Top and Bottom Spray Fluid Bed Granulation Process

Use of Lasentec® FBRM® In-Process Particle Sizing PAT Technique to Study Top and Bottom Spray Fluid Bed Granulation Process

Presented November 10, 2005 American Association of Pharmaceutical Scientists (AAPS) Nashville, Tennessee

B. Smith, METTLER TOLEDO, Ben.Smith@mt.com		K. Walter, Niro Pharma Systems, GEA
E. Dycus, METTLER TOLEDO		M. Hubert, METTLER TOLEDO
A. Birkmire, Niro Pharma Systems, GEA

Purpose

Granulation of fine powders can be performed in a fluid bed processor by spraying solvent or a solvent/binder solution onto a fluidized powder bed (top spray granulation) or by concurrently spraying solvent or a solvent/binder solution into a segregated dilute phase powder stream (bottom spray or "GEA Precision Granulation™"). As the powder characteristics change, the process variables must be adjusted to account for changes in the product particle size, bulk density, and flow characteristics. Using Focused Beam Reflectance Measurement (FBRM®), the chord length distribution (CLD) of particles in a fluid bed processor can be precisely measured in real time without sampling or extracting product. The CLD measurement is a function of particle size, particle shape, and particle population. Obtaining this Process Analytical Technology (PAT) measurement in real time allows for continuous adjustment of process variables and detection of processing problems, and can be part of a process where product is suited for real-time release. Influences of spray rates, airflow rates, and filter cleaning can be quantified and optimized to improve product quality and yield. Combining this measurement with real-time microscopy, Particle Vision and Measurement (PVM®) allows real-time, in-line verification of the chord length distribution measurement.

Methods

A matrix of top spray and bottom spray fluid bed granulation experiments were performed. Critical parameters included:

• Liquid/binder spray flow rate
• Airflow rates
• Atomization pressure
• Method of spray addition - top spray versus bottom spray
• Changing excipient distributions

The in-line real-time particle distributions were characterized using in situ PAT, Lasentec® FBRM® and PVM® technology. Lasentec® FBRM® and PVM® were used to ultimately control the real-time particle distribution during the fluidized bed process.

Experimental description

• Probes used in the experiment:
  • FBRM® D600L - Focused Beam Reflectance Measurementprovides real-time, in-process measurement of particle dimension, number, and shape.
  • PVM® 700 - Particle Vision and Measurement (video microscopy)-provides in-process images of particle dimension and shape.

Figure 2: FBRM® and PVM® Installation Images-Top Spray

Figure 3: Bottom Spray-Precision Granulation™ Process

Figure 4: FBRM® Installation Images -Bottom Spray

Results

Figure 5: Tracking Bottom Spray Precision Fluid Bed Granulation

In the bottom spray granulation process, liquid binder and water are continuously sprayed during three stages: 135 g/min (spray 1), 145 g/min (spray 2), and 160 g/min (spray 3). Following moisture addition, the bed is dried. Using real-time PAT, one can track the number of fine particles in the range 0-100µm decrease in number over time as the number of coarse particles in the range 200-300µm and 500-600µm grow over time. The square weighted (volume based) mean eventually reaches a steady state. However, when filter blow back occurs during the third spraying regime, there is an obvious upset of fines seen by the increase in particle counts/second 0-100µm. The particle growth and fines upset were verified by the in-line PVM® microscope. Attrition and a corresponding increase in fines was noticed during the drying phase of the granulation.

Figure 6 depicts FBRM® distributions highlighting the beginning and endpoints of the bottom spray granulation progress. FBRM® square weighted volume distributions show the increase in granule dimension from spray one to the end of spray three. Statistics calculated from the FBRM® distribution quantify the changes that occurred.

Figure 6: Real-Time FBRM® Distributions-Bottom Spray Granulation

Statistic Spray 1 Spray 3 End Unweighted Median (1-1000µm) 44.3 54.3 Square Weighted Mean (1-1000µm) 190.1 285.1 Number of chords per second in range (1-300µm) 12231 7166 Number of chords per second in range (300-1000µm) 77 198

Real-time in situ PVM® images provide a quick snapshot and a true understanding of how the process is changing in real time. Images can quickly validate the FBRM® measurement and provide information on particle size, shape, and population.

Figure 7: Real-Time PVM® Images

Figure 8 shows the effect of atomizing pressure on CLD in bottom spray granulation. Higher atomization pressures produced finer granules. By changing the atomization pressure, one can adjust the droplet size and associated wetting behavior. FBRM® distributions taken at the end point of two bottom spray granulation batches quantified a 20µm difference in the square weighted (volume) mean. Spray rates were identical between batches.

Figure 8: Effect of Changing Process Variables: Atomization Pressurer

Experiment PG 04091701 PG 04091702 Spray Rate Low Low Atom. Pressure 0.25 Bar 0.5 Bar Mean Square Weight Microns 250.5 231.7

Figure 9 shows the effect of spray rate on CLD in bottom spray granulation. Keeping atomization pressure constant, one can look at the influence of spray rate on granule growth. As expected, higher spray rate produces coarser granules. The real-time response of in-line FBRM® can be used to quickly quantify granule growth and breakage kinetics. This can be used to provide feedback control signals to adjust critical control parameters.

Figure 9: Effect of Spray Rate on Bottom Spray Granulation

Experiment PG 04091701 PG 04091001 PG 04091501 Spray Rate Low Intermediate High Atom. Pressure 0.25 Bar 0.25 Bar 0.25 Bar Mean Square Weight Microns 207.9 242.7 279.7

Figure 10 shows the effect of atomizing pressure on CLD in top spray granulation. Higher atomization pressure produced finer granules. Significant breakage/attrition was observed during drying stage for top spray granulation.

Figure 10: Top Spray Granulation
Experiment Spray Rate Atom. Pressure Square Weighted Mean at Spray Endpoint Square Weighted Mean After Drying TSG 04090902 High 2.5 Bar 272.8 185.9 TSG 04090903 High 1.5 Bar 314.2 203.5

Figure 11 depicts the endpoint comparison between top and bottom spray granulations via FBRM® distributions. Bottom spray granulation had much shorter drying times. This corresponded to less attrition/breakage during drying. Thus, when comparing top to bottom spray granulation, it was easier to target the endpoint particle distribution with the bottom spray process. In this example, when comparing endpoints, bottom spray granulation targeted a larger endpoint mean compared to top spray granulation.

Figure 11: Top Spray Versus Bottom Spray Granulation - Targeting Endpoint Distribution

Experiment Top Spray Granulation 04091602 Bottom Spray Granulation 04091501 Atom. Pressure 1.5 Bar 0.25 Bar Mean Square Weight Microns 245.8 286.1

Conclusions

Figure 12: PVM® Images - Top and Bottom Spray Endpoints

Bottom spray endpoint

Top spray endpoint

• Top spray and bottom spray granulation are dynamically changing processes. Real-time in situ particle characterization with Lasentec® FBRM® and PVM® provides insight into the particle dynamics. This allows the critical process variables to be adjusted to achieve the desired downstream particle characteristics and avoid downstream processing problems. This real-time monitoring could ultimately lead to real-time particle distribution release.
• Attrition, a key upset during drying, can be minimized using real-time FBRM® to detect and avoid the process upsets. Bottom spray granulation also observed much less attrition than top spray granulation during the drying phase of processing.
• In bottom spray granulation, the spray rate has a much greater effect on granulation than atomizing air pressure in the ranges studied.
• In-line PAT instruments, such as FBRM® and PVM®, will allow users to hit target particle distribution specification and reduce or eliminate the number of failed or reworked batches.
• PAT will reduce analytical costs by in-process testing, improve batch-to-batch consistency, optimize and troubleshoot scale-up, and improve product quality.