Advancing Approaches in Detecting Polymorphism

September 1, 2007
Patricia Van Arnum
Pharmaceutical Technology

Volume 2007 Supplement, Issue 4

Identifying polymorphs is a crucial part of the drug-development process as researchers forward select methods to improve detection.

Screening and detecting polymorphism when developing and manufacturing active pharmaceutical ingredients (APIs) is an ongoing challenge for pharmaceutical manufacturers. Polymorphism is the ability of a compound to exist in more than one crystalline structure. Different solid forms can possess different properties, including solubility, which can in turn affect the bioavailability of the drug.

Polymorph screening and detection

"The investigation of polymorphic forms of a drug substance is valuable to the pharmaceutical industry, as the differences in the physical properties of different solid forms can be problematic in the later stages of development and formulation," says Chris Frampton, chief scientific officer at SAFC-Pharmorphix (Cambridge, UK). "Different solid forms can show differences in properties such as aqueous thermodynamic solubility, melting point and hygroscopicity, which in turn can affect bioavailability, processability, and stability, respectively."

One of the more well-chronicled examples of polymorphism occurred in ritonavir, the API in "Norvir," a protease inhibitor developed by A bbott Laboratories (Abbott Park, IL). The drug was approved in 1996, and in mid-1998, Abbott encountered manufacturing difficulties with the capsule formulation, according to the company's 1998 annual report. Ritonavir exhibited conformational polymorphism of two unique crystal lattices that had significantly different solubility properties (1). The formation of the polymorph caused Abbott to pull the drug from the market and reformulate.

A recent analysis by SSCI, the solid-state chemistry business of Aptuit (Greenwich, CT), showed that of 245 compounds it has screened, 89% had multiple solid forms (see Figure 1). Approximately 50% of the compounds showed polymorphism, 37% were hydrates, and 31 were solvates, outlines Stephen Byrn, head of the scientific advisory board of Aptuit and head of the Department of Industrial and Physical Pharmacy at Purdue University.

Figure 1

Other research conducted by Professor Ulrich Griesser at the University of Innsbruck shows a lower incidence of multiple solid forms in organic molecules: polymorphism (36%), hydrates (28%), and solvates (10%) (2).

"The difference in prevalence of 51% versus 36% probably reflects the fact that many organic chemists do not search for polymorphs when preparing new compounds," explains Byrn. "In contrast, compounds submitted to SSCI/Aptuit are drug candidates. Another factor that may explain the differences in prevalence of polymorphs is the fact that SSCI/Aptuit would tend to get problem compounds, including materials that have a large numbers of polymorphs."

Challenges in polymorph screening

A review of several APIs and new chemical entities (NCEs) shows not only the prevalence of other solid forms but also the challenges in detecting these forms.

Acetaminophen. The polymorphism of acetaminophen shows the challenges in screening for polymorphs (see Figure 2).

Figure 2

"The crystal structures for two known crystal forms of acetaminophen (monoclinic and orthorhombic) have been published," explains Frampton (3, 4). "The monoclinic form is the thermodynamically stable crystal form at room temperature with respect to the orthorhombic form. Evidence for a third, metastable, form has been published (5, 6). To date, however, it has not been possible to isolate any crystalline material to enable a full crystal structure or physicochemical property determination."

The problem with acetaminophen was the repeated failures of research groups to isolate the orthorhombic form using the method of slow evaporation from ethanol as described by Haisa et al. (3), explains Frampton. Eventually, it was discovered that the orthorhombic form could be produced from solution by using seeds of the orthorhombic modification that were isolated from melt-crystallized acetaminophen (7).

Figure 3

"It is clear that in this case, conventional solvent-screening techniques of the time would fail to produce any evidence of the less-stable orthorhombic form since it was only possible to produce seeds of the orthorhombic form under somewhat extreme conditions," says Frampton. "Given that the crystal structures of the two modifications were available, it was reasonably straightforward to identify the pure forms from their X-ray powder diffraction (XRPD) patterns (see Figures 3a and 3b). The form assignment from XRPD was subsequently confirmed by single-crystal X-ray analysis.

Sodium diclofenac. Sodium diclofenac (see Figure 4) is a nonsteroidal anti-inflammatory drug that shows polymorphism. The sodium salt of the drug is manufactured by Novartis as "Voltaren." Frampton says there are three known anhydrous crystal modifications of the free base: two monoclinic forms (C2/c, P21/c), and an orthorhombic form (Pcan) (8–10).

Figure 4

"Sodium diclofenac has also been shown to readily form stable hydrate phases when crystallized from aqueous miscible solvents," explains Frampton (11). "Although these hydrated materials are not technically polymorphic phases since they are solvates, they do present interesting solid-form challenges."

He explains that the current literature reports a sodium diclofenac pentahydrate phase that was derived from the results of a single crystal X-ray diffraction study (11). The pentahydrate phase was based on the crystalline asymmetric unit of the structural model containing two molecules of sodium diclofenac and 10 molecules of water in the monoclinic space group P21/m. The results of the structural investigation, however, showed that the solved structure had a disordered sodium-water linkage and a high overall crystallographic R factor of 7.06%.

A decision framework for on polymorphism for generic drug applications

"We were prompted to reinterpret the single-crystal X-ray data after repeated thermogravimetric analysis (TGA) and Karl Fischer (KF) measurements indicated that the water content was reproducibly always slightly less (e.g., 21.48% and 21.0% from TGA and KF, respectively) than the theoretical content of 22.07% based a fully hydrated pentahydrate structure," says Frampton. These values equate to 4.83 and 4.69 moles of water respectively. Figure 5 shows an example TGA.

Figure 5

"Redetermination of the single-crystal X-ray structure reveals that the structure contains only a pseudomirror plane, and as such, the true space group of the structure is P21 and not P21/m. The asymmetric unit of the structure now contains four molecules of sodium diclofenac and 19 molecules of water, yielding an overall stoichiometry of a 4.75 hydrate," says Frampton.

All the atoms in the crystalline asymmetric unit are now fully ordered, and all hydrogen atoms, including those on the 19 water molecules, were located in the Fourier difference maps and were included within the refinement. The theoretical water content for a fully hydrated 4.75 hydrate is 21.19%, which is in accord with the other analytical data, Frampton explains. The structure now has a much-improved crystallographic R factor of 3.04%.

"It is important to stress in this case that the correct result was only obtained after the consideration of all the analytical data and that we should not be surprised by the noninteger value of the water content," he says. A further hydrated species, a triclinic 3.50 hydrate, was also identified in the study.

Andolast. Andolast, an NCE being developed by Rottapharm S.P.A. (Monza, Italy) for certain respiratory disorders, shows the challenge in polymorph screening and the potential opportunities for discovering and patenting a polymorph.

Andolast is administered as a dry powder for inhalation. "When a product is used in the form of micronized powder for the treatment of respiratory tract disease, the particle-size distribution (PSD) of the inhaled drug is crucial to its performance," explains Markus von Raumer, product manager of the solid-state development unit at Solvias AG (Basel, Switzerland). "Physical chemical characteristics play a fundamental role in the case of inhaled drugs not only for the solubility issues shared with oral drugs but also due to the aggregation phenomena, which can impact negatively the performance of the powder, ultimately affecting its pharmacological activity," he says (12).

During the early pharmaceutical development of andolast, deviations in the consistency of the API's physical chemical characteristics were observed, and a review of the solid-state issues was required, explains von Raumer. These deviations related to: batch-to-batch PSD variability (which affected released fraction and therefore drug efficacy); batch-to-batch variability in hygroscopicity (which affected released fraction and hence manufacturability); and batch-to-batch variability in chemical–physical characteristics that reflected in flow and blend properties (which affected manufacturability and therefore product cost).

Analysis of the complete batch history, combined with fresh physicochemical analysis of key samples, revealed several issues. Deviations occurred independently upon the supplier, upon the batch scale, and upon the batch purity. For all batches, the same chemistry and the same isolation procedure was used.

"Sequential to compilation of historic data, the crystallization conditions and a crystallization process investigation was carried out," says von Raumer. The screening was based upon product precipitation from several water-miscible solvents changing concentration, water content, and temperature. Phase-equilibration experiments were carried out as well. In these experiments, material from the previous batches was slurried in different solvents, at different water concentrations, different temperatures, and for different times. Crystallization from different media gave different solids, in terms of XRPD, Fourier transform (FT)-Raman spectra, and hygroscopicity.

Crystallization in thermodynamic conditions, however, afforded the same product when a proper amount of water was present, points out von Raumer. This product was not hygroscopic, was stable, and its water content determined by KF was about 20%. Crystallization under kinetic conditions, especially when the water activity in the medium was poor, afforded a hygroscopic product with KF lower than 20%. This product was not stable and converted under appropriate conditions into a stable product with KF of approximately 20% and XRPD and Raman results identical with the precipitated product. All the products obtained from crystallization trials that gave KF around 20% displayed the same XRPD spectrum.

"According to the above experiments, it was proven that andolast disodium crystallized in only one preferred hydrated form: the penthaydrate. All other hydrated forms were not stable and converted under appropriate conditions into the pentahydrate," notes von Raumer. "The single-crystal structure was resolved, and molecular-mechanics simulation showed that, upon water removal, a certain degree of distortion was introduced within the crystal structure, with the distortion being proportional to the number of water molecules removed."

From this solid-state study, a new crystalline form, andolast pentahydrate, was revealed, which displayed certain superior properties, says von Raumer. Andolast pentahydrate is not a hygroscopic solid, is stable to mechanical milling, and is stable to as much as 85% relative humidity. It also can consistently be prepared according to a standard process, has constant mechanical properties such as flowability, and has advantages in formulation in comparison with previous material because of a better performance in delivered dose, fine particles delivered, and fine-particle fraction. Aside from improved performance, the new crystalline form of andolast also led to patent applications in the United States and Europe (13, 14). The European patent application calls for an expiration date of 2026 compared with the original expiration date of 2010 that is specified in an earlier patent (15).

"While the development of the andolast drug product was hampered due to inconsistencies and was slowed down by a non-negligible factor, the actual situation allows for a rapid continuation with a possibly much longer time horizon of exclusivity," says von Raumer. "This situation once again demonstrates the utility and opportunities that are created by carefully investigating the solid-state aspects of drug candidates."

Analytical methods for polymorphs

XRPD is the most common analytical method used in polymorph screening, although certain researchers use Raman spectroscopy as their primary screening method. "Each method has its value, and more recently SAFC-Pharmorphix has been experimenting on the use of both methods in combination for polymorph and cocrystal screening using a combined X-ray powder diffractometer with a Raman attachment," says Frampton. "This instrument has been developed in collaboration with Bruker AXS, and a patent covering this technology has recently been published (16). This technology has been further enhanced by a new release of the clustering program, Polysnap2, which can cluster any two-dimensional data and produce weighted combinations." SAFC-Pharmorphix has an active collaboration with Professor Chris Gilmore and his group at the University of Glasgow (Glasgow, UK).

In a typical study, a crystallographer places a compound in a range of solvents and subjects them to a range of crystallization conditions in hopes of obtaining single crystals, Aptuit's Byrn explains. He points out that the number of solvents used in the screening varies. A small polymorph screen should include 8–10 solvents (17). "A more complete screen would require more than 50 solvents," he adds. In recent research, Xu and Redman-Furey outlined an approach to narrow their selection of 57 solvents to 20 (18). Byrn points out that in a related study, Miller suggested an approach of finding the most stable polymorphs by slurring compounds in a variety of solvents (19).

"We now know that solvent-based crystallizations alone do not find all of the forms. In several cases, solvent-based polymorph screens have missed very important solid forms," explains Byrn. In another example, Yu discovered the second most stable form of 5-methyl-2-[(2-nitrophenyl)amino]-3-thiophenecarbonitrile (ROY) using a melt-recrystallization technique (20). "It is well established that ROY had previously been screened by all of the best-known high-throughput techniques, and this form had not been found," explains Byrn, adding, solvent-free techniques such as drying, sublimation, and amorphous crystallization can produce new forms not discovered by solvent-based crystallization.

Byrn offers several examples of techniques used to find polymorphs. In a study at Aptuit/SSCI, nabumetone was recrystallized from a wide range of solvents under different conditions. About 250 experiments were performed, yielding Form I. When nabumetone was recrystallized in capillaries, a procedure known to produce very high supersaturations, a new form was discovered in about 70 of 400 crystallizations. Analysis of these crystallization conditions shows that the appearance of the new form depends on superaturation and quiescence and not on solvent content, Bryn explains.

In another study, SSCI/Aptuit compared the polymorphs formed in a common high-throughput technology with those from capillaries and traditional crystallizations under a broad range of conditions. "It was clear that the plate crystallization technology, which is commonly used in high-throughput polymorph screens, tends to overproduce one form relative to other methods," says Byrn.

Other methods at work

Although XRPD is the most common technique used in the overall solid-state characterization of pharmaceutical materials, other techniques are needed to understand the form, determine how it behaves under stress conditions, discern the relationship between forms, and decide which form is suitable for development.

"The powder pattern will tell you if it is crystalline but will not provide critical information such as solvation state, melting point, water uptake, solubility, and physical stability, for example," explains Byrn. "Thermal data such as differential scanning calorimetry, thermogravimetry, and hot-stage microscopy are used to determine melting temperature, solvation state, desolvation, and form changes upon drying. This information can be directly related to processing such as drying."

Gravimetric vapor sorption is used to measure water sorption and desorption, which can lead to environmental handling guidelines to prevent hydrate formation or dehydration upon exposure to various relative humidity conditions. Other methods such as infrared, Raman, and nuclear magnetic resonance (NMR) spectroscopies can often show specificity between forms that may be more difficult to see with XRPD.

"These techniques, along with XRPD and others, are commonly used for quantitative solid-state method development to quantitate the amount of different forms present in API or drug product or confirm that only one form is present. It is usually desired to have only one form produced, but processing such as drying, milling, and granulation, can commonly lead to multiple forms, indicating that an understanding of form changes during the process and better control of the process is needed,"explains Byrn.

He points to recent advances in screening methodology using Raman microscopy to determine the presence of multiple forms in the same crystallization vessel. "Additionally, NIR is finding some use as a tool for differentiating polymorphs during screening. Finally, the developments of X-ray microscopy at Argonne National Laboratories [Argonne, IL] and other synchrotron sources promises to increase our ability to detect solid forms in small samples."

Other spectroscopic methods may be used in detecting polymorphs. "Like infrared spectroscopy, where the absorption of infrared energy is characteristic of a particular molecular vibration, the absorption of the radio-frequency radiation is characteristic of the magnetic environment of the nuclei, explains SAFC-Pharmorphix's Frampton. When applied to solids, cross-polarization magic-angle spinning solid-state nuclear magnetic resonance spectroscopy (CPMAS SS NMR) can give useful information regarding the crystal structure and polymorphic form of the compound. It also can also reveal the degree of crystallinity of the sample under investigation, he says. "Crystalline samples yield narrow lines as equivalent nuclei are in a constant magnetic environment dictated by the crystal structure, whereas amorphous solids usually yield broader lines reflecting the range of environments found in the solid. CPMAS SS NMR spectroscopy can also give valuable information on the crystalline asymmetric unit of the crystal structure."


1. J. Bauer et al., "Ritonavir: An Extraordinary Example of Conformational Polymorphism," Pharm. Res. 18 (6), 859–866 (2001).

2. U. Griesser, "Relevance and Analysis of Polymorphism in Drug Development," presented at British Association of Crystal Growth Spring Meeting. Lancaster, UK, Apr. 4–6, 2006.

3. M. Haisa, S. Kashino, and H. Maeda, "The Orthorhombic Form of p-Hydroxyacetanilide," Acta Cryst. B30, 2510–2512 (1974).

4. M. Haisa et al., "The Monoclinic Form of p-Hydroxyacetanilide," Acta Cryst. B32, 1283–1285 (1976).

5. A. Burger in Acta Pharm. Technol. 28 (1), 1–20 (1982).

6. P. DiMartino et al., " Preparation and Physical Characterization of Forms II and III of Paracetamol," J. Therm. Anal. 48 (3), 447–458 (1997).

7. G. Nichols and C.S. Frampton, "Physicochemical Characterization of the Orthorhombic Polymorph of Paracetamol Crystallized from Solution," J. Pharm. Sci. 87 (6), 684–693 (1998).

8. D Kovala-Demertzi, D. Mentzafos, and A. Terzis, "Metal Complexes of the Anti-inflammatory Drug Sodium [2-[(2,6-dichlorophenyl)amino]phenyl]acetate (diclofenac sodium," Polyhedron 12 (11), 1361–1370 (1993).

9. C. Castellari and S. Ottani, "Two Monoclinic Forms of Diclofenac Acid," Acta Cryst. C53, 794–797 (1997).

10. N. Jaiboon et al., "New Orthorhombic Form of 2-[(2,6-dichlorophenyl)amino]benzeneacetic acid (Diclofenac Acid)," Anal. Sci. 17, 1465–1466 (2001).

11. N. Muangsin et al., "Crystal Structure of a Unique Sodium Distorted Linkage in Diclofenac Sodium Pentahydrate," Anal. Sci. 18, 967–968 (2002).

12. P.J. Akins, "Dry Powder Inhalers: An Overview," Resp. Care 50 (10), 1314–1312 (2005).

13. A. Giordani et al., "New Crystalline and Stable form of Andolast," US Patent Application 20070149586, June 28, 2007.

14. A. Giordani et al.,"A New Crystalline and Stable Form of Andolast," EP Application 06112427.7, Apr. 10, 2006.

15. F. Makovec et al., "Derivatives of N-Phenylbenzamide with Anti-Ulcer and Anti-Allergy Activity and a Method for their Preparation, " WO 9009989 PCT, Sept. 7, 1990.

16. C.S. Frampton et al., "Combinatorial Screening System with X-ray Diffraction and Raman Spectroscopy," US Patent US 2006/0023837 A1, Feb. 2, 2006.

17. S.R. Byrn, "Pharmaceutical Solids: A Strategic Approach to Regulatory Considerations," Pharm. Res. 12, 945–954 (1995).

18. D. Xu and N. Redman-Furey, "Statistical Cluster Analysis of Pharmaceutical Solvents," Intl. J. of Pharm. 339 (1-2), 175–188 (2007).

19. J.M. Miller et al., "Identifying the Stable Polymorph Early in the Drug Discovery-Development Process," Pharm. Dev. Technol. 10 (2), 291–297 (2005).

20. S. Chen, I.A. Guzei, and L. Yu, "New Polymorphs of ROY and New Record for Coexisting Polymorphs of Solved Structures," J. Am. Chem. Soc. 127 (27), 9881–9885 (2005).