(203e) Evaluation of Experimental Methods for Solubility Determination of Carbamazepine in Ethanol for Cooling Crystallization Process Design 

Authors: 
Wu, W. L., Food and Drug Administration
Acevedo, D. A., U.S. Food and Drug Administration
Yang, X., U.S. Food and Drug Administration
O'Connor, T., U.S. Food and Drug Administration
Cruz, C. N., U.S. Food and Drug Administration
Solubility data is important for pharmaceutical formulation and process design. For example, solubility data is essential for estimating the supersaturation and understanding the relative stability of crystal forms. The determination of an accurate solubility curve is critical for efficient design of the crystallization process. Inaccurate determination can lead to use of a higher or lower supersaturation level that would result in undesired extra nucleation or crystal growth. However, there are various solubility measurement techniques. Here we will evaluate several popular ones and discuss how to obtain accurate solubility data.

In this study, carbamazepine (CBZ) is chosen as the model drug due to its various polymorphic forms. The goal was to evaluate the effect of (a) dissolving methods, (b) sampling separation, and (c) quantification technique on the accuracy of solubility data. Also, the solubility curve was developed for form III and II at various temperatures for future crystallization applications. The two dissolving methods examined for Form III CBZ are: (1) the shake-flask, (2) mechanical stirring. The shake-flask method was performed as the base scenario to examine solubility differences by two off-line sampling separation techniques: (1) centrifuging or (2) filtering. The concentration was determined by three quantification techniques: (1) off-line gravimetrically, or (2) off-line high liquid chromatography method (HPLC) or (3) on-line Raman method. In addition, mechanical stirring, filtering and gravimetrical options were chosen to measure the solubility data of form II for a comprehensive understanding of solubility curves of CBZ form II and III.

The mechanic stirring method demonstrated similar mean values but with less variance compared to the shake-flask method. Through analysis of means, there were no significant differences between using centrifugation or filtration as the sampling separation method. But it is noted that a minimum of 30 minutes is required for settling time to avoid significant effect on the equilibrium concentration. Between the two off-line concentration quantification methods (gravimetric and HPLC), the HPLC method always showed lower mean values, but their difference from the gravimetric values are < 4.6%; the standard deviation for gravimetric methods is ~ 1.5% while the standard deviation for HPLC is ~ 3%. The reason for the standard deviation difference is probably due to the multiple-step sample dilution procedure for HPLC. On-line measurements in mechanic stirring with Raman spectroscopy was done to determine the solubility curve of form III as a function temperature (5 - 35 °C). The solubility curve was compared to off-line concentration quantification methods, and the differences for the mean values are normally < 5%. This on-line Raman method allows for quick determination of solubility curve from a single experimental study, and could be used as a quick solubility method if a low accuracy is acceptable. In addition, the dissolution profile of the metastable form (form II) was analyzed between 5 °C and 35 °C in order to develop determine whether the solubility curve for CBZ form II and form III are mono-tropic or enantio-tropic. The results suggest that the system shows a mono-tropic system between 5 °C and 35 °C, while the solubility ratio decreases from 1.20 at 35 °C to 1.12 at 5 °C. This work presents a comprehensive study on the solubility data of CBZ and the solubility determination methods that are important during the process development and design of the crystallization process.

Topics: