The Hel Crystalscan system has been used, in conjunction with the polythermal method, to create solubility and metastability design space information for two API compounds.
The resulting knowledge allowed crystallisation process selection to be based on process benefits such as favourable yields and conditions which would be beneficial for crystal growth.
Known concentrations of API and solvent mixtures were prepared within the Crystalscan system and heated to dissolution.
Variable cooling rates (between 0.1 to 0.6C/min) were used to quantify the shift in the metastable zone.
Solubility information was assessed by heating the obtained slurries at a rate of 0.1C/min.
Heating rates from 0.1 to 0.5C/min were examined, however, the low heating rate of 0.1C/min was deemed adequate to prevent any over estimation of the saturation temperature.
Up to eight different concentrations of API in solution were assessed for solubility.
Differing solvent/anti-solvent composition were obtained automatically by the attached dosing unit to the Crystalscan.
The outputted information was coupled with the Van't Hoff expression for solubility so the saturation temperatures for different concentration could be predicted over the entire design space.
This information, in addition to the nucleation temperatures, were then inputted to a Matlab program which allowed the production of the desired solubility and metastability design spaces.
In was concluded that the product x is not very sensitive to temperature changes in the region between 40 and 10C.
Although the process may exhibit low process yields (0.18g product x/g solvent) it can provide conditions that are favourable for particulate growth.
This is in comparison to the left-hand side of the design space (g anti-solvent/g solvent), which allows anti-solvent crystallisations to be assessed.
In this case the solubility curve is steep at isothermal temperatures.
A notable increase in process yield is observed for the anti-solvent (i.e.
0.38g product/solvent).
Although high in yield, the steep slope of the solubility curve indicates that the anti-solvent system would be an aggressive nucleator and a reduction in particle size, when compared to the cooling technique, may be evident.
The right-hand side of the design space, which represents the solubility of the compound with respect to temperature (cooling crystallisation), and the left-hand side, which represents the solubility with respect to anti-solvent concentration (anti-solvent crystallisations), the solubility slope with respect to each crystallisation technique is similar.
As a result of this, the crystallisation process can be decided on based on process benefits and costs.
In this example an anti-solvent crystallisation may be favoured due to the shortened time required to crystallise the batch or a cooling crystallisation favoured if anti-solvent cost are an issue.
In comparison to product x, this product should exhibit more favourable attributes for the control of anti-solvent crystallisation as its solubility, with respect to the chosen solvent, is relatively flat.
In addition to solubility information, the Crystalscan was used to generate a metastability design space for API y as concerns were raised that product nucleation would require metastable zones up to, and in excess of, 25C.
To combat the negative nucleation attributes these driving forces would bring, the process was designed to minimise driving force while maintaining sufficient crystallisation yields.
The metastable zone, or driving force, required to nucleate product y is high at low-solute concentrations, with values in excess of 20C typically observed.
On scale, this driving force may be exacerbated with the reduction in turbulent mixing which is often observed on scale-up.
In turn, supersaturation values may be high and lead to excessive product nucleation.
As a result of this analysis, a concentration of 1g anti-solvent/g solvent and 0.28g solute/solvent was deemed adequate as the system's metastable zone is approximately 10C.
Although seeding would be deemed favourable to prevent the exacerbation of the metastable zone on scale, it was not permitted.
In this case a slight increase in operating concentration from 0.12g solute/g solvent to 0.25g solute/g solvent was deemed adequate to control particle nucleation on scale.
The Hel Crystalscan turbidity-screening device, when combined with polythermal methods, enhances process understanding and enhances quality by design (QbD).
By gaining a understanding of the limitations and benefits of a solubility and metastability design space, a suitable crystallisation technique can be designed.
