Kinematics and collision kinetics in high shear granulation

High shear granulation (HSG) is an important unit operation in the Pharmaceutieal,
Detergent and Food industries. The desired output of HSG is generally a specified
granule size distribution (GSD) and bulk density of the granular product. Depending on
the final application, homogeneity, strength and internal structure of granules also need
to be of desired quality. In industry this operation suffers from problems such as poor
reproducibility and high rejection rate. Knowledge in this area is mainly empirical for
both operation and scaling up. A generalised universal approach needs to be developed.
This is possible if the rate processes in granulation are understood.
In this dissertation a new approach to granulation is suggested as follows. The operating
conditions and properties of the agglomerating substances dictate the bulk flow and thus
the particle relative motion responsible for particle-particle collisions. The particleparticle collisions result in aggregation, breakage or rebound of particles. So only a
fraction of colliding particles will result in aggregating successfully. If the relationship
between these individual processes is understood, given the initial conditions, it would
be possible to predict the output.
A novel cost-effective technique to measure surface velocities in HSG has been
developed by combining High Speed Imaging and Particle Image Velocimetry. This
technique resulted in efficient characterization of the mean and the particulate flow in a
vertical axis high shear mixer with high temporally and spatially resolved velocity data.
The equipment used for HSG is referred to as high shear mixer (HSM) in the chemical
industry. The experimental mixer (Roto Junior, Zanchetta, Italy) was a 10 litre, vertical
axis HSM with a provision to change the mixing blade. The test feed material was
Calcium carbonate (Durcal 40) and binders used were Polyethylene Glycols (PEG, avg.
molar masses 400, 1500 and 4000 g/mol) and Glycerol. Effect of operating conditions
and material properties on the flow has been established in the high shear mixer
equipped with a 3-blade impeller. The average bed velocities were found to increase
with increase in impeller speed, although approaching a stable value at higher impeller
speeds. A higher viscosity binder considerably slowed the bed. Also it was noticed
during a wet granulation (granulation wherein a liquid binder is used) experiment that the bed velocities were a function of availability of surface binder and the average
granule size. The results were found to be highly reproducible for dry granules.
A disc impeller was used for further experiments with dry granules mainly to simplify
the mean flow in order to extract velocity fluctuations to calculate collision frequencies.
The Kinetic Theory of Granular Flow (KTGF) was found to predict particle collision
frequencies three orders of magnitude higher than those predicted by the Shear theory.
This observation with disc impeller experiments provided a basis to perform further
experiments with the 3-blade impeller.
In a wet granulation experiment with the 3-blade impeller, simultaneous observation of
particle dynamics and GSD was made. By extracting temporal velocities from a single
interrogation area, granular temperature and subsequently collision frequencies based on
the KTGF were calculated. A discretised population balance equation was fitted to the
GSD data by minimising overall sum of square errors to calculate the aggregation rate
constant. From the knowledge of the collision rate and the aggregation rate, aggregation
efficiency was estimated to be of the order of 4 in 100 million in the initial stages of
granulation with gradual decrease in this value as granulation proceeded. This method
of extracting particle collision rate, aggregation rate and estimated aggregation
efficiency provides a basis for predictive high shear granulation model using population
balance modelling and knowledge of particle dynamics.