Scale-up challenge in reactive extrusion and compounding processes: 1D numerical simulation as a powerful tool

Reactive extrusion …

is more and more used for compounding high performance polymers which have complex formulations that imply potential compatibilization issues. Therefore, using twin-screw extruder (TSE) as a continuous reactor in which chemical reactions facilitating the compatibilization is a solution. The main issue could turn up when passing from a well-controlled formulation at lab scale to the production scale.
Indeed, scale-up from a lab TSE scale to an industrial one is one of the biggest challenge that the industry is facing for reactive extrusion applications.

In this study…

, the Ludovic© code, a 1D numerical simulation software dedicated to co-rotating twin screw extrusion process, is used as a tool for making the scale-up easier. There are scale-up laws existing such as constant SME (Specific Mechanical Energy) for example. But in some cases, this kind of extrapolation could present some deficiency. This study shows the powerful potential of 1D modeling to help scale-up in reactive process.
In this way, Ludovic© is used to optimize the process conditions to ensure the best scale-up product at the industrial scale.

Both extruders are Coperion’s co-rotating TSE. The lab scale TSE is a Ø26 mm with a throughput around 50 kg/h and the industrial scale is a Ø92 mm with a throughput around 1000 kg/h. The extruded polymer is a LLDPE described in term of viscosity by a Carreau-Yasuda law.

The first step is the validation of predicted results by comparison with the lab experiments. As we can observe on table 1, the global results (torque, SME, melting temperature) show good accuracy with the experiment.

Table 1In a second time, once the model is validated, Ludovic© draws time and temperature along the screw profile for the lab scale. On figure 1, we can observe temperature, residence time distribution and total residence time for different operating conditions at lab scale.

Figure 1

Well-known extrusion trends are confirmed, i.e temperature is increasing at high screw speed and high throughput, mean RTD, variance and total residence time are decreasing at high throughput.

Going further, then the software computes any known kinetic reaction. For example, we choose an Arrhenius law with a kinetic constant k following equation 1, corresponding to a peroxide decomposition kinetic.

Equation 1Taking into account that kinetic and previous computations of residence time and temperature, the Ludovic© software computes and displays peroxide decomposition curves on lab extruder at different screw speeds and throughputs (see figure 2).

Figure 2

Even if the average residence time is strongly reduced from 35.4 s to 8.2 s from 200 RPM to 1000 RPM conditions, the temperature at the melting zone exit strongly increases from 183°C to 231°C at the same time.
Ludovic© as a powerful and efficient tool for scale-up, predicts conditions on industrial extruder by fitting the total residence time and temperature computed curves.
This fitting is done with the help of Design of Experiments, which runs automatically hundreds of processing scenarios in few minutes.
The most equivalent conditions to 1000 RPM lab extruder are obtained considering temperature profile likeness as first criterion and residence time distribution equal or higher to the lab extruder’s one.

Figure 3

This analysis gives conditions at 840 RPM and 3000 kg/h on the industrial extruder. We can observe on figure 3, compared kinetics on lab and industrial extruders in extrapolated conditions.

This study showed …

that 1D numerical simulation can be a powerful tool to scale-up and optimize reactive extrusion process from lab to industrial extrusion lines. This tool gives good starting process conditions to prepare industrial test runs and save expensive tests.

*based on: «Scale-up tools in reactive extrusion and compounding processes. Could 1D-computer modeling be helpful?» by J-L Pradel, Quinebeche, P. Blondel (Arkema) C. David (Sciences Computers Consultants), (Proceedings of the Polymer Processing Society 29th Annual Meeting – 2013).

Leave a Comment