Product temperature in a BFS process
One of the major challenges for the BFS operation compared with a conventional filling process is the high temperature involved
in the plastic extrusion step. As illustrated in Figure 1, the plastic granules are melted and extruded at temperature in
excess of 160 °C. The hot parison will come into contact with a metal mold, be transported to the filling shroud, and be filled
with drug product. Cool water circulating inside the mold helps to lower the ampul temperature. The cooling process, however,
is limited by the contact time between the ampul and the mold (a total of about 10 s). In addition, the ampul cooling is restricted
by the minimum temperature needed to form a hermetic seal at the end of the filling step. The ampul temperature near the neck
area has to remain high enough to ensure an appropriate seal at the end of the filling process.
In the development phase of a biological drug product, when using a plastic ampul as the primary container, it is important
to understand the temperature to which the product could be exposed. The folded conformations of proteins are only marginally
stable, and a rise in temperature can denature the molecule leading to loss of activity. Elucidating the temperature profile
during the BFS process will enable formulation scientists to develop more robust formulations.
In a BFS manufacturing process, product temperature during the filling process is not an easily controlled parameter, and
it cannot be directly measured. Instead, the controllable parameters are the starting-product solution temperature in the
bulk tank, plastic extrusion temperature, mold-cooling water temperature, and the cycle time for each BFS step (e.g., molding,
filling, and sealing).
The authors have used a computational fluid dynamic (CFD) model (ANSYS Fluent 12) to simulate the BFS process using the Pulmozyme
ampul as a model. Initial simulation results (not shown) showed that ampul-wall temperature reaches mold temperature quickly
(within 1 s) during Step 2 of Figure 1. If the actual mold temperature is at cooling-water temperature (usually near room
temperature of ~ 25 °C), the finished ampuls at the end of BFS process should not be significantly hotter than the mold cavity.
It is the authors' experience, however, that product ampuls usually feel warm right after the BFS process. This observation
suggests that the actual mold temperature is much higher than the cooling-water temperature (the actual mold temperature during
BFS production is not usually measured).
Figure 3: Simulation of product temperature in a blow–fill–seal process.
|
For the purpose of the CFD simulation, the authors adjusted the mold temperature to different settings. Figure 3 illustrates
the product-temperature profile when the mold temperature was set at 60 °C. The simulation assumes that it takes approximately
11 s for the entire BFS process including ~3 s for the molding step, ~3 s for filling, and ~5 s for forming a hermetic seal.
The upper portion of Figure 3 shows the liquid volume in the ampul, as it is being filled (from time = 3 s). The temperature
profiles of the entire ampul (including wall, solution, and air) throughout the filling and sealing step are shown in the
bottom sequence of Figure 3. It should be noted that in the actual manufacturing process, the filling nozzle is retracted
from the ampoule after filling is complete to allow for sealing. Since this is a nonproduct contact part the current CFD simulation
simply turns off the solution flow from the inlet nozzle, instead of modeling the actual nozzle retraction.
The simulation revealed that the ampul is cooled from the initial extrusion temperature of ~160 °C to 60 °C (assumed actual
mold temperature) in less than 2 s. Forced convection dominates heat transfer during filling (from 3 s to 6 s), thereafter
natural convection becomes more prominent. By the time the filled ampul was ejected from the mold (~ 11 s), only the fluid
layer near the ampul walls had been heated to within 10 °C of the mold temperature. The majority of the liquid remains considerably
cooler.
In the current simulation the authors assumed a worst-case scenario of liquid being filled at ~25 °C. In the actual production
process, the drug substance is typically maintained at 2–8 °C and would be introduced in the ampoule close to these temperatures.
Therefore, in a production BFS process, it is expected that the drug substance would experience temperatures that are lower
than these shown in the CFD simulation results.
Based on the observation of higher actual mold temperature compared with that of the cooling water from the simulation results,
it is highly recommended to monitor the actual mold temperature during the development or characterization phase of the BFS
process. Monitoring could potentially be achieved by attaching thermocouples to the mold or by infrared temperature measurement
of the mold surface.
|
