(402c) Foam Granulation Extrusion: A Novel Method to Continuous Wet Granulation of Powder Drug Formulations

The present work seeks to demonstrate a new continuous method for wet granulation using an aqueous foam binder to produce stable continuous operations in a twin screw extruder.  Furthermore, we endeavor to understand the influence that foam properties have on the performance of a binder during foam granulation.

EXPERIMENTAL: The study was conducted using α-lactose monohydrate (Pharma FlowLac® 100, Meggle) as a model excipient to be granulated in a co-rotating twin screw extruder (ZSE 27HP, Leistritz) with side stuffer at 10-40 kg/h and 220-320 RPM. The extruder barrel was controlled to 25^oC using a cooling system. An aqueous binder solution, composed of 4% (w/w) hypromellose (either METHOCEL™ A15PLV, E3PLV, E6PLV or E15PLV; Dow Chemical, Midland, MI), was foamed to 75-95% foam quality (FQ) using a mechanical foam generator and added at 6% or 11% (w/w) to the extruder. Foam properties were evaluated using two shear stability tests. Evaluation of granule properties was done by particle size analysis and determination of the characteristic fracture strength.

RESULTS: The initial studies to realize the method were done with A15 binder.  Using a side stuffer (auxiliary feeder affixed to the extruder) it was found that the foamed binder could be accurately metered into the extruder at all operating conditions.  A visualization window showed spreading of the foam over the surface of the lactose bed along an axial distance more than 4 L/D. To determine the value of the new method, wet granulation was compared by introducing the same aqueous binder solution at 6% or 11% (w/w) either as a foam (FA) or by direct addition using a pulseless metering pump (LI). The FA process demonstrated excellent stability over different flow rates (up to 40 kg/h) and screw speeds with no surging, attributed to uniform wetting of the powder; no segregated dry regions were noted, unlike by the LI process.  Exiting temperatures of granules varied from 23-40^oC (FA) versus 23-80^oC (LI). Final granule properties were shown to be comparable between the two methods of binder addition though not at similar operating conditions and fewer operating conditions produced good yields with the LI approach. Processing was only found possible for the drier formulation with 6% binder solution using the FA approach.

A second set of studies looked more directly at the influence of the foamed binder on the granulation process to better understand the mechanism.  In this case, granulation was done using binders of comparable surface tension but different molecular weights selected from the E-series of METHOCEL products. The half-life drainage time for the low MW binder, E3, was 9±3 min at 75% FQ (70 Pa-s), increasing to 23±2 min at 95% FQ (109 Pa-s); the value in brackets was the apparent viscosity of the foam to indicate its stiffness.  For the highest MW binder, E15, its foam half-life was 82±6 min at 75% FQ (103 Pa-s) and 165±12 min at 95% FQ (232 Pa-s). Under shear in a cylinder viscometer, the average collapse rate of foam (based on foam stiffness) was 60% greater using E15 than E3. In comparison, the average collapse rate was 168% higher with foams of 95% FQ versus 75% FQ. The total fraction of particles larger than 1180µm was a good differentiating response variable to determine effectiveness of granulation.  This fraction was nominally 54% (75% FQ) to 52% (95% FQ) for E3, and 42% (75% FQ) to 48% (95% FQ) for E15. Flow rate and screw speed were determined to be the most significant factors (P<0.05) affecting both particle size and granule fracture strength whereas binder MW had a minor influence on granule properties and foam quality showed no significant influence.

The negligible influence of foam quality on granule properties was reconciled with the earlier mentioned attributes of foam stability and visual observations of the process to present a proposed two-stage nucleation regime inside the extruder.   The talk will present this nucleation mechanism and how foam properties can control the spreading length of the binder inside the extruder.

CONCLUSIONS: A robust continuous process for wet granulation using a twin screw extruder was presented using a foamed binder. Lower material temperatures resulted and a broader set of operating conditions were possible using the foamed binder in comparison to direct liquid addition to the extruder.  The negligible influence of foam quality to granule properties points to a consistent wetting mechanism where the foam collapses at a similar spreading length regardless of foam stiffness as it is introduced to the extruder. This regulation of wetting rate is related to the environment of the extruder where foams of lower stiffness favor collapse over stagnant solids and foams of higher stiffness, near the high shear planes of the barrel.  In comparison to foam properties, operating conditions (flow rate and screw speed) most strongly affected granule properties suggesting a mechanical dispersion dominated nucleation mechanism.