Recrystallization is a method employed to purify solid compounds. It involves dissolving the impure solid in a heated solvent until the solution reaches saturation, followed by gradual cooling. As the solution cools, pure crystals form while impurities are left behind. These crystals are then collected, washed, and dried, resulting in a purified solid product. Recrystallization is a crucial technique for obtaining high levels of purity in solid compounds.
The underlying mechanisms, dissolution, and recrystallization can also minimize the crystal's internal energy in order to reach a more global energy equilibrium resulting a stable polymorph. While recrystallization is usually applied deliberately to optimize crystals and processes, uncontrolled recrystallization can lead to the unwanted formation of hydrates and solvates or polymorph transformation.
7 Key Crystallization Mechanisms
Free Guide
Crystallization proceeds through a series of interdependent mechanisms uniquely influenced by the choice of process parameters: nucleation, growth, oiling out, agglomeration, breakage, seeding, and polymorph transition.
Often, these mechanisms occur simultaneously, making effective crystallization design challenging. Without mechanistic understanding, scientists rely on trial and error to adjust process parameters and optimize yield, purity, and crystal size. By understanding which mechanisms occur during crystallization, scientists can deploy strategies to deliver a crystal product with the desired properties.
Read our 7 Crystallization Mechanisms Hiding Between Your Samples Guide to learn how mechanisms can influence the outcome of your crystallization process.
7 Step Recrystallization Procedure
This seven-step recrystallization guide provides essential information to design efficient recrystallization processes with optimal product and process performance.
Step 1: Recrystallization Solvent Screening
The first step of recrystallization is solvent selection. To identify the characteristics of a good recrystallization solvent, several factors should be considered:
Solubility: The solvent should have high solubility for the compound to be purified at elevated temperatures, allowing for the efficient dissolution of the impure solid.
Selectivity: The solvent should selectively dissolve the desired compound while leaving impurities insoluble or minimally soluble. This ensures effective separation during the recrystallization process.
Temperature Range: The solvent should have a suitable boiling point and melting point range to facilitate dissolution at elevated temperatures and crystallization upon cooling.
Inertness: The solvent should be chemically inert towards the compound and impurities, ensuring that it does not react or introduce additional impurities during the recrystallization process.
Volatility: The solvent should possess adequate volatility to allow for easy removal after recrystallization, typically through evaporation, leaving behind pure crystals.
Safety: The solvent should be non-toxic, non-flammable, and environmentally friendly to ensure safe handling during the recrystallization process.
Depending on its molecular structure, a solute can be classified as soluble, partially soluble, or insoluble in these solvents or solvent mixtures. To understand this thermodynamic relationship, it is necessary to assess the solute's temperature-dependent solubility behavior in a wide range of different solvents. Solubility can be high or low at room temperature, a strongly increasing function of temperature, or almost independent of temperature at all. High throughput screening methods quickly determine solubility data for a large number of solute/solvent systems.
Cooling crystallisation is commonly applied when the solubility of a compound in a given solvent is a strong function of temperature. A hot, fully dissolved solution is cooled to a lower temperature where the obtained supersaturation initiates recrystallisation of the desired compound. The applied cooling rate and obtained supersaturation to a large extent determines product quality and downstream performance of the product crystals. Anti-solvent crystallization is carried out when the solubility of a compound is a strong function of the solvent composition. The compound is dissolved in a good solvent and immiscible second solvent with very low solubility for the compound acts as anti-solvent. In normal operation, the anti-solvent is added into the solution to reduce solubility in the solvent mixture and to increase super saturation required for re-crystallization. In a process called reverse edition the anti-solvent is provided and the solution is added to it. Both methods, normal operation or reverse edition as well as the rate of edition have a major influence on the obtained supersaturation profile in the bulk mixture and properties of the final crystal product. Systems where the solubility is neither a strong function of temperature nor solvent composition are suitable for evaporative recrystallization. The fully dissolved solution is heated to its boiling point where super saturation for recrystallization is created by evaporating off a significant portion of the solvent. The rate of solvent evaporation and creation of super saturation have a major impact. on process robustness and final product properties obtained. Reactive recrystallization involves the chemical reaction between two complex molecules or an acid-based neutralization to create the desired compound. The dosing and mixing of the reagent increases the compound concentration in the solvent mixture until sufficient supersaturation is created for recrystallization to occur. The rate of addition and the degree of mixing have a large influence on the properties of the final crystal product. Based on solubility data, one or more recrystallization methods can be applied to obtain crystals with desired properties or to maximize yield. Careful control over the degree of supersaturation and an understanding what particle mechanisms crystals go through, for example nucleation, growth, secondary nucleation, agglomeration, or polymorph transformation is essential to design efficient recrystallization. Processes with optimal downstream performance
Step 2: Selection of Crystallization Method
A. Cooling Recrystallization A solute/solvent system with low solubility at low temperature, but with solubility being a strongly increasing function of temperature, is suitable for cooling recrystallization. A large amount of solute can be dissolved at high temperature and due to lower solubility at low temperature, controlled cooling can initiate recrystallization. Temperature sensitive solutes that decompose at elevated temperatures are not suitable for cooling recrystallization.
B. AntiSolvent Recrystallization Solute/solvent systems with high solubility at low temperature and the availability of a miscible antisolvent meet the requirements for antisolvent recrystallization. The controlled addition of antisolvent reduces solubility in the mixture and triggers recrystallization. Two common ways of operation are either antisolvent addition to product solution or product solution addition to antisolvent (reverse addition). The disadvantages of antisolvent recrystallization are introduction of an additional solvent, local high supersaturation at the point of addition, reduced volumetric productivity, and the requirement for solvent separation downstream.
C. Evaporative Recrystallization High solubility at low temperature and the unavailability of an antisolvent often requires evaporative recrystallization. The removal of solvent reduces solubility in the remaining mixture and once sufficient supersaturation is created, recrystallization occurs. Challenges in evaporative recrystallization are the introduction of gas bubbles, which can act as a source of nucleation, difficult to predict seeding points and unpredictable scale-up.
D. Reaction (Precipitation) Recrystallization When the desired solute is generated through a chemical reaction between two complex compounds or an acid/base neutralization, the method is referred to as reactive recrystallization. The progressing chemical reaction increases supersaturation of the solute, which eventually recrystallizes. Creation of supersaturation can be extremely fast, leading to local high supersaturation at the point of mixing, extensive nucleation, poor process control, and difficult downstream handling.
Step 3: Thermodynamics and Kinetics
Understanding the solubility behavior of a compound is an important requirement for successful recrystallization process development. Knowledge about how much solute can be dissolved in a solvent and how much solute will remain in the mother liquor at the end is essential to assess recrystallization efficiency. For solution crystallization, the solubility is the maximum amount of solute that can be dissolved in a given amount of solvent at a specific temperature.
A system is supersaturated when the concentration of dissolved solute exceeds the solubility limit at a given temperature. Depending on kinetics, the solution has the ability to remain supersaturated over a range of temperature and time before it recrystallizes. The elapsed time between the creation of supersaturation and the formation of the first crystals is called induction time. Increasing supersaturation reduces induction time to a point where the formation of crystals happens spontaneously as soon as supersaturation is further increased. This point is defined as the Metastable Boundary, the difference between solubility curve and Metastable Curve being the Metastable Zone Width.
Step 4: Determine a Seeding Strategy
During recrystallization, tight control over supersaturation, nucleation, and growth is required to achieve a final product with desired physical properties. The help of seed crystals is often critical to start the recrystallization processes robustly and repeatedly. Unseeded processes often show uncontrolled spontaneous nucleation, which can lead to extreme process variation, especially during scale-up and large-scale manufacturing. Typically, the following seeding strategies are applied:
Tiny (0.1-1 %): Hopefully avoiding oiling out, uncontrolled nucleation, or crashing out. It may be satisfactory in the laboratory, but is rarely effective or robust at larger scale.
Small (1-5 %): Controlled nucleation but not enough to exclusively obtain crystal growth. Secondary nucleation along the process and bimodal distributions are likely.
Large (5-10 %): Improved probability of crystal growth with the chance to suppress secondary nucleation to avoid bimodal distributions.
In this short video, I'd like to demonstrate that significant process knowledge can be obtained pretty easily by visualizing seeding mechanisms during crystallization development. The seeding step during crystallization is an important one. Seed crystals are added to the crystallization process in the metastable zone width in order to help induce nucleation. There are a number of variables that can be adjusted during seeding in order to produce different results. These can include the size of the seeds used, the shape of the seeds, the mass of seed added to the seed loading, and indeed the polymorphic form of the seed chosen for a crystallization process where the polymorphic form needs to be controlled. Really seeds are added for two reasons. The first might be to ensure that the crystallization starts at the same temperature or concentration for every experiment or every batch. This can really help with consistency and repeatability for the crystallization process. The next reason that seeds might be added in addition to this would be to provide surface area in the crystallizer to drive kinetics in favor of either growth or nucleation. And in order to optimize the seeding step, visualizing the seeding event itself can really help provide significant process knowledge that can support optimization. So there's two examples where we'll actually directly visualize the seeding step for two different crystallizations. In the first case study, we'll look at how secondary nucleation can occur during seeding. And in the second case study, we'll actually look at seed crystals agglomerating and the impact that can have on the process. In our first example, compound A is crystallized at a one-liter scale in an OptiMax reactor, and particle view with PVM technology is used to visualize the crystallization process using real-time microscopy. This is a seeded crystallization, obviously, followed by an antisolvent addition. So here we see one of our first real-time microscopy images collected using particle view, and we can actually see what the seed looks like from a microscopic perspective directly in the process. This is how seeds actually are existing in the process itself, rather than what they might look like after sampling and preparation for visualization under an offline microscope. So what we can see is the seeds are added to the crystallizer, they form these loosely bound flocks. these loosely bound flocks appear to kind of form more densely formed aggregates or agglomerates. And then after 7 minutes, we really begin to see the direction this crystallization process is going to take. The seed crystals have agglomerated or aggregated, and now what we're seeing is what appears to be surface nucleation with small needle-shaped crystals growing out from the center of these seed aggregates. Dendritic growth, maybe aware that you've you could apply in this case. After 10 minutes, we begin to see these seed aggregates have grown even more, but now what we're beginning to see is these small needle-shaped crystals are really beginning to grow from the aggregates and are actually beginning to shear off into the crystallizer itself. And this is the beginning of a secondary nucleation event. And this occurs, as can be seen from the timestamp, at about 10 minutes. Now if we look later in the process after 30 minutes, you can see we've actually ended up with a highly bimodal system where we have the seed aggregates, large seed aggregates in the background combined with a very high quantity of small needle-shaped crystals which have been formed due to the secondary nucleation event. And really what this means is this product is difficult to filter and handle due to the bimodal properties. And then secondly, the root cause of this poor process quality and product quality can directly be attributed to the seed itself. So the process knowledge that's obtained here by visualizing the seeding step is very important because now an optimization is possible. Clearly, to avoid this secondary nucleation event and this bimodal result, the seeding step must be characterized in more detail and improved in order to improve the process and the product itself. So, in our next example, we'll actually visualize seed agglomeration. So, in this case, we actually used two technologies, particle track and particle view, in an EasyMax reactor. And this is actually looking at a seeded crystallization where there is an isothermal hold for 60 minutes. Three experiments were conducted with different seed loadings. The first was actually an unseeded crystallization process, so 0% seed, followed by a seed loading of 0.1% and 2%. And really what we're trying to do here is study the effect of seed loading on the crystallization mechanism itself, but then also the final crystal size. So what we can see for the unseeded crystallization is that after 5 minutes, we begin to see the crystals forming. After 10 minutes, you can see they've actually grown quite a lot. And then after 45 minutes, you can actually see they appear to have broken apart quite a bit and are actually much smaller. So what's happening here for the unseeded crystallization is really a two-step process. The first is agglomeration and growth of those initial crystals, followed by the breakage of those aggregates or agglomerates into a much finer crystal suspension. So the question in this case was if seeding would actually help avoid this intermediate agglomeration step. So at the bottom here we have our unseeded crystallization process and what that looks like as time progresses. And then as we increase seed loading, we can actually begin to see differences in the mechanisms that occur during crystallization. So what's apparent from this set of images is that agglomeration occurs every time regardless of of the seed loading used. But the visualization here really helps us understand that this agglomeration step has occurred, and we can also see qualitative differences in what the crystallization looks like with these real-time microscopy images. Now if we look at the 2% seed crystallization process, we look at our particle track data, which is actually tracking the rate and degree of change to particle size and count in process, we can begin to obtain more quantitative information about what's happening during this seeding step. So we can see the point at which seed is added right here around 35 minutes, And then what we begin to see is these very large counts increase, indicating this agglomeration step right up to about 50 minutes. And then those large counts, the orange trend, begin to decrease, and then the smaller counts between 0 and 50 microns increase before we reach a steady state at around one hour or one hour and 5 minutes. What's really important here to show is that the dynamics of this process around seeding, there really is a significant agglomeration and breakage event that occurs, and it's important to understand this as this crystallization process gets developed and ultimately scaled up. Next, we can look at the three crystallization processes in a slightly different data representation, looking in three dimensions. So here we're using what's essentially a waterfall plot to study our chord length distributions measured by particle track and how they evolve over time. So essentially what we're looking at here is a heat map indicates the particle counts in different size ranges and then vertically we're progressing over time here. So what we see is that for the unseeded crystallization process we get nucleation followed by breakage. It's the same for the seeded processes but there's less breakage that occurs when we use 2% seed. Also when we have an unseeded process we actually start off with quite large particles initially. For the.1% seed, they start off smaller, and then for the 2% seed, they actually start a lot smaller than that even again. However, the end point for each of these is reversed. So when we have the unseeded process, we actually end up with small particles, but for the 2% seed, we end up with the largest crystals for all three experiments conducted. So what we conclude from this is that increasing the seed loading reduces the agglomerate size that's formed and potentially increases the agglomerate strength. Maybe the binding of these agglomerates is stronger, meaning that the mixing inside the crystallizer is less able to break them apart. So what we've really determined here is that seed loading is a critical parameter to control the crystallization and potentially the final particle size, especially as mixing conditions change, if the strength of those agglomerates is really influenced by the seed loading chosen. So quickly to conclude, visualization of the seeding step during crystallization process can really deliver significant process knowledge. We've shown that secondary nucleation and agglomeration are common mechanisms that often occur during seeding. Uh, careful observation and control of mechanisms such as this can really help improve performance, uh, especially at scale. And then finally, just a couple of other, uh, really useful seeding references that, uh, may be interesting, uh, to review. Thanks very much.
Effective seeding depends on the method and time of seed addition, quality of seed material, and the rate at which supersaturation is generated. Seeding is most effective if crystals are added as a seed slurry halfway into the Metastable Zone. The seed crystals should be uniform in size and provide smooth surfaces to activate recrystallization. Many recrystallization processes benefit from seed aging, where the temperature is held constant after seeding until supersaturation is depleted before continuing. Effective seeding can be accomplished using particle size analyzers.
Minimizing process time, energy consumption, and waste, optimal resource allocation and high process yield are important parameters to produce sellable crystal product at the lowest cost. Based on the solubility data one or more recrystallization methods (e.g. cooling, antisolvent, evaporative, reactive) are applied to reach an endpoint of high yield in the phase diagram.
Careful control over the degree of supersaturation and an understanding what particle mechanisms crystals go through is essential to design efficient recrystallization processes with optimal downstream performance.
Step 6: Solid/Liquid Separation
In most recrystallization processes, the solid particles are the desired product, which need to be separated from the mother liquor by filtration. The basic requirements for an efficient filtration process are:
Crystal suspension with a low number of fine particles
Suitable filter medium
Driving force (pressure or vacuum)
Filtration equipment (for collecting the fluid and retaining the filter cake)
After filtration, the cake is commonly washed with an easy to evaporate antisolvent to remove remaining mother liquor and to help with the drying process.
Step 7: Drying
The evaporative removal of moisture from the wet filter cake is often the final production step before a usable crystal product is obtained. Depending on the thermal and mechanical stability of the Active Pharmaceutical Ingredient (API) molecule and crystal, the solvent type and the risk of polymorphic transformation, a suitable drying method (atmospheric or vacuum) is applied.
Introducing EasyViewer for Recrystallization
Accelerate the Design of Fit-for-Purpose Particles
EasyViewer is a probe-based imaging tool that captures high-resolution images of crystals, particles, and droplets as they exist in process.
Exceptional information-gathering capacity combined with excellent usability makes EasyViewer a compelling tool that scientists will enjoy using to accelerate decision making and speed process development.
Below is a selection of recrystallization articles in publications.
Palbociclib Commercial Manufacturing Process Development. Part II: Regioselective Heck Coupling with Polymorph Control for Processability, Mark T. Maloney, Brian P. Jones, Mark A. Olivier, Javier Magano, Ke Wang, Nathan D. Ide, Andrew S. Palm, David R. Bill, Kyle R. Leeman, Karen Sutherland, John Draper, Adrian M. Daly, Joseph Keane, Denis Lynch, Marie O’Brien, and Joanne Tuohy, Organic Process Research & Development 2016 20 (7), 1203-1216, DOI: 10.1021/acs.oprd.6b00069
Operating Strategy to Produce Consistent CSD in Combined Antisolvent-Cooling Crystallization Using FBRM, Martin Wijaya Hermanto, Pui Shan Chow, and Reginald B. H. Tan, Industrial & Engineering Chemistry Research 2012 51 (42), 13773-13783, DOI: 10.1021/ie301626c
Furushima, Y., Kumazawa, S., Umetsu, H., Toda, A., Zhuravlev, E., & Schick, C. (2017). Melting and recrystallization kinetics of poly (butylene terephthalate). Polymer, 109, 307-314.
Pearsall, A., Hanson, K. M., Lad, S. P., & Salan, J. (2010). Synthesis, Scale-Up, and Recrystallization Studies of 2, 6-Diamino-3, 5-Dinitropyrazine-1-Oxide (LLM-105). Insensitive Munitions and Energetic Materials IMEMTS, 11-14.
Freitas, C., & Müller, R. H. (1999). Correlation between long-term stability of solid lipid nanoparticles (SLN™) and crystallinity of the lipid phase. European journal of pharmaceutics and biopharmaceutics, 47(2), 125-132.
In-situ FTIR spectrometers enable scientists to gain insight into their reactions and processes in a wide range of applications. Optimize reaction variables with inline FTIR instru...
Request Quote or InfoGet a QuoteInstant QuoteSolution InformationRequest a Demo
I want to…
Need assistance? Our team is here to help you achieve your goals! Speak with our experts.