Crystal Clear: Perfecting Product Purity Through Recrystallization

In the development of new pharmaceutical ingredients and other high-purity products, carefully controlled crystallization processes can help to eliminate critical impurities and zero in on optimal formulations

Historically, small molecules have played a significant role in the pharmaceutical landscape, representing approximately 90% of all marketed drugs and 70% of new drugs that reach the market [1]. While the pharmaceutical landscape constantly evolves to include newer modalities for the treatment of patients, small molecules continue to play a significant role in treating a diverse array of indications, including unmet medical needs within the fields of neurology, oncology and infectious disease.

However, the development of small-molecule drugs and other high-purity substances is becoming increasingly complex. Although the use of technology, such as in silico molecular screening tools alongside artificial intelligence (AI), is streamlining the process of discovering and developing new drug candidates, challenges remain [2, 3, 4].

Many emerging small molecules are complex and exhibit development challenges, such as low solubility and poor permeability. For example, approximately 70% of emerging drug candidates demonstrate extremely limiting solubility [5] — potentially limiting the utility of the drug product through a lack of uptake in the patient.

To address these challenges, a key process in high-purity product development is controlled crystallization. This technique is essential for achieving the desired critical quality attributes (CQAs) of a substance, which ensures its quality, safety and efficacy. By carefully controlling the crystallization process, scientists can remove impurities, ensure the substance exists in the most stable polymorphic form and optimize particle size and morphology. Identifying and achieving these CQAs ensures drug suitability for human use while meeting regulatory needs.

This article explores the intricacies of controlled crystallization, emphasizing its importance in achieving desired product-quality attributes (Figure 1) and navigating the specific challenges of small-molecule drug development.

 

Strategic classification systems

Achieving the desired CQAs often requires careful consideration of a drug’s physicochemical properties. The Developability Classification System (DCS), based on the Biopharmaceutics Classification System (BCS), is used to classify drugs based on their solubility and permeability limitations (Figure 2) [6]. This classification may then directly impact crystallization strategies applied to drug substances to enable drug product development. An understanding of BCS and DCS intricacies is crucial for drug development, an example being the application of biowaivers for demonstrating in vivo bioequivalence of drug products [7].

Strategically, an integrated, iterative approach is the best way to deliver a high-purity substance from a crystallization process. This enables maximum efficiency and risk mitigation during the overall development process, leading to a desirable, high-quality product.

 

Form-specific product recrystallizations

The temperature and solubility behavior of a compound in solvents is the foundation upon which any crystallization process is built. Provision of these data requires an understanding of the polymorphic landscape of a molecule, because this directly impacts solubility. Complex molecular behaviors do not always go hand in hand with complex structures, as some very simple molecules have numerous polymorphic forms, solvates and enantiotropic form relationships. Therefore, care is required when assessing and planning the development requirements of any new drug substance.

Regardless of the complexity, any recrystallization process will require solvent selection that defines which solvents efficiently remove impurities. This is the first step toward the provision of the desired high-purity substance. Ideally, this enables a scientist to develop an ordered crystallization via cooling as the molecule is removed from solution to provide a thermodynamically stable crystalline form. However, sometimes metastable phases are required, adding another layer of difficulty to the design and control of a process.

It is also a regulatory requirement to define a stable polymorphic form within a chemistry, manufacturing and control (CMC) dossier that allows for a suitably controlled production method. Groups that bypass gaining a sufficient and fundamental understanding are at risk of isolating an alternative form with altered solubility, stability and dissolution properties. These are only a small number of the associated problems, and the potential risk should not be underestimated. In extreme cases, the identification of a new form late in development or once commercialized can lead to product recall and significant loss of earnings — the case of Ritonavir being well-documented [7].

A suitable and phase-appropriate screening strategy is usually employed to define the most thermodynamically favored form ahead of Phase 1 clinical trials. At this juncture, the production route is in its early stages of optimization, and batches of material are limited to low-gram amounts. Although the route may not fundamentally change, the specifics relating to reagents, stoichiometry and solvents employed often do change as a compound progresses into late-stage development and quality by design (QbD) investigations. This leads to a process that is appropriate for validation. Such changes during early development can have a significant impact on retention, purge of impurities and the composition of the batch. Targeting a thermodynamically favored form based on polymorphism studies is a gold standard for development. This form usually packs efficiently within the crystalline lattice, avoiding imperfections. Exclusion of imperfections reduces the risk of entrapping solvent — and more importantly, compound-related impurities. From the drug developer’s perspective, the selection of the stable form supports the achievement of CQAs and avoids the risk of form change within the drug product.

 

Achieving form and purity

The composition of early development batches is a critical element of focus for the development chemist and crystallization scientists. For this reason, polymorphic screening for stable forms prior to Phase 1 will use high-purity input materials to reduce the impact of impurities on material behavior. Some impurities of note to consider when looking at what constitutes a process-typical batch for early-phase development are:

• Oligomers
• Polymers
• Trace metals (various oxidation states) and ligands
• Process-related impurities
• Inorganic content
• Partial salt forms

Not all of these impurities are readily detectable, so multiple analysis modes should be employed (Figure 3). It is best practice to confirm form behavior during early development with “process typical” specification to try and reduce the risk of failure due to the impact of impurities at a later stage. Process impurities that are structurally related can interfere with growth sites and spoil what might otherwise be a reliable crystallization process. They may be trapped via inclusion during a process that is not operating under controlled conditions, or may spontaneously nucleate and crystallize to provide an unwanted heteronuclear seed bed. Appraising which mechanisms are at play is not always simple under the time and material constraints (including the number of available batch streams) encountered during pre-clinical development.

During these investigations, the ability to integrate early-phase batch profiling with preliminary crystallization development and polymorphism screens becomes essential to the smooth progression of the project. Form fates, impurity behavior and solubility data are as important to the process chemist trying to define a suitable reaction work-up as they are to the crystallization chemist. Both have the same goal: isolation of solids with a profile that excludes as many impurities as possible (see later discussion regarding toxicological cover).

 

Pre-clinical production scaleup

A more detailed analysis of solubility and chemical stability for both pure and process-typical behavior is required for successful scaleup. This determines which type of crystallization methods to evaluate and guides scaleup strategy and operational temperature limits. The ideal is a simple and controllable cooling protocol that allows for clarification (such as screening filtration to remove debris). This usually offers the most efficient vessel occupancy times and throughput if solubility aligns for a process operating at 5–10 volumes/weight. It is typical to run through a small number of demonstrative crystallizations on 0.5–3 g of crude material prior to selecting a lead system. Following these assessments, unwanted features, such as phase separation and unacceptable slurry mobility, can be correlated to solvent type and avoided. Defining an accurate temperature-solubility curve and metastable zone width (MZW), and demonstrating crystallization seed control are usual in the provision of a robust process.

Understanding the critical process parameters through a pragmatic yet robust study period should create the basis for the successful isolation of the desired form with an appropriate purity profile.

After defining solvents and anti-solvents, calculating cooling or anti-solvent driven processes, and conducting a short series of demonstrative recrystallizations, the limitations and potential points of failure with regard to maintaining or losing control of the process become clearer. The application of parallel screening tools, such as the Crystal16 platform for a temperature-solubility study and form screening, in parallel to small-scale batch crystallization with full-beam reflectance mode/turbidity and optical microscopy probes, all help to de-risk system selection. These probes are especially useful to visualize and measure the growing crystals and gain a fundamental understanding of how to drive the process to the desired endpoint. These data will provide information that is critical for optimization and control, including the following:

• Crystallization kinetics
• Impact of cooling and anti-solvent upon solubility and control of supersaturation
• The impact of supersaturation control on:
  • Morphology
  • Impurities
  • Agglomeration
• Ostwald ripening (a phenomenon where the smallest crystallites dissolve preferentially and stimulate growth of the larger, existing crystal population [9]) is of benefit at a given point in the process to manipulate morphology, particle size or purity.
• Isolation temperature and wash regimes to control purity

Although the primary goal is to derive a product batch that is suitable for clinical evaluation, material that contains impurities or byproducts that are likely to be present due to process or stability issues is required for toxicological evaluation. A typical initial specification allows a small number of such impurities at 0.2–1.0 area % by high-performance liquid chromatography (HPLC). Therefore, communication and integration with process chemists can help in the design of a protocol with isolation for toxicological coverage in mind. Common questions and considerations include the following:

• What sort of batch variance is anticipated?
• Can a tight specification be justified with only one to two small- or intermediate-scale test batches in hand to base a risk assessment upon?
• What is the ultimate fate of the material, will some element of pre-formulation design be based upon the batch profile destined for toxicological evaluation?
• Will particle size and habit be taken as typical of the process?

These considerations often necessitate a “quick and dirty” crystallization or a slurry, but not all quality attributes may be met by this approach, especially regarding particle size and morphology.

As mentioned, discussing risk and behavior with an integrated development team brings significant benefits. Spiking of crystallization liquors back into a process is often discussed. This activity is often disfavored, but can offer a path forward under the right circumstances so long as batch homogeneity is stringently considered. Inefficient wash regimes or pre-slurry steps are another useful option to list.

In some instances, a simple change in cooling rate, anti-solvent addition rate or isolation temperature is all that is required. These instances are ideal, because the process is being conducted to mimic but not replicate the clinical process, so residual solvents remain the same.

 

Integrated development

Successful crystallization development relies on a robust investigation of form and solubility behavior, with impurities playing a critical role. An integrated approach, combining chemical and solid-form development teams, is crucial for understanding material behavior and impurity profiles. This collaborative approach enables the rationalization of polymorphic fate and the control of impurity and morphology attributes during recrystallization.

In simple systems, a well-controlled seeded crystallization will often proceed to pack tightly and exclude impurities. Unfortunately, a significant number of molecules do not behave in this manner. They demonstrate disorder in their crystalline lattice or display stability and solubility properties that make removing impurities challenging. Impurities that have very similar and limiting, or lower solubility versus the parent molecule can be particularly difficult to purge and can negatively impact performance of a crystallization. Using an integrated approach promotes the discussion of where troublesome impurities should be removed and what is tolerable. A robust understanding of chemical stability and the limitations and tolerances of large-scale equipment (heat or mass transfer, mixing and suitable hold points) is also essential for successful scaleup.

There is no value in over-engineering smaller-scale batch operations with addition rate and temperature profile parameters beyond the accuracy of the large-scale equipment. It is usually the case for early-phase development that an alternative exists and this should constantly be challenged ahead of committing to an expensive and complex protocol. The inclusion of process analytical tools, such as focused-beam reflectance measurement (FBRM) or particle vision and measurement (PVM), is encouraged during early screening runs to enhance understanding. They allow the scientist to very accurately visualize and measure the growing crystals from various test systems and aid in the selection of an optimal process. When available, the study of single-crystal structural information can further optimize the process, assess form landscape risks and inform in silico modeling to identify potentially more stable forms [ 8].

Ultimately, pragmatic project management and cooperative discussion are vital for identifying critical quality attributes and producing a viable drug product. Designing a crystallization process that effectively eliminates impurities and understands their impact on form remains crucial for any development program. ■

Acknowledgement

All images provided by author

References

1. Hartmut, B., Harter, M., Hab, B., Schmeck, C. Baerfacker, L., Drug Discovery Today, vol. 27, No. 6, 2022, pp. 1,560-1,574.

2. Abbott, A., Dementia: A problem for our age, Nature, vol. 475, July 2011, pp. 52–54.

3. Ngo, H.X., Garneau-Tsodikan, S., What are the drugs of the future?, Med. Chem. Comm., Issue 5, May 2018, pp. 757–758.

4. Mullard, A., The Drugmaker’s Guide to the Galaxy, Nature, vol. 549, September 2017, pp. 445–447.

5. L. Benet, Predicting drug disposition via application of BCS: transport/absorption/elimination interplay and development of a biopharmaceutics drug disposition classification system, Pharm. Res., January 2005, pp. 11–23.

6. Butler, J.M., Dressman, J.B, The developability classification system: Application of biopharmaceutics concepts to formulation development, J. Pharm. Sci., vol. 99, Issue 12, December 2010, pp. 4,940–4,954

7 Bauer, J., Spanton, S., others, Ritonavir: an extraordinary example of conformational polymorphism, Pharm. Res., June 2001, pp. 859–866.

8. Sadiq, G; Sharma, S., others, An integrated approach combining experimental, informatics and energetic methods for solid form derisking of PF-06282999, J. Pharm. Sci., Vol. 114, Issue 1, January 2025, pp. 371–382.

9. Olsson, U., others, On the Ostwald ripening of crystalline and amorphous nanoparticles, Soft Matter, Issue 12, February 2025.

Author

Julian Northen is currently the Solid State manager at Onyx Scientific Ltd. (Sunderland, U.K.; Phone: +44 (0) 1915166516; Website: www.onyxipca.com). He graduated from the University of Newcastle upon Tyne and holds a Ph.D. in medicinal chemistry and anti-cancer drug design. He also held two post-doctoral positions before a move from academia to Onyx. He has over 20 years of industrial experience in product research and development. In his current role at Onyx, he is responsible for all solid-form development, crystallization development and preformulation activities.