(M1530-06-33) Dimensional Conformity of Filaments in Hot-Melt Extrusion: Design of Experiment and Process Optimization

Purpose: Hot-melt extrusion (HME) is a widely used technology for the formulation of poorly soluble drugs. In HME, the drug is incorporated and stabilized in a polymer matrix through mixing in the extruder barrel at elevated temperatures, preferably in an amorphous state. A homogenous extrudate or filament diameter is an indicator for a steady process and can be a critical quality parameter, e.g., in the production of uniform pellets [1]. To better understand the mechanisms affecting filament diameter fluctuations, a design of experiment (DoE) was performed to assess the influence of process parameters and optimize the process to achieve highly uniform filament.

Methods: Polyvinyl alcohol (PVA, Parteck® MXP, Merck KGaA, Darmstadt, Germany) was used for the DoE. Extrusion was performed on a twin-screw extruder (Pharma 11, ThermoFisher Scientific, Waltham, USA) equipped with a 2 mm die. A screw configuration with three kneading elements was used and heating zone 1 was set to 75 °C, zone 2 to 140 °C and zones 3 – 7 to 210 °C. Gravimetric feeding was performed using a Congrav OP1-S feeding system (Brabender GmbH, Duisburg, Germany). Process data were recorded during extrusion by the provided data logging software of the extruder. A full factorial DoE was performed with three repetitions at the center points. The parameters investigated regarding their influence on filament diameter fluctuations, expressed as coefficient of variation (CV), are displayed in Table 1. Data were analyzed using MODDE (version 12.01, Sartorius, Göttingen, Germany). Filament diameter was adapted using a conveyor belt (Brabender GmbH) by adjusting the conveyor belt speed in a manner that the filament diameter was 1.75 mm, which was measured with a three-axis laser measurement system (Odac 33 Trio, Zumbach, Orpund, Switzerland). Therefore, conveyor belt speeds had to be adapted to the respective throughput and was set to 1.6, 2.4 and 3.1 for throughputs of 0.2, 0.3 and 0.4 kg/h, respectively. The melt viscosity of neat PVA was determined using a plate-plate rheometer (Haake™ Mars™, Thermofisher, Waltham, USA). The sample was heated up with constant oscillation (1 Hz) from 170 °C to 250 °C at a rate of 2 K/min (n = 3). The gap between the plates (Ø = 25 mm) was 1 mm. Differential scanning calorimetry measurement were performed on a DSC 3+ (Mettler Toledo, Columbus, USA). Approximately 5 mg were heated and cooled in sealed aluminum pans with pierced lid at 30 K/min from 25 – 210 °C down to -10 °C and up to 250 °C.

Results: Statistical analysis in Modde showed that the model had a good fit, high predictability and high reproducibility (values > 0.9) and showed no significant statistical problems. The main factors affecting filament diameter fluctuations were throughput and die temperature as displayed in the contour plot (Figure 1), of which the throughput showed the highest influence. No significant influence of the screw speed was observed. The filament diameter fluctuations increased with increasing throughput due to discontinuous material transport. The log data indicated that the actual feed rate was consistent. It was observed, though, that at high throughputs, the screws in the feeding area were not capable of pulling the material consistently into the extruder barrel, leading to periodic fluctuations and CV of the diameter up to 10 – 12%. In addition, the die temperature showed a small but significant influence on the diameter fluctuations. Higher temperatures of the die also resulted in higher fluctuations. PVA is a semicrystalline polymer and showed a melting point 195.0 °C (± 0.5 °C) and accordingly a severe drop of complex melt viscosity starting at approximately 200 °C. (Figure 2). It is likely that, due to the low melt viscosity of the extrudate strand exiting the die, the dimensions are affected to a higher extend by volume fluctuations in the barrel. A low met viscosity of extrudate is also likely to result in flattening of the extrudate upon contact with the conveyor belt.

Conclusion: Extrusion parameters highly affected the dimensional quality of filaments. The throughput was identified as the most influencing factor, highlighting the importance of consistent feeding but also consistent transport of the material in the extruder. Even though it was possible to produce very homogenous filament, this was possible only at low throughputs. Further process development is necessary in order to reduce process times and ensure high quality of filaments.

References: [1] Rehrl, J., Kirchengast, M., Celikovic, S., Sacher, S., Kruisz, J., Khinast, J., Horn, M., 2020. Improving Pellet Quality in a Pharmaceutical Hot Melt Extrusion Process via PID Control and LOLIMOT-Based MPC. J. Pharm. Innov. 15, 678–689.


Table 1: Process parameters for DoE


Figure 1: Series of contour plots at different die temperatures displaying the filament diameter fluctuations expressed as CV (%) in color as a result of the factors throughput and screw speed


Figure 2: Complex melt viscosity of PVA ± SD (n = 3)