(13a) Surface Velocity Measurements of a Powder Bed in a High Shear Wet Granulator Using High Speed Video Analysis

Current scale up strategies for wet granulation methods from lab scale (5 L) to plant scale (300+ L) normally involve maintaining a constant impeller tip speed and operating at a Froude number which ensures the granulator is operating in the roping regime. However, for highly soluble drugs, powder wetting and water distribution are the most important parameters. Therefore an alternative approach is to improve water distribution through the implementation of spraying apparatus and then to subsequently scale up using the spray flux density. The spray flux density is a dimensionless parameter which can be used to maintain equivalent liquid distribution across different scales of granulator. An important value used for calculating this parameter is the powder surface velocity, which is significantly lower than the tip speed of the impeller or the velocity of the bulk powder. This work tested the application of a high speed video camera running at up to 750 frames per second to film the movement of a large fraction of the area of the powder bed. A range of bowl sizes was studied, with various bowl fill levels using a lactose based placebo blend. To analyse the data, a combination of ImageJ with the TrackMate plug-in and Matlab were used. This allowed the determination of powder flow properties as a function of impeller speed and bowl fill in great detail. This approach was able to accurately measure powder bed velocity, using thousands of tracked particles. As a result, the transition from the bumping to the roping regime, with increasing impeller speed, was clearly seen as a distinct plateau in the data. It also revealed that the speed at which the transition between regimes occurred increased with increasing bowl fill. This suggests that the Froude number could be modified for use with wet granulators to take bowl fill into account. Further to this the video data was able to reveal the oscillatory nature of the velocity of the powder bed caused but the periodic movement of the impeller blades under the bed. It was also established from this information that powder movement varied distinctly as a function of powder bed location, both radially and with relative impeller position. The results were also compared against DEM simulations and other measurement approaches to demonstrate the accuracy and repeatability of the method. This work has demonstrated that the high speed video camera is a useful tool for monitoring the movement of powder in a wet granulator and the data can be used to assist in scale up. Moreover, careful analysis of the wealth of data provided will help to adjust system parameters as needed.