(350h) Isothermal By Design: Accelerated Nucleation of Aqueous and Non-Aqueous Solution Systems

1. Introduction

Nucleation is a phenomenon that can occur in a matter of seconds or a period of years depending upon the ease with which a system will crystallise. Conventional crystallisation methods do not allow for slow nucleating systems to be studied in a relatively short and useful timeframe, instead systems have to be left supersaturated for an inordinate time period or the nucleation events are not observed. Also, a vast number of pharmaceutical crystallisations are performed under high supersaturation conditions using antisolvent (drown out) crystallisation methodologies, but currently it is not easy to determine key kinetic crystallisation parameters at such high supersaturations prior to industrial scale crystallisations. Both of these issues are addressed and overcome in this work.

Isothermal by Design (IbD) is a methodology used for studying nucleation kinetics over a large range of supersaturations, utilising an antisolvent crystallisation procedure. It has been shown to achieve high levels of supersaturation that are unachievable through conventional crystallisation methodologies, whilst enabling key nucleation parameters to be calculated¹. A key aspect of IbD is its ability to take into account the enthalpy of mixing associated with adding an antisolvent to a solution through simple calorimetry calibrations, ensuring isothermal conditions are maintained.

This work focuses on results obtained utilising IbD for an aqueous pharmaceutically relevant system of p-aminobenzoic acid (pABA) in ethanol:water mixtures and a non-aqueous fuel and hydrocarbons relevant system, eicosane:additive (C20:A) in toluene and acetone mixtures. Nucleation kinetics for both systems were examined using classical nucleation theory (CNT).

2. Materials & Methods

2.1. Materials

pABA was supplied by Alfa Aesar, ethanol absolute and acetone were supplied by VWR, toluene was supplied by Honeywell, C20 was subpplied by Sigma-Aldrich, the additive was supplied by Infineum Uk Ltd. and deionized water was sourced on site at the University of Leeds.

2.2. Experimental Procedure

All experiments were performed on the Technobis Crystal16 system² at a 1 ml solution scale.

Calorimetry calibrations were performed in order to understand and take account of the enthalpy of mixing associated with antisolvent addition processes. Solutions of pABA in ethanol:water 70:30 wt% saturated at 293 K were held between 293 - 301 K in 2 K increments, containing various initial solution volumes from 0.2 - 0.8 ml. Water (antisolvent), held at different temperatures, was added to the solutions to make the final solution volume up to 1 ml. The exothermic spike in solution temperature was recorded, using a digital thermometer inserted into the solutions, and recorded. Final solution temperature was plotted as a function of antisolvent temperature for each volume studied and a linear regression was fitted through the points. Extrapolation of the linear regression to the initial solution hold temperature determined at what temperature the antisolvent must be to maintain solution temperature and ensure isothermal conditions were achieved throughout the experiments. For C20:A (99.9wt%:0.1wt%) in toluene solutions, saturated at 278 K, the same procedure was completed using acetone as an antisolvent, with solutions held at 283 K.

Induction time experiments were performed to enable key nucleation kinetic parameters to be calculated through CNT analysis. These experiments were performed in the same manner as the calorimetry calibrations, but the determined ‘isothermal’ antisolvent temperature was used for each temperature and initial volume of solution, no digital thermometer was present to ensure nucleation was not affected by the large temperature probe heterogeneous surface and induction times were determined through transmission data.

Data analysis was performed as outlined in a previous research article¹.

3. Results & Discussion

3.1. Calorimetry Calibrations

The antisolvent temperature was found to have a pronounced effect on the solution temperature of both the aqueous and non-aqueous systems. For both systems studied, at the solution temperatures used, antisolvent addition caused an exothermic effect, in which, the solution temperature was raised, even with a lower temperature antisolvent. Therefore, in order to ensure an isothermal system was used, antisolvent temperature had to be offset (lowered) to negate the exothermic enthalpy of mixing.

3.2. Induction Time

Induction times were reduced from around 10,000 s - 1 s, for the pABA system studied, over a range of supersaturations from 1.1 - 3.7. A similar range of induction times were probed for the C20:A system, but over a much smaller range of supersaturations, due to the metastable zone width (MSZW) of the C20:A system being much smaller than for the pABA system, 3 - 4 K in comparison to 4 - 8 K, respectively. The higher supersaturations achieved for the pABA system using IbD were much higher than those that could be achieved through conventional isothermal crystallisation methods, where a solution is cooled to a set supersaturation and the induction time to nucleation is measured. This is due to the initial cooling period required to reach a certain level of supersaturation, which limits achievable supersaturations. In comparison, for the pABA system, IbD achieved levels of supersaturation up to 5, but above 3.7 the system nucleated too fast to be detected by the instrumentation used.

The ability to accelerate the induction time by orders of magnitude with relative ease, for both aqueous and non-aqueous crystallising systems, demonstrates IbD’s efficacy with regards to accelerated nucleation testing.

3.3. Nucleation Kinetic Parameters

Values of effective interfacial tension (), critical nucleus radius () and number of molecules within the critical cluster () were calculated from the induction time vs. supersaturation data and CNT. For all solutions studied, and were found to decrease, in a non-linear fashion, as a function of supersaturation. Vales of showed strong compositional dependence, with large ranges of values of the mixed solvent systems studied as well as between the aqueous and non-aqueous systems.

The ability to determine nucleation kinetic information at high supersaturations for both aqueous and non-aqueous crystallisation systems is very important in an industrial setting, enabling accurate information about the system to be known and the operation conditions to be fine-tuned to improve crystallisation processes. Furthermore, this is extremely important for accelerated testing, where accurate kinetic data must be obtained at high supersaturations, so that accurate extrapolations of solution behaviour can be made to ‘real life’ solution conditions.

3.4. Nucleation Mechanism

The large range of supersaturations studied through IbD for pABA solutions enabled a nucleation mechanism change from heterogeneous nucleation (HEN) to homogeneous nucleation (HON) to be observed at a transition supersaturation of around 1.5. This was indicated through a sharp change in over values collected for a given solution temperature^3,4. No such mechanism change was observed over the range of supersaturations studied for the non-aqueous C20:A system.

The knowledge of nucleation mechanism changes within a crystallising system is very important in an industrial setting as it enables changes in operating conditions to account for potential beneficial crystallisation control.

4. Conclusions

IbD was used to study two contrasting crystallisation systems, an aqueous system of pABA in ethanol:water mixtures with water antisolvent and a non-aqueous system of C20:A in toluene with acetone antisolvent. Antisolvent calorimetry calibrations showed that the antisolvent temperatures used had to be reduced in comparison to the solution temperature, to take into account the exothermic enthalpy of mixing associated with antisolvent addition, and ensure isothermal conditions were maintained for induction time analysis. Induction times were reduced by orders of magnitude for both systems studied, with the higher supersaturations and lower induction times being accessible compared to conventional isothermal crystallisation methodologies. This is particularly important with regards to accelerated nucleation testing capabilities and probing nucleation kinetics at higher supersaturations. Key nucleation kinetic parameters were determined through induction time data and CNT, allowing a large range of solution conditions to be probed and understood, as well as demonstrating IbD’s ability to determine key nucleation information over a vast range of supersaturations. A nucleation mechanism change was also observed for the aqueous pABA system studied.

References

1) Kaskiewicz, P. L. et al. Isothermal by Design: An Accelerated Approach to the Prediction of the Crystallizability of Slowly Nucleating Systems. Organic Process Research and Development 23, 1948–1959 (2019).

2) Technobis. Crystal16. Available at: https://www.crystallizationsystems.com/crystal16.

3) Söhnel, O. & Mullin, J. W. A method for the determination of precipitation induction periods. Journal of Crystal Growth 44, 377–382 (1978).

4) Kim, K. J. & Kim, J. K. Nucleation and supersaturation in drowning-out crystallization using a T-mixer. Chemical Engineering and Technology 29, 951–956 (2006).