Drying of Pharmaceutical Powders Using An Agitated Filter Dryer

The drying of pharmaceutical powders following their isolation via crystallization and filtration is examined using a laboratory-scale agitated filter dryer (AFD). Vacuum contact and through-circulation convective drying with and without agitation of aspirin powders with pure and mixed solvents are studied through an integrated experimental and modelling approach. Experiments were carried out under different operating conditions by using aspirin particles and water/water-ethanol as typical solid-solvent systems. Two approaches were used to study the drying behaviour of aspirin: 100 µL small-scale drying using thermal gravimetric analysis (TGA) and 5 L medium-scale drying in a laboratory AFD. While the TGA tests provided a better understanding of drying kinetics as functions of temperature, vacuum level, solvent type and particle size, the AFD study revealed the effects of industry-relevant operating parameters (such as the heating rate, vacuum level, agitation speed and regime, gas flow rate, and initial solvent content) on the drying of aspirin particles. In the TGA experiments, the constant and falling rate periods are observed for drying with pure water, while only the falling rate period is observed for pure ethanol and water-ethanol mixed solvents. In the laboratory-scale AFD experiments, drying cycle time is found to decrease with increasing agitation speed, vacuum level, heating power supplied and gas flow. The critical solvent content is determined by identifying the transition point on the temperature-time and the torque-time plots during the drying process. Particle size distribution and morphology analysis of aspirin particles before and after drying with agitation indicate that the particles tend to agglomeration at lower stirring speeds while they are prone to attrition at higher speeds. An increase of the circularity of the particles with the increase of the agitation speed is observed due to attrition effect. Drying models based on a lumped-parameter (LPM) approach are developed for the static and agitated bed vacuum contact drying and combined LMP and distributed-parameter model (DPM) for the convective drying. Within this modelling approach, only one parameter that is the critical solvent content needs to be estimated from experimental results compared to other LPMs which greatly reduced the degree of parameter estimation. These computationally expedient models can be used for the prediction of drying time in a fast and reliable way. The model equations are implemented in gPROMS software and the models are applied to simulate the experiments carried out in the 5 L AFD. The vacuum contact drying experiments were also simulated using a DPM embedded in the gPROMS software. The modelling results of LPM were compared with that obtained using the DPM. Overall, a good agreement between the calculated and measured drying curves is obtained and the measured drying times are well predicted at varied drying conditions. This research presented the following improvements over the current studies on powder bed drying with limited research on organic crystalline materials and mixed solvent: holistic studies on drying behaviour with varied external drying conditions as well as internal material properties using different scale apparatus: TGA and AFD. Over the previous LPMs on powder bed drying, the modelling approach presented the following improvements: the models developed in this work need less parameter estimation. Three consistent approaches to measure the required parameter at the laboratory scale have been described in this work. The models provide a potential for industrial application on AFD drying.