Scholarly article on topic 'Template-induced polymorphic selectivity: the effects of surface chemistry and solute concentration on carbamazepine crystallisation'

Template-induced polymorphic selectivity: the effects of surface chemistry and solute concentration on carbamazepine crystallisation Academic research paper on "Chemical engineering"

Share paper
Academic journal
OECD Field of science

Academic research paper on topic "Template-induced polymorphic selectivity: the effects of surface chemistry and solute concentration on carbamazepine crystallisation"

Volume 16 Number 23 21 June 2014 Pages 4895-5166




Jerry Y. Y. Heng et al.

Template-induced polymorphic selectivity: the effects of surface chemistry and solute concentration on carbamazepine crystallisation



View Article Online

View Journal | View Issue

Cite this: CrystEngComm, 2014, 16, 4927

Received 24th December 2013, Accepted 4th March 2014

DOI: 10.1039/c3ce42622j

Template-induced polymorphic selectivity: the effects of surface chemistry and solute concentration on carbamazepine crystallisation!

Jose V. Parambil,ab Sendhil K. Poornachary,c Reginald B. H. Tan and Jerry Y. Y. Heng*a

Cooling crystallisation of carbamazepine solution in glass vials functionalised with different silane molecules resulted in preferential nucleation of metastable form II or stable form III polymorphs within a definite range of supersaturation. In contrast, the two crystal forms nucleated concomitantly on a control substrate under similar solution conditions.

Polymorph control in pharmaceutical crystallisation is mainly ensured by seeding with the required crystal form and ensuring strict control of solution supersaturation. Seeding not only helps in reducing the crystallisation batch time through secondary nucleation but also provides an additional control over particle size distribution.1 Nevertheless, the properties of seed crystals, such as surface area, roughness and defects, have to be consistent throughout in order to minimise batch-to-batch variation in product quality. Heterogeneous nucleation is thermodynamically favoured over homogeneous nucleation since the high surface energy component required to produce a stable nucleus is reduced by the contribution from a heterogeneous nucleating surface. As a result of the reduced surface energy component associated with heterogeneous nucleation, most nucleation events can be envisaged to occur via heterogeneous nucleation following a two stage nucleation pathway.2 One of the motivations for developing a template-induced crystallisation process using tailor-made surfaces is to improve upon reproducibility with respect to the crystal form obtained. Epitaxial relationship between a template surface and nucleating crystal has been

a Surfaces and Particle Engineering Laboratory, Department of Chemical Engineering Imperial College London, South Kensington Campus, London SW7 2AZ, UK. E-mail:; Tel: +44 (0)20 7594 0784 b Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117585

c Institute of Chemical and Engineering Sciences, A*STAR (Agency for Science, Technology and Research), 1 Pesek Road, Jurong Island, Singapore 627833 ! Electronic supplementary information (ESI) available: Details on silanisation method, silane structure, advancing contact angle measurements of four probe liquids on template surface, ellipsometry and PXRD data are provided in ESI. See DOI: 10.1039/c3ce42622j

singled out to be the most critical factor in polymorph selective crystallisation by heterogeneous nucleation.3 However, for certain systems, it has been reported that the influence of chemical interactions at the heterogeneous nucleant interface may dominate in directing polymorphic nucleation.4

Template substrates used for inducing crystallisation of a specific polymorph range from crystalline inorganic surfaces,5 single crystals or surfaces adsorbed with structurally-related molecule or a different crystal form of the same molecule,6 self-assembled monolayers (SAMs) of thiols7 and chiral molecules,8 polymer9 and organosilane substrates.4 Besides molecular functionality, porosity of the template substrate has been reported to influence crystallisation behaviour of both small molecules and proteins.10 However, the effect of solute concentration on template-mediated nucleation is often overlooked and no studies have been reported to date.

In this study, glass vials were chemically modified via silanisation procedure reported by Shah et al.11 The silanised surfaces were analysed using contact angle measurements and ellipsometry (see ESIf). Cooling crystallisation of carba-mazepine (CBZ), an anticonvulsant and mood-stabilising drug, from ethanol solutions of different initial concentrations was carried out in the silanised glass vials. CBZ was selected as the model compound in these studies because of its well characterised polymorphic forms12 and crystallisation behaviour from common organic solvents.13 Moreover, recently a new polymorphic form of CBZ was reported to form on the crystal facet of an analogous molecule directed by epitaxial relationship.6

Surface energy of the modified glass substrates were calculated using Owens-Wendt method,14 and the polar and dispersive components of surface energy values are reported in Table 1. It can be seen that the cyano-functionalised surface is the most polar and also exhibits highest surface energy amongst the three functionalised templates. The thickness of the silane layer on a silicon wafer substrate as determined by ellipsometryj was found to be 1.37 ± 0.03 nm. Considering that the linear length of a stretched out silane molecule equals

This journal is © The Royal Society of Chemistry 2014

CrystEngComm, 2014, 16, 4927-4930 | 4927



Table 1 Surface energy components for functionalised templates calculated using the Owens-Wendt method"

Cyano Mercapto Fluoro

Head group (-CN) (-SH) (-CF3)

Yd (mJ m-2) 38.0 37.6 19.2

Yp (mJ m_2) 14.2 5.0 2.5

Y (mJ m_2) 52.3 42.6 21.7

a Yd, and Y stand for dispersive, polar and total surface energy of the template surface respectively.

to ca. 7 A, the silane layer thickness should correspond to 2-3 layers of silane molecules depending on its molecular orientation on the surface.

The crystallisation experiments were conducted at 20 °C with 3 mL solution with initial solution concentrations in the range of 45-65 mg ml-1 in glass vials placed inside a temperature controlled incubator. Induction time measurements correspond to the time for observation of a first crystal (<10 ^m) nucleating from the supersaturated solution using a digital handheld optical microscope (VMS-004D, Veho U.K.). Induction time measurements can be affected by sample volume and experimental setup.15 Hence, precautions were taken to ensure identical conditions between control and silanised templates with surface chemistry being the only variable between vials. CBZ crystal polymorphs were distinguished from its growth morphology, with needle shaped form II and tetrahedral form III crystals. The polymorphic forms were further confirmed through powder X-ray diffraction (see ESIf). Throughout this study, at least 10 trials of crystallisation experiments were carried out for each solute concentration, with each trial comprising 3 replicates of the same template surface. Control experiments were performed in clean glass vials at similar solution conditions.

Results of CBZ crystallisation conducted in glass vials from an initial concentration of 60 mg ml-1 are shown in Fig. 1. In

the control glass vials, both form II and form III polymorphs crystallised concomitantly, eventually transforming into stable form III crystals in about 24 h. Crystallisation in mercapto and fluorosilanised glass vials nucleated more of form III crystals and very few form II crystals compared to control vials. In contrast, the cyano silanised vials predominantly resulted in nucleation of metastable form II crystals. By monitoring the images captured at different time intervals, it was revealed that the metastable crystals formed on cyano functionalised vials did not transform to the stable form even after 24 h. This could be reasoned due to desupersaturation of the solution primarily through nucleation of form II crystals. Consequently, form II crystals are prevented from undergoing polymorphic transformation in the absence of any stable form III crystal nuclei. On the other hand, on introducing form III seed crystals into the system, form II polymorph completely transformed to the stable form within 24 h.

The induction periods for nucleation of CBZ crystals at various solute concentrations and under the influence of different template surface chemistries are shown in Fig. 2. With the increase in solute concentration from 50 mg ml-1 to 60 mg ml-1, the induction period decreased from 172 ± 24 min to 115 ± 17 min. This data qualitatively agrees with the classical nucleation relationship between induction period and solution supersaturation. For an initial solute concentration of 60 mg ml-1, the observed induction period was ~2 h in control glass vials as compared to ~45 min in each of the silanised glass vials. Likewise, at 55 mg ml-1 and 50 mg ml-1 solute concentrations, the induction period was shortened by 2-3 times, due to the influence of templates. The effect of template surfaces on the number of crystals formed was also pronounced. While a large number of either form II or form III crystals were observed in functionalised glass vials, only a few big crystals of form III together with form II needles were observed on the control surfaces (cf. Fig. 1). While these kinetic data clearly shows that the silanised surfaces in general enhanced heterogeneous nucleation of CBZ crystals, no specific effect of the template surface chemistry on the induction times were observed.

Furthermore, template-mediated nucleation of CBZ crystal polymorphs was found to be influenced by the initial solute

Fig. 1 Carbamazepine crystallised from 60 mg ml-1 solution in glass vials silanised with different functional groups: i) control; ii) mercapto; iii) fluoro; iv) cyano. The crystals were photographed after 24 h from the start of crystallisation. Scale bar represent 200 ^m length.

Initial concentration (mgfml)

Fig. 2 Induction period for nucleation of carbamazepine crystals in silanised glass vials at various solute concentrations.

4928 | CrystEngComm, 2014, 16, 4927-4930

This journal is © The Royal Society of Chemistry 2014



c (mg/ml) 45

Fig. 3 The effects of surface chemistry and initial supersaturation (c/c*, calculated based on the solubility of form III) on crystallisation of carbamazepine polymorphs from ethanol at 20 °C. Initial solute concentration (mg ml-1) is also shown on the horizontal axis for reference.

concentration (Fig. 3). At 55 mg ml-1, the polymorphic forms crystallised in the control and silanised glass vials respectively were akin to that observed at 60 mg ml-1. At 50 mg ml-1, nucleation in the control vessel was dominated by form III crystals, along with a few form II polymorph crystallised con-comitantly. On mercapto and fluoro-functionalised surfaces, form III nucleation was clearly dominant over form II at this solute concentration. In contrast, cyano-functionalised surfaces produced form II crystals initially followed by nucle-ation and growth of form III. Consequently, the form II crystals underwent solution-mediated transformation to the stable form III within 24 h. Upon further reducing the initial solute concentration to 45 mg ml-1, no apparent differences were observed in the polymorphic forms crystallised on the control and functionalised surfaces. Above 60 mg ml-1, CBZ crystallisation was dominated by form II nucleation on both the control and functionalised surfaces. Thereafter the form II crystals remained stable for more than 24 h without undergoing polymorphic transformation.

Preferential nucleation of polymorphs on templates with different chemistry could be attributed to specific molecular interaction at template interface. Due to the multilayer thickness of silane layer, an epitaxial relationship between crystal facet and template surface appear highly improbable. However, facet specific electrostatic interactions involving dipole moments and hydrogen bonding potentials emerge as the most probable candidates that can result in polymorph selective nucleation. The polarity of the template surface can also play a role in resulting in a preferential orientation of crystals as reported for polymer heteronuclei.16 Also, polarity of the solvents used in CBZ crystallisation has been established to affect the structure of prenucleation aggregates.17 Nevertheless, a direct correlation between nucleant surface polarity and facet specific crystal properties is not identified.

From the experimental observations, it is conclusive that the template surface chemistry had effectively mediated preferential nucleation of CBZ crystal polymorphs from ethanol solutions only within a definite solute concentration range of 50-60 mg ml-1. This corresponds to a supersaturation (c/c*)

window of 2.1-2.6. Outside this solute concentration range, the kinetics of homogeneous and/or heterogeneous primary nucleation was dominant over template-mediated nucleation resulting in crystallisation of the stable and metastable poly-morphs at the lower and higher supersaturations respectively. As the functional groups on silanised templates differ in polar and non-polar surface energies, it is likely that intermolecular interactions with nucleating cluster had potentially influenced the template-mediated nucleation of CBZ crystal forms. However, an epitaxial relationship between the template substrate and crystal polymorph seems elusive. As a logical extension of this work, chemical interactions between the template surface and various facets of CBZ crystal poly-morph(s) are currently being analysed to gain further insights into the template-induced nucleation process. Building on this study, a design space for selective crystallisation of CBZ polymorphs is being developed by incorporating surface chemistry as a process parameter in the concentration-temperature phase diagram.


The authors acknowledge Dr. Stephane Kena-Cohen and Tarik Abdelmoula, Department of Physics, Imperial College London, for help with ellipsometry.

Notes and references

J Ellipsometric data was fitted using combined Cauchy layer model with a refractive index of 1.45 for the silane and SiO2 layers on Si wafer surface. The thickness of the silane layer was calculated by subtracting the SiO2 layer thickness from the total Cauchy layer thickness.

1 S. L. Morissette, O. Almarsson, M. L. Peterson, J. F. Remenar, M. J. Read, A. V. Lemmo, S. Ellis, M. J. Cima and

C. R. Gardner, Adv. Drug Delivery Rev., 2004, 56, 275-300.

2 D. Erdemir, A. Y. Lee and A. S. Myerson, Acc. Chem. Res., 2009, 42, 621-629.

3 K. Chadwick, A. Myerson and B. Trout, CrystEngComm, 2011, 13, 6625; R. Hiremath, J. A. Basile, S. W. Varney and J. A. Swift, J. Am. Chem. Soc., 2005, 127, 18321-18327; A. Y. Lee, A. Ulman and A. S. Myerson, Langmuir, 2002, 18, 5886-5898.

4 J. Zhang, A. Liu, Y. Han, Y. Ren, J. Gong, W. Li and J. Wang, Cryst. Growth Des., 2011, 11, 5498-5506; T. Delmas, U. V. Shah, M. M. Roberts, D. R. Williams and J. Y. Y. Heng, Powder Technol., 2013, 236, 24-29; B. K. Olmsted and M. D. Ward, CrystEngComm, 2011, 13, 1070.

5 L. Ruan, H. Ramezani-Dakhel, C. Y. Chiu, E. Zhu, Y. Li, H. Heinz and Y. Huang, Nano Lett., 2013, 13, 840-846; O. Werzer, B. Kunert, E. Roblegg, A. Zimmer, M. Oehzelt and R. Resel, Cryst. Growth Des., 2013, 13, 1322-1328.

6 J. Urbanus, J. Laven, C. P. M. Roelands, J. H. ter Horst,

D. Verdoes and P. J. Jansens, Cryst. Growth Des., 2009, 9, 2762; J. Tao, K. J. Jones and L. Yu, Cryst. Growth Des.,

This journal is © The Royal Society of Chemistry 2014

CrystEngComm, 2014, 16, 4927-4930 | 4929



2007, 7, 2410; J. B. Arlin, L. S. Price, S. L. Price and A. J. Florence, Chem. Commun., 2011, 47, 7074-7076.

7 A. Kwokal, T. T. H. Nguyen and K. J. Roberts, Cryst. Growth Des., 2009, 9, 4324; R. Hiremath, S. W. Varney and J. A. Swift, Chem. Commun., 2004, 2676-2677.

8 D. H. Dressler, I. Hod and Y. Mastai, J. Cryst. Growth,

2008, 310, 1718-1724.

9 M. D. Lang, A. L. Grzesiak and A. J. Matzger, J. Am. Chem. Soc., 2002, 124, 14834-14835.

10 U. V. Shah, D. R. Williams and J. Y. Y. Heng, Cryst. Growth Des., 2012, 12, 1362-1369.

11 U. V. Shah, M. C. Allenby, D. R. Williams and J. Y. Y. Heng, Cryst. Growth Des., 2012, 12, 1772-1777.

12 A. L. Grzesiak, M. Lang, K. Kim and A. J. Matzger, J. Pharm. Sci., 2003, 92, 2260-2271.

13 C. A. Hunter, J. F. McCabe and A. Spitaleri, CrystEngComm, 2012, 14, 7115.

14 D. K. Owens and R. C. Wendt, J. Appl. Polym. Sci., 1969, 13, 1741-1747.

15 R. J. Davey, S. L. Schroeder and J. H. ter Horst, Angew. Chem., Int. Ed., 2013, 52, 2166-2179.

16 E. Curcio, V. Lopez-Mejias, G. Di Profio, E. Fontananova, E. Drioli, B. L. Trout and A. S. Myerson, Cryst. Growth Des., 2014, 14, 678-686.

17 C. A. Hunter, J. F. McCabe and A. Spitaleri, CrystEngComm, 2012, 14, 7115-7117.

4930 | CrystEngComm, 2014, 16, 4927-4930

This journal is © The Royal Society of Chemistry 2014