461527 The Effects of Crystal Size and Habit on the Dehydration Induced Polymorphism; The Case of Carbamazepine Dihydrate

Friday, November 18, 2016: 1:50 PM
Continental 5 (Hilton San Francisco Union Square)
Eftychios Hadjittofis, Chemical Engineering Department, Imperial College London, London, United Kingdom, Geoff G. Z. Zhang, Global Pharmaceutical R&D, AbbVie Inc., North Chicago, IL and Jerry Heng, Department of Chemical Engineering, Imperial College London, London, United Kingdom

The process conditions influence the kinetics and thermodynamics during drying of hydrated materials. The water removal from hydrated crystals induces mechanical stress on the crystal lattice leading to structural deformations. Concurrently, the nucleation of the anhydrous form alters material's properties1. Thus, material properties influence dehydration, but also they are influenced by it. This work examines the role of interfacial phenomena of hydrated materials, using carbamazepine dihydrate as model compound.

Three methods were used for the crystallization of carbamazepine dihydrate:

  1. 0.075 g*mL-1 of anhydrous carbamazepine (as received) were dissolved under heating in a methanol-water (60:40 %v/v) mixture. Then the mixture was left to cool down for 12 hours under stirring (500 rpm). The solution was filtered out and the crystals were collected2.
  2. 0.05 g*mL-1 of anhydrous carbamazepine (as received) were dissolved under heating in a methanol-water (50:50 %v/v) mixture. Then the mixture was left to cool down for 12 hours under stirring (500 rpm). The solution was filtered out and the crystals were collected.
  3. 0.025 g*mL-1 of anhydrous carbamazepine (as received) were dissolved under heating in a methanol-water (40:60 %v/v) mixture. Then the mixture was left to cool down for 12 hours under stirring (500 rpm). The solution was filtered out and the crystals were collected.
X-Ray powder diffraction (XRPD) patterns agreed that all methods result to carbamazepine dihydrate crystals, however optical examination of the crystals, using Scanning Electron Microscope (SEM), revealed differences in their crystal size and habits. Crystals from methods one are smaller in size and they have a smaller aspect ratio, leading to a more compact shape and better flowability and cohesion behaviour, compared with the crystals obtained from the other two methods. It is important, for later on, to mention that the dominant face in the crystals prepared via methods two and three facet (0 2 0) is by far the dominant one.

The three types of carbamazepine dihydrate were exposed in dehydration conditions of 30, 40, 60 and 80oC for 24 hours and the anhydrous material was examined with XRPD and SEM. The material from methods two and three dehydrated to the Triclinic polymorph, the less stable carbamazepine polymorph3. The crystals obtained from method one dehydrated concomitantly to the P-Monoclinic and Triclinic polymorphs. Increasing temperature leads to increasing amount of the Triclinic polymorph, as shown from the SEM images. The domination of the Triclinic polymorph, in the high aspect ratio crystals is associated with the formation of well-ordered macroscopic cracks perpendicular to the (0 2 0) facet of the crystal, followed by the growth of Triclinic carbamazepine whiskers. The high aspect ratio of the crystals obtained from methods two and three appears to be quite attractive for crack nucleation and propagation. Triclinic carbamazepine whiskers grow in needle-shaped structures with high surface area to volume ratio to enhance the dissipation of heat from the crystals. In the, smaller and more compact, particles obtained from method one smaller cracks, formed randomly on the crystal facets, seem to be sufficient to dissipate the heat from the system without the need of the whiskers. The smaller size of the particles, probably, contributes to better heat dissipation. From the above, it can be concluded that energy dissipation drives dehydration induced polymorphism in this system.

Material from method one was placed in a glass vial with six milling balls (5 mm diameter)4-6left to roll for 12 hours. XRPD examination of the material obtained did not reveal dehydration of the crystals, although the SEM images showed that defects were formed on the crystals. The grinded material was subjected to the aforementioned drying conditions, however the trend in the dehydration-induced polymorphism did not change dramatically, the dried material was again a mixture of the two polymorphs, P-Monoclinic and Triclinic. This highlights the importance of powder’s bulk properties (size and habit) over surface properties in the determination of drying induced polymorphism. The effects of stirring were also examined in the context of method one. It was shown that low Reynolds number favors the formation of high aspect ratio dihydrates which give a Triclinic polymorph upon dehydration. Using the same liquid composition as in method one the effect of supersaturation on crystal habit were examined. It was shown that low supersaturation, also, leads to the formation of high aspect ratio dihydrates.

FD-IGC measurements were performed on samples from method 1 dried at the four different temperatures. The shift from lower to higher surface energies is correlated with the increasing amount of Triclinic material observed with SEM. Computational models and FD-IGC measurements on pure P-Monoclinic and Triclinic can be combined to give approximate values of the polymorphic content of each sample.

This work, establishes pathways to obtain pure Triclinic carbamazepine and P-Monoclinic carbamazepine via the dehydration carbamazepine dehydrate. It critically revisits previous studies suggesting that the dehydration follows the Ostwald2 rule of stages. A mechanism for the dehydration-induced polymorphism, suggesting a concomitant nucleation scheme is proposed. According to this mechanism crystal size and habit determine the dehydration induced polymorphism. The mechanism is driven by the energy dissipation mechanism of the system. This drying mechanism can be mechanistically understood using notions similar to those employed by Machon to describe the influence of defect concentration of nanoparticle’s phase transitions using the one dimensional Ginzburg-Landau equation7,8. It is, also, demonstrated that FD-IGC measurements can be used to distinguish the polymorphic content of the dried material.


Imperial College London and AbbVie jointly participated in study design, research, data collection, analysis and interpretation of data, writing, reviewing, and approving the publication. Eftychios Hadjittofis is a graduate student at Imperial College London; Jerry Y. Y. Heng is a professor at Imperial College London. They all have no additional conflicts of interest to report. Geoff G. Z. Zhang is an employee of AbbVie and may own AbbVie stock.


1. Galwey AK. The topochemistry of the reaction interface in decomposition processes in solids. Reactivity of Solids. 1990;8(3–4):211-230.

2. Khoo JY, Shah UV, Schaepertoens M, Williams DR, Heng JYY. Process-induced phase transformation of carbamazepine dihydrate to its polymorphic anhydrates. Powder Technol. 2013;236:114-121.

3. Grzesiak AL, Lang M, Kibum K, Matzger AJ. Comparison of the four anhydrous polymorphs of carbamazepine and the crystal structure of form I. Journal of Pharmaceutical Sciences. 2003;92(11):2260-2271.

4. Wang Z, Shah UV, Olusanmi D, et al. Measuring the sticking of mefenamic acid powders on stainless steel surface. Int J Pharm. 2015;496(2):407-413.

5. Shah UV, Olusanmi D, Narang AS, et al. Effect of crystal habits on the surface energy and cohesion of crystalline powders. Int J Pharm. 2014;472(1–2):140-147.

6. Shah UV, Olusanmi D, Narang AS, Hussain MA, Hinder SJ, Heng JYY. Decoupling the contribution of surface energy and surface area on the cohesion of pharmaceutical powders. Pharmaceutical Research. 2014;32:248-259.

7. Machon D, Melinion P. Size-dependent pressure-induced amorphization: A thermodynamic panorama. Physical Chemistry Chemical Physics. 2014;17(2):903.

8. Machon D, Piot L, Hapiuk D, et al. Thermodynamics of nanoparticles: Experimental protocol based on a comprehensive ginzburg-landau interpretation. Nano Letters. 2014;14(1):269-276.

Extended Abstract: File Not Uploaded