Modelling Recrystallization of Amorphous Lactose
Examensarbete för masterexamen
Engineering mathematics and computational science (MPENM), MSc
When producing powders within the inhalable range, particles often need to be micronized to achieve a small enough size. During this process, amorphous content is introduced. The amorphous state is a higher energy state than the crystalline one, and therefore such content tends to recrystallize over time, especially if exposed to elevated temperature and/or humidity. During recrystallization of partly amorphous lactose, unwanted properties such as stickiness, caking and particle growth may occur. This thesis sets up a model for simulating recrystallization of partly amorphous lactose at different conditions. To be able to compare results with experiments, the model is created as to simulate the conditions of an isothermal microcalorimetry, TAM. The driving force for recrystallization is the difference between the glass transition temperature (TG) of the amorphous lactose and the operating temperature. In this work, the Gordon & Taylor equation is used to describe TG. Glass transition temperature is directly connected to moisture content and temperature of the powder, and so moisture sorption isotherms of the crystalline and amorphous content is of the essence. Guggenheim–Anderson–de Boer (GAB) equations with parameters fitted for the different states of lactose are utilized in this project to estimate saturated moisture content at different humidity levels. The crystallization kinetics used are the ones proposed by Bronlund , namely Avrami-Bronlund. The problem solved by the model is a set of partial differential equations for mass and energy balances. The results are compared with a series of TAM experiments, covering a range of T − TG at equilibrium. The model manages to cover the main phenomena of the sorption and recrystallization processes, and, when changing factors such as weight or relative humidity (RH), answers well to what is expected. The exact results are highly dependent on many model and material parameters, where the precise value of some are unknown. The experimental data show variation even for identical setups, and it is therefore difficult to tune the model. Whilst the general outcome of the model are good, it does not fully manage to capture the sorption and recrystallization behaviour. The sorption process appear differently in the simulations compared to the experiments, and the crystallization kinetics seem to be too slow at lower relative humidities (RH 53 and below). Keywords: amorphous lactose, DPI, recrystallization, powder bed, microcalorimetry, conditioning, annealing, pharmaceutical.