This work has characterised the way wet granules break and deform during processing, and investigated the effect of granule properties and equipment parameters on granule breakage.
Three distinct sets of experiments were undertaken. In the first two sets of experiments, pellets of granular material were compressed in an Instron DynaMightTM load frame to characterise granule mechanical properties. The pellet deformation was filmed to link granule breakage behaviour with mechanical properties.
Cylindrical pellets (20 mm diameter, 25 mm height) of granular material were compressed axially at velocities ranging from 0.1 to 180 mm/s. Three sieve fractions of glass ballotini and two fractions of lactose powder were used to vary particle size and shape. Four binders were used to vary binder viscosity: water, lactose solution (for use with lactose formulations only), 1 Pa.s Silicone oil and 30 Pa.s Silicone oil.
When the results were plotted as dimensionless peak flow stress (Str*, defined below) against Capillary number (Ca), 2 regions appear: a Ca independent region and a Ca dependent region. Earlier work suggested a transition from semi-brittle to plastic granule deformation corresponding to the transition between Ca independent and dependent regions. This work has partly confirmed this observation with the use of Phantom 4 and 5 high speed cameras. The effect is stronger in spherical ballotini than angular lactose.
The effect of particle shape on peak flow stress was studied in detail. Using these experiments and literature data, the following empirical relationship was found to best describe the results.
.0 58 −3.4
Str* =(0.7 +Ca221 )AR Where Str * =σ pd32
where AR is the aspect ratio, σ is the peak flow stress, d32 is the specific surface
mean particle size, γ is the binder surface tension and θ is the surface-liquid contact angle.
Increasing the pore saturation of the pellets increased the dimensionless strength of the pellets. The relative effect of pore saturation changed with Ca, due to the transition between friction and viscous forces. At high Ca, as much as a four fold increase in peak flow stress was found when increasing pore saturation from 39 to 68 %.
The effect of Capillary number, pore saturation, pellet porosity, particle shape and particle distribution breadth were incorporated into the following empirical equation.
* =4346 Ca .0 401 .1 577 .2 059 −.2 250 S .0 799
Str s ε AR
where s is the liquid pore saturation, ε is the pellet porosity and Sp is the span of the particle distribution. This equation is applicable for Capillary numbers above 10-2.
Diametrical compression studies were also performed using the Instron DynaMightTM load frame. Pellets 20 mm in diameter and 10 mm high were compressed diametrically at speeds ranging from 0.1 to 180 mm/s. These experiments were also filmed with high speed cameras. The geometry of these pellets suited the study of granule deformation and mode of failure. Two types of granule deformation were observed: semi-brittle and plastic. Semi-brittle deformation occurred when a peak flow stress was observed at strains below 5 %. Plastic deformation occurred when no peak flow stress was observed, or the peak flow stress occurred at strains greater than 5 %. Capillary number alone could not predict failure mode. Non-spherical particles showed semi-brittle fracture at much lower values of Ca. It is likely that the failure
mode is determined by a balance between interparticle friction and viscous liquid dissipation.
In the final experiments, a Breakage Only Granulator was designed and built. The Breakage Only Granulator is a modified high shear granulator with interchangeable impeller blades to vary the shear and impact forces in the granulator. A powder bed of sticky sand (sand coated with 0.1 Pa.s Silicone oil) was used as the flow medium. Pre-formed granules were placed on the sand bed and mixed for 15 s, then sieved to recover the surviving granules and granule fragments. Two impellers were used in the Breakage Only Granulator Experiments. The 11˚ bevelled blade was designed to give a mixture of shear and impact forces in the granulator, and the frictional flat plate impeller was used to minimize impact and maximize shear in the granulator.
With the bevelled blade impeller, the amount of breakage decreased with increasing peak flow stress with some considerable scatter in the results. -20 �m ballotini showed different behaviour to the other granules, probably due to the small particle size and low pore saturation of the granules compared to the pellets used to measure strength in the Instron testing machine. There was considerably more granule breakage at 750 rpm impeller speed than at 500 rpm.
The frictional flat plate impeller at 500 rpm showed very little breakage for all formulations. At 750 rpm, only the weakest formulation showed considerable breakage. These results emphasise the profound effect of impeller geometry on the breakage process and suggest that impact is the primary cause of granule breakage in high shear mixers.