Particle shape monitoring and control in crystallization processes using MD
Limitation of experimental techniques
for the on-line measurement of solid-phase properties have
restricted the development and implementation of optimal
design, monitoring and control methods for crystallizers and
other particle formation processes. Research on particle
shape is extremely important to industrial applications. For
example, in the pharmaceutical industry, morphology can
affect important properties such as dry powder density,
cohesion, and flowability, that can have major impact on a
company's ability to formulate drug particles into finished
products. Moreover, crystal morphology can affect drug
dissolution, potentially affecting finished product
performance and, in extreme, resulting in a companies loss
of the license to making the drg product. Experimental
measurement and control of crystal shape usually present
several difficulties. Particle shape measurements are
affected by unmeasured disturbance variables that cannot be
controlled. Very often the images are replete with bad
data. Particles sampled from a suspension crystallizer, for
example, can contain broken, agglomerated, aggregated, and
irregularly grown crystals besides the correctly grown
crystals. Particles fuse or particle boundaries overlap. It
is also difficult to obtain representative samples and to
sample enough images to remove the effects of noise through
averaging. In all cases the measurements are replete with
bad data and contain significant noise. Experiments show
that crystals can display different morphologies depending
on which faces grow faster. While it is not clear how to
control the rate of growth of each face, it is known that
the presence of certain impurities can affect crystal
morphology. A reasonable hypothesis is that impurity
molecules that bind preferentially to certain crystal faces
can be used to block or delay growth along those faces, thus
providing the means to control crystal morphology. While
some work has been done in this area, providing preliminary
confirmation of this hypothesis, at present the
pharmaceutical industry lacks the means to select
systematically the "impurities" and processing
conditions that would yield the desired morphology for a
given crystal. Thus, our goals are to understand the
fundamental phenomena that control preferential binding of
impurities to crystal faces, in order to develop methods for
controlling the morphology of crystals as they grow. We will
create a library of biocompatible "impurity"
molecules, and use Molecular Dynamics to simulate attachment
of these molecules to crystal seed faces in order to develop
a method for understanding a priori the factors that
control the crystal growth process. Binding energies will be
determined, and used to fine-tune methods and
models. Appropriate case studies will be selected in
consultation with industrial partners.
[ Top ]
