OR WAIT null SECS
© 2023 MJH Life Sciences™ and Pharmaceutical Technology. All rights reserved.
The pharmaceutical manufacturing platforms of fluid-bed granulation and milling are widely used to modify particle size. However, the adoption of process analytical technology to monitor and control these processes is difficult because of their dynamic nature. This study examines the efficacy of a particle characterizing technology to capture particle images under dynamic conditions and to calculate particle size distribution data both in-line and at-line during fluid-bed granulation and milling.
Submitted: Feb. 18, 2014. Accepted: May 2, 2014.
Granulation methods are used in the pharmaceutical sector to enlarge and densify small powder particles into larger ones, typically to improve powder flow so that the material can be processed effectively and efficiently into solid dosage forms. In addition, granulation methods are important for “locking in” the API within the formulation, thereby reducing the risk of segregation. Granulated material is attained by direct size enlargement of primary particles or size reduction from dry compacted material (1).
Granule properties play a pivotal role in the final performance of the tablet; for example, granule size can affect the flowability and hence the average tablet weight and weight variation. The effect of granule size and size distribution on final blend properties and tablet characteristics is dependent on formulation ingredients, their concentration, and the type of granulating equipment and processing conditions employed. Granulation and sizing of granulation are therefore critical unit operations in the manufacture of oral solid dosage forms (2).
Granulation methods can be divided into two types-wet methods, which use some form of liquid to bind particles together, and dry methods, which do not use liquid (1).
The objectives of this study were to assess the ability to measure particle size distributions under static (offline) and dynamic (on-line) conditions for milled and fluidized particles as well as to determine the following:
Materials and methods
For the off-line analysis, aluminium oxide (Al2O3) ceramic microspheres (Brace GMBH) of size ranges 212–250 µm, 500–560 µm and 900–1120 µm were tested to determine the precision of the instrument. For in-line evaluation within a fluidized-bed chamber, a placebo of microcrystalline cellulose (Avicel) and lactose was granulated. Four granulation experiments were carried out in a laboratory scale fluid-bed granulator (Glatt, model GPCG15). The formulation comprised of 33.33% Avicel 101 (BASF), 66.66% lactose 200 (Meggle). The spray solution comprised of water and 5.5% PVP 90 (supplier BASF) with an addition rate of 220 g/min using a Watson Marlow 505S peristaltic pump.
Measurements were taken in real-time and the granulate growth per batch was evaluated. For the in-line evaluation of the rotary milling process, granulate that was processed by twin-screw granulation and fluid-bed drying was measured during rotary milling. During this study, granules of PVP (Kollidon 30, supplier BASF) were milled through mesh sizes of 2000, 1575, 1397, and 991 µm.
The imaging technology (the Eyecon particle characterizer) employed in this study is based on high-speed machine vision. It enables the capture of both size and shape of particles between 50 and 3000 microns. A continuous image sequence of the particles is captured using illumination pulses with a length of one microsecond for freezing the movement of particles which are moving at a speed up to several meters per second. The illumination is arranged according to the principle of photometric stereo for capturing the 3-D features of the particles in addition to a regular 2-D image. The particle size is calculated from the images using the 2-D and 3-D information, applying novel image analysis methods and direct geometrical measurement. As the approach is based on direct measurement instead of indirect, such as laser diffraction, there is no need for material based calibration. In addition, the method is non-contact and can be applied, for example, behind a view glass on a granulator without physical modification of the process equipment. Direct imaging technologies may be considered more accurate methods of measurement compared with laser and sieve analysis techniques because of the two dimensional method of measurement. These advantages of non-product contact and more accurate size estimation for in-line measurement are considered beneficial to the pharmaceutical development and manufacturing sector where reduction of cross contamination risks and accuracy of measurement are paramount. Figure 1 provides an example of an integrated technology on a Glatt fluidized bed granulator.
Results and discussionSummary of data from benchtop, fluidized bed and milling platformsExperiment 1: Benchtop evaluation. Figures 2a, b, and c show the size distributions from Al2O3-sintered spheres plotted against their certified size ranges. The red line indicates a normal distribution of the sample based on defined lower and upper range specifications to 3σ. Representative images for each distribution are also presented. During experiment 1, the at-line benchtop analysis using certified Al2O3-sintered spheres of known size distributions indicates a significant level of accuracy of the imaging technology. Figures 2a, b, and c indicate that a minimum of 91% of spheres were calculated as being within the pre-defined supplier certified specifications. The number of Al2O3-sintered spheres measured were 4572 (212 µm-250 µm), 440 (500 µm-560 µm), and 314 (900 µm-1120 µm). Because of the tight certified distribution range for these samples, it was deemed that a low sample population was acceptable. A larger data set may be required if greater variability is present within the sample set, for example, when evaluating granulates or milled materials.
Experiment 2: Fluidized bed evaluation. Figures 3a, b, and c show the D10 and D90 (i.e., the diameters at which 10% and 90% of the samples mass is comprised of smaller particles and mean particle distributions during fluid-bed granulation trials. All fluid-bed granulation trials were circa 40 minutes in duration. Graphs illustrating the D10, D90, and mean indicate the wetting, agglomeration, and drying phases that are typical of a fluid-bed granulation process. Other D values (D25, D50, and D75) were also tracked and similar profiles were identified. During the process, the technology successfully captured images, measured granulate size, and tracked a typical fluid-bed granulation process. The granulate size for all four batches reflected the typical granulate growth profile for a fluid-bed granulation process, including the wetting, agglomeration, and drying phases. The granulation growth trajectory (3) and end points (as illustrated in Figure 3a) could be monitored with this data. The variability between batches 1 and 4 versus 2 and 3 is most likely linked to variability in humidity levels during the trials as the dew point was not a controlled parameter and all other parameters remained fixed. These changes were confirmed through off line sieve analysis.
It can be seen from the particle size data illustrated in Figures 3a-3c that the dried granulate is of a smaller size compared to wet granulates earlier in the agglomeration process. This observation is to be expected when considering the removal of moisture and granule-granule attrition during the fluidized bed drying phase. All of the information presented in Figures 3a-3c could be used to develop, model, and control fluid-bed granulation processes.
Experiment 3: Rotary cone milling evaluation. Figures 4a, b, c, and d show the results achieved with the imaging technology during a rotary cone milling evaluation. The technology was capable of detecting the reduction in particle size of the milled material as the mesh size was reduced from 2000 µm to 991 µm as measured with D90 values. A positive correlation (weighted R2 = 0.8696) was calculated.
The reduction in particle size was also evident in the images that were captured during each milling run. A sample of images are included in Figures 4a-4d. There is a visible reduction of larger particles as the mesh size was reduced.
It can be seen in Figure 5 that the level of particle size variability reduces with the reduction in mesh sizes. Run 1 with a mesh size of 2000 µm mesh generated a D90 particle size range from 385 µm to 1196 µm (standard deviation = 199 µm). Run 4 with a mesh size of 991 µm generated a D90 particle size range from 427 µm to 707 µm (standard deviation = 69 µm).
The sample size was considerable with over 46,000 particles measured during each run, which was between 90 and 120 seconds in duration. The number of particles analyzed will vary depending on the speed of material flow, the volume of material transferred to the interface window for analysis, and the particle size range being analyzed. The density, hardness, and moisture content of particles were also considered during the development of the integration device to ensure a repeatable and representative sample for analysis.
The imaging technology (the Eyecon particle characterizer) successfully captured images and subsequently calculated particle size distributions for the sample materials in a rapid and accurate manner. As expected, there will be a degree of modification required as part of integration on new equipment types to ensure a representative and consistent sample presentation--which is critical to enabling accurate sample measurement. It can be concluded that the Eyecon particle characterizer can be successfully integrated within typical pharmaceutical fluid-bed granulators and milling environments in addition to accurately measuring samples in an at-line benchtop set-up.
1. D.M. Parikh, Chapter 1 “Introduction,” in Handbook of Pharmaceutical Granulation Technology, pp. 1-3 (Tailor and Francis Group, 2005).
2. G. Singh Rekhi and R. Sidwell, Chapter 17 “Sizing of Granulation,” in Handbook of Pharmaceutical Granulation Technology (Tailor and Francis Group, 2005).
3. J. Huang and M. Moshgbar, “Platform Technologies for Manufacturing Process Optimization through Integration of PAT and Control System,” accessed Dec. 18, 2014.
About the Authors
*I. Jones is founder and CEO of Innopharma Labs, firstname.lastname@example.org
Matti-Antero Okkonen is research team leader, VTT Research Centre of Finland
A. Greene is senior lecturer, School of Chemical and Pharmaceutical Sciences, Dublin Institute of Technology, Kevin Street, Dublin 2, Ireland
P.J. Cullen is senior lecturer, School of Chemical Engineering, UNSW Australia, Sydney NSW 2052 Australia.
Article DetailsPharmaceutical Technology
Vol. 39, No. 3
Citation: When referring to this article, please cite it as I. Jones, et al., “Monitoring Fluid-Bed Granulation and Milling Processes In-Line with Real-Time Imaging,” Pharmaceutical Technology39 (3) 2015.