Modelling Recrystallization of Amorphous Lactose
Typ
Examensarbete för masterexamen
Program
Engineering mathematics and computational science (MPENM), MSc
Publicerad
2021
Författare
Björling, Elsa
Modellbyggare
Tidskriftstitel
ISSN
Volymtitel
Utgivare
Sammanfattning
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 [1], 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.