Oiling Out in Crystallization

Oiling out in crystallization is a process in which supersaturation causes the initially dissolved compound to separate from solution by creating a secondary liquid phase (emulsion) (image 1a) instead of a solid, crystalline phase (suspension) (image 1c). Oiling out mostly occurs when the integration of solute molecules into the crystal lattice is kinetically hindered, delayed or the system experiences very high supersaturation.

Solute molecules in the product-rich oiling out droplets can arrange themselves randomly and show much higher mobility than in a rigid crystal lattice. Often, this makes oiling out droplets a good solvent for unwanted impurities. Oiling out droplets are usually unstable and can transform to amorphous material (image 1b) over time, or spontaneously solidify into crystalline solid.

Oiling out is often not easily detected and can be missed, accounting for many process challenges during crystallization development and manufacturing.

Oiling out comes with a number of process challenges. In many cases, the final product depends on the starting point in the phase diagram and the evolution of this unpredictable particle mechanism.

• Oiling out is difficult to control as some systems first form small droplets which can coalesce later (image 2a), whereas others initially form large droplets which can disperse later (image 2b). Either way, oiling out droplet size distributions are highly dependent on mixing performance and hydrodynamics in the crystallizer. These circumstances can lead to unsuccessful scale-up when changes in vessel geometry and mixing performance lead to out-of-specification final product.
• The transformation of oiling out droplets to solid particles is often spontaneous and can bypass the desired impurity rejection process during nucleation and slow particle growth. Spontaneous solidification and a tendency for agglomeration (image 2c) can lead to undesired high impurity levels in the final crystal product.
• Going through an oiling out particle mechanism can lead to sticky, amorphous, gel or gum-like final product, which can be difficult to isolate or process. Common challenges are yield losses, slow downstream process performance or batch rework.

Identifying oiling out in crystallization requires the ability to robustly detect the undesired oiling out particle mechanism. Like nucleation, oiling out causes turbidity in a crystallizer and traditional methods (human eye or turbidity sensors) either miss the event, mistake one for the other, or struggle to distinguish between them when they occur simultaneously. Sampling an emulsion during oiling out is challenging because often the oil exists only for a short time before it solidifies. The change in process conditions during sampling can lead to coalescence, crystal nucleation, droplet dissolution or solidification, leaving a high degree of doubt and uncertainty about the results of any offline measurements.

This example shows how the combination of EasyViewer (iC Vision with droplets image analysis) and a chemical reactor (highly accurate temperature control) robustly detect and identify an undesired oiling out event (image 3a). The initial increase in turbidity coincides with a sharp increase in droplet count but does not show a nucleation exotherm. Trends, images and lack of a nucleation exotherm confirm the oiling out particle mechanism.

In the next step, the droplet count trend drops (image 3b), the turbidity trend increases further and the nucleation exotherm causes a heat-bump in the temperature profile. Trends, images and the nucleation exotherm are evidence that crystallization has occurred (image 3c).

In-situ video microscopy has proven successful to reliably differentiate between oiling out and nucleation in real time. The following control and optimization strategies can reduce the risk of oiling out or eliminate this undesired particle mechanism entirely:

• Utilize solubility and Metastable Zone Width (MSZW) data and avoid working at high supersaturation
• Generate supersaturation slowly by reducing cooling or antisolvent dosing rates
• Control the generation of supersaturation very tightly (very precise temperature and dosing control)
• Reduce impurity levels by utilizing pure raw materials
• Apply effective seeding strategy by adding seed crystals halfway into the Metastable Zone
• Perform mixing studies and mix adequately to avoid local high supersaturation

In Figure 3, an unseeded crystallization process is cooled down relatively quickly. The turbidity signal increases, EasyViewer identifies that turbidity is created by oiling out (not by nucleation) and a little later, the droplet solidification creates an exotherm.

In the application shown in Figure 4, the same process is initially cooled to a higher temperature (lower supersaturation) than that at which the oiling out occurred previously, seed crystals are added, followed by very slow cooling. The droplet count trend stays low and does not indicate oiling out droplet formation.

Seeding and slower cooling successfully prevented the undesired oiling out process and led to an in-specification final product.

A miscibility gap (or stable Liquid-Liquid Phase Separation (LLPS)) is an area in the phase diagram where a mixture of two or more components splits into two or more liquid phases, usually leading to an emulsion system.

Oiling out is usually a kinetic phenomenon (depending on cooling rate, supersaturation, nucleation kinetics, stirring and impurity levels). Reliable oiling out detection, for example with EasyViewer, an effective seeding strategy, and slow creation of supersaturation are often successful measures to avoid oiling out.

A LLPS due to a miscibility gap is a thermodynamic phenomenon (depending on system temperature, concentration and miscibility capability) and phase separation will occur independent of stirring, impurities, nucleation kinetics or rate of supersaturation creation.

Adding seeds to an oiling out process is usually a good measure to prevent oiling out. However, adding seeds into an emulsion caused by a miscibility gap will lead to seed dissolution and will not start the crystallization process.

The example on the left demonstrates that in most cases the only options to bypass the miscibility gap are to change the solute concentration, or to operate the system at a different temperature profile. At 25 % solids concentration, the temperature trajectory hits the miscibility gap an emulsion is formed (image 5c), which later solidifies (image 5d), leaving very little solution concentration for crystal growth. At 5 % solids concentration, the temperature trajectory does not hit the miscibility gap (image 5a). Crystals nucleate just below 0° C and grow to a large particle size.

METTLER TOLEDO offers a suite of tools to create a crystallization workstation, allowing scientists to obtain maximum scientific information from a single experiment in a centralized software suite. Process Analytical Technology (PAT) provides the following information:

• Automated Reactors: EasyMax, OptiMax, RC1 reaction calorimeters and RX-10 provide 24/7 precision control and recording of process parameters, including the capture of crystallization enthalpies so that scientists can confidently identify Critical Process Parameters (CPPs)
  
  
• EasyViewer: A probe-based imaging tool with image analysis to capture high-resolution images and calculate crystal, particle and droplet size distributions as they naturally exist in process
  
  
• ParticleTrack: Statistically, highly robust particle mechanism fingerprinting through particle size and count expedites successful scale-up from lab to fully ATEX rate plant scale
  
  
• ReactRaman: In-situ Raman spectroscopy provides molecular and structural information to help scientists understand complex polymorphic landscapes and how to select process parameters to always achieve the desired crystal form
  
  
• ReactIR: Real-time FTIR spectroscopy provides key information for concentration and supersaturation levels, metastable zone width, desupersaturation kinetics and crystallization endpoint so that crystallization processes can be run repeatedly to a desired target endpoint every time
  
  
• iC Software: Enables interoperability between all PAT tools so that all probes and reactors can communicate seamlessly and all analytical information (size, form, supersaturation, etc.) becomes a process control parameter

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.

In this webinar, a leading scientist from Eli Lilly explains their experiences with Liquid-Liquid Phase Separation and the undesired consequences, such as secondary nucleation, poor filtration performance, low crystal purity and unsatisfactory yield. Furthermore, a methodology is presented on how to systematically study oiling out phenomena to overcome these challenges and to potentially benefit from the oiling out effect. Two case studies describe solutions for:

• A spherical solidification and reduced surface area which led to secondary nucleation and poor filtration
• A strong degree of agglomeration after oiling out which led to a bad impurity profile and reduced yield

This presentation by Bristol-Myers Squibb explains how working at high solute concentrations may lead to metastable Liquid-Liquid Phase Separations (LLPS) in miscible solvent systems during batch operation. The talk describes potential side effects of this phenomenon as uncontrolled nucleation and growth, inadequate impurity rejection, solvent entrainment in the product and morphological inconsistencies in the solids.

A combined approach of Population Balance Modelling (PBM), Process Analytical Technology (PAT) and development of a continuous crystallization process was successfully able to overcome the challenges of oiling out, which assisted process engineers in creating a robust crystallization procedure.

• Adlington, N. K., Black, S. N., & Adshead, D. L. (2013). How To Use the Lasentec FBRM Probe on Manufacturing Scale. Organic Process Research & Development, 17(3), 557–567. https://doi.org/10.1021/op300326b
• Deneau, E., & Steele, G. (2005). An In-Line Study of Oiling Out and Crystallization. Organic Process Research & Development, 9(6), 943–950. https://doi.org/10.1021/op050107c
• Desikan, S., Parsons, R. L., Davis, W. P., Ward, J. E., Marshall, W. J., & Toma, P. H. (2005). Process Development Challenges to Accommodate A Late-Appearing Stable Polymorph: A Case Study on the Polymorphism and Crystallization of a Fast-Track Drug Development Compound. Organic Process Research & Development, 9(6), 933–942. https://doi.org/10.1021/op0501287
• Duffy, D., Cremin, N., Napier, M., Robinson, S., Barrett, M., Hao, H., & Glennon, B. (2012). In situ monitoring, control and optimization of a liquid–liquid phase separation crystallization. Chemical Engineering Science, 77, 112–121. https://doi.org/10.1016/j.ces.2012.01.047
• Zhao, H., Xie, C., Xu, Z., Wang, Y., Bian, L., Chen, Z., & Hao, H. (2012). Solution Crystallization of Vanillin in the Presence of a Liquid–Liquid Phase Separation. Industrial & Engineering Chemistry Research, 51(45), 14646–14652. https://doi.org/10.1021/ie302360u