(M1330-10-54) DSC vs. DVS: Determining the Moisture-Induced Tg vs. RH of Hygroscopic Materials

Purpose: Processing of hygroscopic materials requires an understanding of the effects of moisture content on bulk physical properties. While polyelectrolytes generally display no thermal transitions in their anhydrous state, a moisture induced glass transition state (Tg) may occur when moisture content exceeds a threshold value. Understanding the Tg as a function of water content is important to identify potential risk in the selection of outlet conditions for aqueous spray drying of hygroscopic materials. However, the Tg of hygroscopic polyelectrolytes can be difficult to measure with standard calorimetric techniques, such as differential scanning calorimetry (DSC), due to the convolution of the heat flow signal with concurrent processes such as thermal degradation or dehydration. Alternatively, changes in relative humidity at isothermal temperatures measured by dynamic vapor sorption (DVS) can induce a phase transition through plasticization of the material. This study compared the capability of DSC and DVS techniques for determining the moisture-induced glass transition and plasticization of GLY-200, a Phase 2 cationic polymer drug, to inform spray drying parameters.

Methods: DSC analyses were performed with a TA Instruments/Waters Discovery DSC 2500. Samples were equilibrated to a specific relative humidity for 24 hours prior to analysis. All samples were sealed with Tzero hermetic lids and analyzed in triplicate. The samples were heated at a rate of 2°C/min with temperature modulation of 1°C/min. TGA was performed with a TA Instruments Discovery Model 5500 employing a 2°C/min ramp rate from 25°C to 200°C. Water content for humidity-equilibrated samples was estimated by measuring the %weight loss up to 100°C. DVS analyses were performed with a Surface Measurements DVS Resolution, sample size 10-15 mg. For isothermal measurements, sorption and desorption data points were collected in increments of 10% relative humidity (RH) from 0%RH to 90%RH with a mass equilibration criterion of 0.002%. Three samples were analyzed at 25 °C, 45 °C, and 65 °C, respectively. Glass transitions were determined from the onset of hysteresis between the sorption and desorption cycles. A %RH ramp measurement was also conducted at 45°C by equilibrating the sample at 0%RH, increasing the humidity by 5% per hour from 0-90%RH, and finally holding at 90%RH. The glass transition was determined from the intercept of two tangential lines corresponding to the rate of sorption before and after the glass transition.

Results: DSC experiments demonstrated an inverse correlation of Tg with water content from samples analyzed as-received with 9% – 13% water, without any further equilibration. However, samples that were pre-equilibrated to defined %RH conditions showed no correlation between the water content and the Tg, measured by TGA and DSC respectively. In fact, all the pre-equilibrated samples exhibited the same severely depressed Tg at 45°C within the error of the measurement. DVS experiments with isothermal holds produced a strong inverse correlation (R² = 0.97) between Tg and water content over a broader range of 10% – 20.5% water content. The %RH ramp DVS method demonstrated good agreement to the isothermal method, both showing a water content of 17% at 45°C. In these samples, the %RH ramp method took about half the amount of time compared to the isothermal method (23h vs 40h), providing a faster alternative method with similar results.

Conclusion: Both the isothermal and ramp DVS methods yielded the expected inverse trend in the Tg vs RH. Glass transitions in the 25-65°C range were measured for samples having 10-20% water content and the Tg decreased proportionally with increasing water content. While DSC methods showed the expected plasticization trend for as-received samples, the equilibrated sample results were dependent on the method of sample preparation and measurement of water content. Conventional calorimetric techniques require the careful preparation of a series of samples with accurately known moisture content and add unnecessary complexities to the determination of Tg. Alternatively, the DVS experiments allow consistent sample preparation and do not require an independent measurement of sample moisture level. As a result, the DVS method can yield more precise results and better Tg values. During spray-drying, droplets of aqueous GLY-200 become wet solid particles. The development of surface stickiness of droplets during spray drying is correlated with Tg. Understanding the glass transition temperature as a function of water content is important to establishing spray-drying conditions for GLY-200 to minimize adhesion of material to the walls of the dryer which results in reduced run times and low yields.  This study successfully identified the processing zone associated with moisture-induced stickiness for GLY-200.

References: (1) Burnett, D. J.; Thielmann, F.; Booth, J. Determining the Critical Relative Humidity for Moisture-Induced Phase Transitions. International Journal of Pharmaceutics 2004, 287 (1–2), 123–133. https://doi.org/10.1016/j.ijpharm.2004.09.009.
(2) Abiad, M. G.; Carvajal, M. T.; Campanella, O. H. A Review on Methods and Theories to Describe the Glass Transition Phenomenon: Applications in Food and Pharmaceutical Products. Food Eng. Rev. 2009, 1 (2), 105–132. https://doi.org/10.1007/s12393-009-9009-1.

Tg vs %water with DSC methods (green) and DVS methods (blue).

DVS isothermal method at a) 25°C, b) 45°C, and c) 60°C showing the hysteresis between the sorption and desorption curves at a mass change, or water content.

DVS ramp method showing tangential lines fits to determine a %mass change, or water content, of 17% at 45°C.