(T1030-11-63) Twin Screw Extrusion Process Design and Die Modeling for Directly Extruded Implants

Purpose: Developed approximately 100 years ago for natural rubber and plastics applications, processes via twin screw extrusion (TSE) now generate some of the most cutting-edge drug delivery systems available. Progress has been made toward improving direct and continuous extrusion of ocular, hormonal, and multi-layer implant devices. The typical small batch sizes associated with several implant devices (50 g/batch) can be processed with approximately 90% yield utilizing a plunger feed method. With the addition of a gear pump at the end of the process section of the TSE, the previous challenge of producing dimensionally stable parts is mitigated. Coextrusion is the simultaneous extrusion of two or more different, yet compatible, materials through the same die to produce a single product with multiple layers in a uniform consistent shape. Detailed modeling of these multi-layer structures dramatically improves the dimensional stability, shear stresses experienced by the material, and layer stability through complex downstream flow paths.

Methods: Various TSE’s were used to develop the characteristics required for directly extruded implants including a Nano 16 mm (co-rotating, 25:1 L/D, 1 cc/dia, 1.2 OD/ID), ZSE 18 mm HP (co-rotating, 25:1 and 40:1 L/D, 3.2 cc/dia, 1.5 OD/ID), and a ZSE 27 MAXX (co-rotating, 40:1 L/D, 14.3 cc/dia, 1.66 OD/ID), 14.3 cc/dia free volume). 5 & 10 MFI polypropylene were used for residence time evaluation on the 27 mm and Copovidone was used on the 18 mm. HPMC-AS + 40% poorly soluble drug was used on the 16 mm to determine waste and minimal batch size required. PLGA and a Zenith Pep-II, 1.5 cc/rev, gear pump at the end of the 18 mm processing section were used to evaluate the dimensional stability of implants. To fully understand the implications of how the material is being provided to the die, detailed flow modeling was performed on a multilayer structure by Compuplast.

Results: Calculations (Eq. 1 – 3) were used as a baseline to understand what type of residence times should be expected at various rates, rpms, and screw designs. This is helpful to understand thermal exposure of any potentially sensitive formulation components. Material was then processed using the 27 mm and 18 mm TSE’s to compare the calculated time to the actual. Overall, the results match with a largest difference between calculated and actual of 13 seconds (figures 1 a-d). Using flow rates of 2, 4, and 8 cc/min, 50 gram batches were processed on the Nano 16 mm using the plunger feeder (figure 2a). Out of these 50 grams, 44 were collected and determined to be usable, an 88% yield. The 6 grams that were lost were: 1 gram at the extruder/plunger interface (figure 2b), 2-3 grams on the screws (figure 2c), and 2 grams in the die/front end adapter. With the addition of the Zenith PEP-II gear pump to the end of the extruder, the dimensional stability of the PLGA implant increased dramatically. Without the gear pump the dimensions were 1.2 mm ± 0.3 mm (figure 3a). With the gear pump the dimensions were 0.75 mm ± 0.03 mm (figure 3b). A crosshead coextrusion die was modeled using PLGA the inner layer and EVA as the outer layer as model polymers. Flow geometries, spiral design, pressure drop and shear stress and shear rate in the die are analyzed for their impact on residence time and distribution as well as the overall product layer uniformity and potential surface defects caused by melt fracture. Initial results indicated a layer distribution problem, most likely related to the large viscosity difference between the layers (figure 5 a – d). After adjusting the die design to account for the viscosity issues, the layer distribution became uniform (figure 6 a – d).

Conclusion: After working through some common implant processing challenges of thermal exposure, minimum batch size, material waste, dimensional stability, and die design, solutions were developed to create a directly extruded usable product. Residence times were calculated to reflect the thermal exposure of a formulation. A plunger feed method and a shallow flight depth were used to minimize batch size and material waste. A gear pump was added to the end of the TSE to dramatically increase dimensional stability. Flow modeling analysis was performed to ensure the dimensional and layer stability through complex downstream flow paths. Pairing these techniques with the processing and screw design variables of the TSE, a robust, repeatable, and dimensionally stable implant device can be produced.


Equations 1 - 3, Figure 1a - d.


Figure 2a - c, Figure 3a - b, and Figure 4.


Figure 5a - d and Figure 6a - d.