening and the heating/cooling steps in temperature cycling-enhanced deracemization, respectively. The experimental results reveal that ultrasound-enhanced grinding combined with seeding can effectively replace the conventional bead grinding in Viedma ripening. Furthermore, a comparison between different types of grinding in deracemization processes sheds significant mechanistic insight into the role of abrasion and fracture of crystalline particles in attrition-enhanced deracemization. Similarly, microwave heating coupled with rapid cooling, allows accessing novel operating windows for temperature cycling, with an order of magnitude higher heating/cooling rates than conventionally applied. This results not only in faster deracemization due to a reduction in the cycle duration, but also in higher selectivity due to the minimization of side reactions that lower productivity. 

Solid-state deracemization methods are only applicable when the enantiomers crystallize as a racemic conglomerate, in which each enantiomer forms individual crystals. However, the majority of racemic mixtures (~90%) form a racemic compound, in which single crystals contain both enantiomers. Consequently, extending deracemization to those compounds is highly desired. By exploiting conditions (temperature, solvent) where the (solvated) racemic compound becomes metastable with respect to the conglomerate, a new process that combines a solvent-mediated crystal transformation of the racemic compound to conglomerate with deracemization is designed and exemplified. The process is then applied to deracemize two chiral model compounds and the results show that it allows for much faster rate compared to the conventional deracemization of conglomerates, to such extent that grinding can be minimized. Besides, the new process extends deracemization to the wider pool of compounds that crystallize as racemic compounds and may undergo such transformations.

Finally, in view of better understanding the mechanisms involved in deracemization phenomena, a simplified kinetic model is developed and validated to account for the physicochemical phenomena thought to occur in Viedma ripening, namely crystal growth/dissolution subject to the size-dependence of solubility, racemization, breakage and enantioselective agglomeration. The model results explain recent experimental observations and confirm that the aforementioned mechanisms combined could explain the deracemization process. Subsequently, the model is extended to account for the combined crystal transformation/deracemization process. For the latter case, it is shown that breakage and agglomeration are not essential ingredients, provided that the racemization reaction converts the counter enantiomer molecules faster than they nucleate, leading to a more direct conversion route for the counter enantiomer. The results presented in this work help increase the understanding of deracemization phenomena and may pave the way towards their industrial application.