(727f) Crystal Regeneration Post-Breakage: Solvent Effects, Multiple Cleavage Sites, and Scale-up Towards Industrial Applications

Crystallisation is acknowledged as the earliest chemical engineering unit operation, and many aspects of the process (e.g., nucleation, crystal growth, particle dissolution, breakage) have been well-studied and relevant mathematical models have been developed [1]. However, understanding of crystals post-breakage has been elusive, and it is usually integrated into process models indifferent from the growth of unbroken crystals. Crystallisation product properties, including crystal shape and size distribution, have profound impact on the efficiency of downstream operations (e.g., processability, clogging issues) as well as the ultimate product qualities (e.g., transfer rate and bioavailability of pharmaceuticals) [1], [2]. Therefore, accurately integrating post-breakage crystal growth into process models is crucial in estimating and optimising commercial processes. Among the breakage mechanisms, impact due to particle-stirrer collisions have been reported to be the predominant breakage mechanism, as compared to particle-particle and particle-wall collusions [3]. Breakage incidence has been reported to be higher in higher agitation rates [4], [5], larger particle sizes [4], [6], and higher magma densities [5]. Moreover, studies have shown that, for crystals featuring a plane direction with the lowest attachment energy that is not featured on the external shape, termed as the internal cleavage plane (ICP), breakage exposing a particle’s ICP is the most prominent breakage direction [7], [8].

In a recent study on paracetamol (PCM) crystallisation in ethanol demonstrated that, when crystals are cleaved to expose their ICP (010), they predominantly grew from their broken side and attained their original pre-breakage shape. This novel finding on post-breakage crystal growth, named as ‘crystal regeneration’, elucidated the particle growth behaviour and opened the door for industrial advancements. Regeneration behaviour was confirmed for a variety of crystal sizes. In addition to evolving into their equilibrium shape, regeneration also featured a higher growth rate compared to the case of unbroken crystal. In the regeneration period, the growth perpendicular to (010) was 2-3 times faster than the growth parallel to it, whereas the growth rate ratio of the dimensions followed unity after the completion of regenerative growth. Also, regeneration was reported to occur singularly from breakages exposing the (010) facet [9]. Building on the initial findings, the current work examines regeneration in different solvents and for a crystal with multiple breakage planes. Moreover, industrial relevance of regeneration is explored via quantifying the translation of regeneration into mass-based growth rates, and monitoring the phenomenon in upscaled multi-crystal settings.

To describe the experimental methodology, initially, PCM crystals were grown to about 3-5 mm size, isolated, and cleaved with a scalpel to expose their (010) facet (Figure 1a). An evaporative crystallisation setup was utilised to explore different aspects (i.e., different solvents, multiple cleavage planes) of regeneration. Single-crystal regrowth experiments were conducted via evaporative crystallisation of broken PCM crystals in 50 ml crystallisation dishes, initiated with saturated solutions. Partially enclosing the dishes enabled slow evaporation rate, thus, preventing nucleation via operating within the metastable zone (Figure 1b). The translation of regeneration into mass growth rates was comparatively studied by ‘two-crystal’ experiments involving broken and unbroken crystal pairs in the same crystallisation dish using the aforementioned setup, combined with periodic mass measurements.

Regarding image analysis, images of growing crystals were taken using an automatised setup consisting of an RS Pro camera attached to a programmable robot arm, both connected to a computer (Figure 1b). This setup allowed for remote and periodic imaging of multiple crystallisation experiments, simultaneously. The images were digitally analysed using ImageJ software to quantify the growth rates of crystals’ specific dimensions.

Regarding the results, first, PCM regeneration was demonstrated in THF and acetone as different solvents, showcasing that a solute’s regeneration capability is not restricted to a specific solvent (i.e., ethanol in the previously reported case). Reflecting the differences in their volatility, and accordingly, evaporation rates, regeneration took around 140h for ethanol, 50h for THF, and 16h for acetone. Therefore, growth rates were compared for the initial half of the regeneration, latter half of the regeneration, and post-regeneration periods (Figure 2). Similar trends were observed across the three solvents. In the initial regeneration stages, the growth rate ratio of the dimension perpendicular to (010) to the dimension parallel to it, was 1.7-2.0 times higher than the growth ratio after the regeneration. The differences in growth ratio diminished through the later regeneration stages, before the crystals proceeded with their post-regeneration growth.

Simultaneous regeneration of multiple surfaces was demonstrated by growth of PCM crystals with 2 parallel ICP fractures in ethanol, resulting in a regenerative growth rate for the dimension orthogonal to (010) 4.5 times faster than the parallel dimension (Figure 3). Comparing with the case with single fracture, the doubled growth rate ratio for the case with two fractures show that regenerating surfaces do not inhibit each other by competition. In contrast, similar to the growth of an unbroken crystal, regenerating surfaces (i.e., (010) in this case) grow by their dependence on the crystallisation environment, independent of their overall surface area.

Results of the two-crystal experiments gave insight into the translation of regeneration into mass-based growth rates. In the case of paracetamol crystallisation in acetone, in the regeneration period, broken crystals of several sizes grew 38 ± 2% faster than their unbroken pairs in the same crystallisation environment, whereas the crystals grew at the same mass growth rate after the completion of regeneration (Figure 4). The experiments with mass-based monitoring are currently expanded into other solvents, to test if the trends can be generalised. The increased mass-based growth rate would open the possibility of making use of the regeneration phenomenon to reduce crystallisation batch process times as an industrial enhancement. For instance, the 38% faster growth rate, combined with the findings that two-sided regeneration doubles the overall regeneration rate (vide supra), would reduce the batch time of paracetamol regeneration in acetone by 43% depending on the weight fraction of one-sided and two-sided regenerating seeds to start the operation (Figure 5), inspiring to explore industrial implications for process-scale economic and environmental improvements. To test such relevant industrial applications, a large-scale crystalliser with an in-situ image monitoring apparatus (e.g., μ-DISCO [10]), is currently in design and development phase.

Through this study, crystal regeneration was demonstrated in a variety of solvents, which resulted in similar trends of doubled growth rate of the regenerating direction compared to its orthogonal direction. Simultaneous regeneration of two parallel sides on a crystal resulted in a quadrupled growth rate for the regenerating direction overall, showcasing the possibility of multiple-facet regeneration in addition to indicating that regenerating facets do not inhibit each other. Regeneration’s translation into mass growth was quantified via monitoring a regenerating and an unbroken crystal in the same crystallisation dish, yielding 38 ± 2% higher mass growth rate for regeneration. Current research directions include studying crystal regeneration in multi-crystal systems utilising a large-scale crystalliser, to quantify the effects of breakage and regeneration in industrial settings. These findings elucidate novel phenomena in crystal breakage and enhance crystal size, shape, and growth rate predictions, with promising implications for industrial processes.

References

[1] W. Beckmann, Crystallization: Basic Concepts and Industrial Applications. Weinheim, Germany: Wiley-VCH Verlag, 2013.

[2] A. S. Myerson, D. Erdermir, and A. Y. Lee, Handbook of Industrial Crystallization, 3rd ed. Cambridge University Press, 2019.

[3] B. Szilágyi and B. G. Lakatos, ‘Model-based analysis of stirred cooling crystallizer of high aspect ratio crystals with linear and nonlinear breakage’, Comput Chem Eng, vol. 98, pp. 180–196, 2017, doi: 10.1016/j.compchemeng.2016.11.028.

[4] Y. Bao, J. Zhang, Q. Yin, and J. Wang, ‘Determination of growth and breakage kinetics of l-threonine crystals’, J Cryst Growth, vol. 289, no. 1, pp. 317–323, 2006, doi: 10.1016/j.jcrysgro.2005.11.001.

[5] S. M. Reeves and P. J. Hill, ‘Mechanisms Influencing Crystal Breakage Experiments in Stirred Vessels’, Cryst Growth Des, vol. 12, pp. 2748–2758, 2012.

[6] A. D. Randolph, ‘Effect Of Crystal Breakage On Crystal Size Distribution In A Mixed Suspension Crystallizer’, I&EC Fundamentals, vol. 8, no. 1, pp. 58–63, 1969.

[7] J. Y. Y. Heng, A. Bismarck, A. F. Lee, K. Wilson, and D. R. Williams, ‘Anisotropic surface energetics and wettability of macroscopic form I paracetamol crystals’, Langmuir, vol. 22, no. 6, pp. 2760–2769, 2006, doi: 10.1021/la0532407.

[8] C. C. Sun and Y. H. Kiang, ‘On the Identification of Slip Planes in Organic Crystals Based on Attachment Energy Calculation’, J Pharm Sci, vol. 97, no. 8, pp. 3456–3461, Aug. 2008, doi: 10.1002/JPS.21234.

[9] I. Bade, V. Verma, I. Rosbottom, and J. Y. Y. Heng, ‘Crystal regeneration - a unique growth phenomenon observed in organic crystals post breakage’, Mater Horiz, vol. 10, pp. 1425–1430, 2023, doi: 10.1039/d2mh01180h.

[10] A. K. Rajagopalan et al., ‘A comprehensive shape analysis pipeline for stereoscopic measurements of particulate populations in suspension’, Powder Technol, vol. 321, pp. 479–493, 2017, doi: 10.1016/j.powtec.2017.08.044.