Crystallisation is used extensively in the pharmaceutical industry for the separation and purification of chemical compounds. Recently, radically new technologies have been developed that allow for step improvements in the performance of pharmaceutical crystallisation. This article provides a brief overview of such technologies and discusses the opportunities for industry.
Solution crystallisation is the formation of a crystalline solid state from a homogeneous solution. Different methods exist to create the driving force for crystal nucleation and growth. Common crystallisation methods are based on strategies to lower the solubility (e.g., via cooling or the addition of a so-called anti-solvent) or to increase the solute concentration via solvent removal. Crystallisation can easily achieve a product purity in excess of 99 per cent in a single step at mild operating conditions, which explains its wide application for the separation and purification of intermediate compounds and Active Pharmaceutical Ingredients (APIs) in the pharmaceutical industry. However, pharmaceutical crystallisation is also a complicated process with many interacting physical phenomena and its design and operation is challenged by the need to optimise process performance criteria related to aspects such as throughput and stability. Furthermore, tight requirements related to intrinsic crystal quality attributes such as size distribution, shape, and solid-state form often have to be considered. Such quality attributes will always have some impact on downstream processes (e.g., filtration, drying) and may also affect the final product performance (e.g., bioavailability).
During the last decades, radically new technologies have been developed to achieve step improvements in the performance of chemical processes for applications in other branches of chemical industry, which fall under the umbrella term of Process Intensification (PI). More recently, such technologies have been adopted and optimised for pharmaceutical crystallisation, which provide new opportunities to improve the design and operation of pharmaceutical crystallisation. The success of those technologies is the result of a significant amount of research, industry and regulatory support, and the recent availability of enabling technologies such as advanced real-time Process Analytical Technologies (PAT), reliable process models, and the trend towards continuous manufacturing in pharmaceutical industry. This article aims to provide a brief but systematic overview of the principles of selected PI technologies for pharmaceutical crystallisation. A more rigorous discussion has recently been provided by Wang et al.
Process intensification aims to enable step changes in the performance of chemical processes by deploying radically different processing technologies. Prominent examples include the development of reactive distillation and catalytic membrane reactors. The pharmaceutical industry is traditionally not among the frontrunners of PI. However, in recent years, driven by competition from the growing generic drug market, increasing cost in drug R&D and incentives from regulatory authorities, pharmaceutical industry becomes more involved in embracing innovations for advanced manufacturing.
PI approaches for pharmaceutical crystallisation (Figure 1) can be organised according to a fundamental view on PI proposed by Van Gerven and Stankiewicz, which is based on a classification into four domains (space, time, function, and energy).
Significant spatial inhomogeneity often exists in crystal suspensions. Besides gradients in concentration, temperature and momentum, the distribution of the solid phase in the vertical direction is also often non-uniform. Such inhomogeneous conditions are usually undesirable, as the crystal formation may occur at unintended conditions leading to an offspec final product quality. Furthermore, inhomogeneous conditions lead to a non-uniform product quality. Various PI approaches have been developed in the space domain, which can be divided into two directions: miniaturisation and structurisation.
Miniaturisation refers to crystallisation in a confined space. The confined space helps to create a well-defined and homogeneous solution environment and offers different crystallisation behavior compared to bulk crystallisation due to inherent surface and volume effects. For example, a miniaturised space can offer a large surface-to-volume ratio, which provides opportunities for surfaceinduced heterogeneous nucleation to control crystal formation. Furthermore, crystal growth can be restricted due to physical barriers and fast depletion of supersaturation in a small volume avoiding large crystals. Commonly applied enabling technologies for miniaturisation include emulsions, patterned surfaces and porous materials. Crystallisation inside the dispersed phase of an emulsion has been used for the production of small crystals, drug encapsulation and drug delivery. A patterned surface offers an array of small wells for crystallisation, which can be fabricated by standard lithographic techniques. Although a patterned surface may not be practical for manufacturing, it can be a useful tool for screening optimal crystallisation conditions in the early R&D phase. Finally, porous materials such as porous glasses, MCM-41 silica and alginate gel offer another possibility to confine crystallisation. The use of biocompatible porous materials offers new opportunities to integrate crystallisation with downstream formulation steps to produce drug products with high bioavailability.
Structurisation refers to the use of specific structures within a crystalliser to intensify the mixing of a crystal suspension, which is particularly important when the supersaturation is generated rapidly. For example, in the case of reactive crystallisation or anti-solvent crystallisation, a lack of fast mixing can lead to high local supersaturation, which makes the crystal product quality poorly predictable. Recently, a new generation of tubular crystallisers has been developed, as mixing problems are frequently encountered when operating a tubular crystalliser at low flow velocities. Different mixing principles have been exploited. Static mixers have been used at the inlet of the crystalliser to improve the radial mixing for homogeneous supersaturation and within the crystalliser to keep crystal suspended. Furthermore, plate-type static mixers or baffles have been adopted in the oscillatory-baffled crystalliser design, which is based on an oscillating flow. In a segmented-flow crystalliser, the slurry is divided into small segments that are separated by gas or immiscible liquid, which improves mixing behaviour in tubular crystallisers. Each segment acts as a tiny crystalliser and mixing within the segments is enabled by internal circulation.
PI approaches in the time domain have been developed along two directions: manipulating process time scales or introducing dynamic states. The switch from batch to continuous operation is an important PI approach for pharmaceutical crystallisation processes, which aligns well with the trend in pharmaceutical industry. Furthermore, periodic operation has been investigated as a PI approach to provide a desired dynamic profile of operating states.
A continuous crystalliser can be a single or cascade of Mixed Suspension Mixed Product Removal Crystallisers (MSMPRC) or a Plug Flow Crystalliser (PFC). A single MSMPRC is robust, but exhibits a broad residence time distribution, which may lead to a non-uniform crystal product quality. In this respect, a cascade of MSMPRCs is preferred to obtain a narrow residence time distribution and more uniform product quality. Moreover, the use of multiple MSMPRC in sequence also gives more flexibility to optimise operation. The PFC inherently provides a narrow residence time distribution and, therefore, usually has a higher throughput and more uniform product quality compared to an MSMPRC. The supersaturation profile along the tube can de designed by adding anti-solvent or manipulating the cooling rate at different segments.
Temperature cycling is an example of periodic operation and involves alternating heating and cooling of the crystal slurry. The solution becomes undersaturated after heating, which leads to partial dissolution of small crystals. After a slight increase in solute concentration, cooling will be employed leading to the growth of the remaining crystals and the possible formation of new nuclei. This successive dissolution-recrystallisation can be considered as an intensified ripening process, which promotes the formation of thermodynamically favoured states. Temperature cycling has already found applications for the control of various crystal quality attributes, such as polymorphic form, chirality, shape, and size distribution.
One important general PI principle relates to the execution of different process functions in a single physical space, which aims to achieve synergistic effects and to reduce the number of equipment. Crystallisation has inherent thermodynamic and kinetic limitations. Therefore, combining crystallisation with other separation technologies to create a so-called hybrid separation process can lead to synergistic effects improving the overall separation performance. Membranes have several advantages for solvent removal, which can be combined with crystallisation to generate supersaturation in a more controlled way and potentially with lower energy consumption. Such membrane-assisted crystallisation based on reverse osmosis or membrane distillation has been successfully demonstrated on a small scale. Compared to solvent evaporation, membranes offer a larger surface area for solvent removal which allows for more compact equipment. Another promising hybrid crystallisation process involves the combination of chromatography and crystallisation, which has demonstrated excellent performance for chiral separations. Chromatography is used to enrich one of the enantiomers, which will then be separated sharply by preferential crystallisation to produce the pure enantiomer.
The integration of crystallisation with downstream operations can lead to a more efficient and compact pharmaceutical manufacturing process. For example, so-called spherical crystallisation, which integrates crystallisation and agglomeration in a single processing step, has been developed. Agglomeration is induced conventionally by granulation to improve the powder properties for tableting. With the help of different spherical crystallisation methods, agglomerates can be formed directly. The commonly used spherical crystallisation methods include spherical agglomeration, quasiemulsion solvent diffusion, crystalloco-agglomeration and ammonia diffusion.
The use of external fields is one of the most studied PI approaches in the energy domain for crystallisation processes, as external fields can manipulate crystallisation conditions on a local scale, which is not well possible in traditional crystallisation equipment. Ultrasound has been investigated as a method to control nucleation in crystallisation processes for many decades. Such so-called sonocrystallisation can lead to either an increased or decreased induction time for crystal nucleation depending on the operating conditions and can manipulate the crystal size distribution. Ultrasonic waves create cavities in solution, enhance mixing and provide heat, which are all effects that can have an important impact on the crystallisation kinetics. Sonocrystallisation is especially useful for making small API crystals. Moreover, sonocrystallisation has also been demonstrated to accelerate the crystallisation of large biomolecules such as proteins, which are often notoriously difficult to nucleate.
The interest in the use of electric fields (both AC and DC) to manipulate crystallisation has increased significantly during the last two decades, which is especially the case for large biomolecules such as proteins. An electrical field can provide some temporal and spatial control over crystallisation kinetics, which can be used to control crystal formation and growth. Traditionally, electric-field assisted crystallisation has been used to produce single highquality protein crystals for X-ray diffraction analysis. However, recently, electric fields have also been used to increase the yield of protein crystallisation in continuous flow.
The application of other external fields for crystallisation, including microwave, magnetic fields and laser, has also been reported. Microwaves can be used for rapid heating, which can be used for fast temperature cycling or solvent evaporation. The use of magnetic fields has been demonstrated to enable better control over the polymorphic form and spatial orientation of crystals for some cases. Finally, laser is mainly used to enhance the nucleation rate by either photochemical or non-photochemical effects.
This article provides a brief overview of PI approaches used for pharmaceutical crystallisation based on a fundamental framework for classification of PI methods. There exists a significant amount of literature on this topic, which demonstrates that much opportunities to intensify pharmaceutical crystallisation processes exist. However, commercial applications of PI for pharmaceutical crystallisation published in open literature are rare. Nevertheless, the current business environment in pharmaceutical industry (a changing mindset, increased competition, good regulatory support) and the development of new technologies provide a favourable ecosystem for future adoption of PI for pharmaceutical crystallisation. Those new technologies include specific technological innovations based on the fundamental principles of PI as well as enabling technologies such as PAT and continuous processing. Much of the work mentioned in this article is recent, which emphasises the strong current interest in the topic. Consequently, a broad window of opportunities exists for enabling novel PI approaches for pharmaceutical crystallisation due to a favourable combination of regulatory, economical, and technological developments.
Richard Lakerveld is an Assistant Professor in the Department of Chemical and Biological Engineering at The Hong Kong University of Science and Technology and an Honorary Assistant Professor in the Department of Pharmacology and Pharmacy at The University of Hong Kong. He received his PhD from Delft University of Technology and was a Postdoctoral Associate at MIT.
Jiayuan Wang is a senior research postgraduate student at the Hong Kong University of Science and Technology. He reveived his BSc degree from East China University of Science and Technology and his MSc degree from Otto von Guericke University Magdeburg.His research and career interests are on industrial pharmaceutical crystallisation.
Fei Li is an Assistant Professor in the School of Chemical Engineering and Technology at Tianjin University. She received her Ph.D. degree from McGill University and was a Postdoctoral Associate at The Hong Kong University of Science and Technology.
