Crystallization is usually the final stage in the manufacture of an active pharmaceutical ingredient (API, the drug substance). Crystallization is commonly achieved by cooling or adding another ingredient to a solution of the API in a suitable solvent, which may be aqueous or organic. In recent years, supercritical fluids (SCFs) have been introduced as the crystallization medium. Crystallization from SCFs may be achieved by gas anti-solvent (GAS) addition, by rapid expansion of supercritical solutions (RESS), or by precipitation with compressed antisolvents using either the supercritical antisolvent (SAS) process or an aerosol spray extraction system (ASES). Certain properties, such as particle size and purity, of the crystalline product may change as the supersaturation changes during crystallization from liquid solutions. However, it is claimed that crystallization from SCFs yields a much more uniform crystalline product. This uniformity appears to result from the high degree of control that can be exerted on the crystallization process, which is usually very rapid.
The properties of the crystalline product depend on the precise conditions of the crystallization process, including the initial and final supersaturation, the desupersaturation and temperature regimes, the nature of the solvent and the impurities.
The development of a given crystal during the crystallization process proceeds through three consecutive stages, namely, nucleation, crystal growth, and Ostwald ripening. In the nucleation stage, molecules of the substance aggregate to form embryos. If the conditions are such as to allow an embryo to reach a critical size, known as a nucleus, this nuclear aggregate can grow to form a macroscopic crystal. If the conditions do not allow the embryo to reach the critical nuclear size, the embryo dissolves. If the crystallizing substance is capable of existing in more than one crystalline phase, such as polymorphs or solvates, each phase is associated with its own specific embryonic aggregate and nucleus. In the supersaturated solution, the different embryos compete for solute molecules. The type of embryo that first reaches the critical nuclear size forms a nucleus for that particular crystalline phase and hence enables that phase to grow into macroscopic crystals. Because of the time element involved in the race for nucleation, this step is controlled by kinetic considerations provided that the thermodynamic driving force for the formation of the crystallizing phase is favorable, i.e., ΔG is negative.
Crystal growth proceeds by successive incorporation of molecules, first onto the surface of the nuclei and subsequently onto the surface of the growing crystals. Provided the solution is supersaturated, this growth process is spontaneous (ΔG is negative) and is accompanied by an increase in molecular order (ΔS is negative). Hence, crystallization is accompanied by the liberation of heat, the exothermic enthalpy of crystallization (ΔH is negative). Within the thermodynamic limits, the kinetics of crystal growth may be explained by various models. If nuclei of more than one phase are present, the relative amount of each crystalline phase that is formed at a given time is determined by the relative number of nuclei of the phases and the relative rates of crystal growth of the phases.
The supersaturation of the crystallization solution is expressed as some function of the ratio of the actual concentration of the solute in the solution at a given time to the solubility of the solute in the solvent at the same temperature. A logarithmic function of this ratio is directly proportional to the negative free energy change (-ΔG), which represents the driving force for the processes of nucleation and crystal growth. The greater the supersaturation, the greater the number of nuclei that are formed per unit volume of solution per unit time and, also, the greater the rate of crystal growth. However, the nucleation rate is much more sensitive than the growth rate to the degree of supersaturation. Hence, a high supersaturation yields many small crystals in such numbers that the solution soon becomes depleted of solute molecules, thereby limiting crystal growth. On the other hand, a solution for which the supersaturation is low may take a long time to nucleate, leading to few nuclei that grow slowly and give rise to a few, relatively large crystals.
The nucleation step can, of course, be eliminated by adding seed crystals to the supersaturation solutions. These seed crystals will then grow as described above. However, nucleation can occur as long as the solution is supersaturated, even during crystal growth. Continuing nucleation is one of the main factors that explains a wide particle size distribution of the crystallized product. For a solute that is capable of existing as different phases, addition of seeds of the desired phase usually results in phase-pure crystals of that form. Hence, crystal growth can be directed and controlled by seeding.
During crystal growth and Ostwald ripening, certain crystal faces may grow more slowly than others, thereby giving rise to a change of crystal habit. The fastest-growing faces grow out of existence, so that the overall crystal habit is dominated by the slowest-growing faces.
Furthermore, during crystal growth and Ostwald ripening, any impurities present in the solution or in the crystals may redistribute themselves between the solution and the crystals and between the crystals themselves. Usually, however, the crystallization process leads to a crystalline product that contains smaller amounts of impurities than the original material that was dissolved to prepare the supersaturated solution. Exceptions to this statement arise when the impurities bear strong similarities in molecular structure and shape to the substance being crystallized or when an impurity forms a solid solution in the crystals.
The formation of solid solutions corresponds to the creation of impurity defects in the crystals. If the solid solution is formed under conditions of crystal growth that are close to equilibrium, such a solid solution will be stable. The intrinsic dissolution rate of the crystals will then be less than that of pure crystals, because of the reduction in free energy caused by the dilution effect of the impurity. On the other hand, if the solid solution is formed under conditions of crystal growth that are far removed from equilibrium, as during high initial supersaturation, such a solid solution will be metastable. The intrinsic dissolution rate of the crystals will then be greater than that of the pure crystals, because of the increased free energy caused by the lattice strain introduced by forcing the impurity molecules into the crystal lattice under non-equilibrium conditions. Under these conditions, impurity defects are formed.
Other types of crystal defects are introduced during crystal growth, such as point defects (vacancies and interstitial sites), line defects (edge and screw dislocations), plane defects (twinning and grain boundaries), and phase defects (solid, liquid, or gaseous inclusions). Usually, more rapid crystallization leads to a greater concentration of defects in the crystals, because of a greater number of errors in incorporation of the solute molecules in to the crystal lattice.
Dr. David J. W. Grant
SSCI Consultant
For more information on SSCI, please see our contacts page.[top]
© 2003 SSCI Inc.