Computational fluid dynamics analysis of mixing and gas–liquid mass transfer in wave bag bioreactor

Abstract

Single use bioreactors provide an attractive alternative to traditional deep-tank stainless steel bioreactors in process development and more recently manufacturing process. Wave bag bioreactors, in particular, have shown potential applications for cultivation of shear sensitive human and animal cells. However, the lack of knowledge about the complex fluid flow environment prevailing in wave bag bioreactors has so far hampered the development of a scientific rationale for their scale up. In this study, we use computational fluid dynamics (CFD) to investigate the details of the flow field in a 20-L wave bag bioreactor as a function of rocking angle and rocking speed. The results are presented in terms of local and mean velocities, mixing, and energy dissipation rates, which are used to create a process engineering framework for the scaleup of wave bag bioreactors. Proof-of-concept analysis of mixing and fluid flow in the 20-L wave bag bioreactor demonstrates the applicability of the CFD methodology and the temporal and spatial energy dissipation rates integrated and averaged over the liquid volume in the bag provide the means to correlate experimental volumetric oxygen transfer rates (k L a) data with power per unit volume. This correlation could be used as a rule of thumb for scaling up and down the wave bag bioreactors.

CFD modelling of a wave-mixed bioreactor with complex geometry and two degrees of freedom motion

“…Thus, Liangchao et al (2018) determined up to 50% lower power inputs compared to the determination via torque (experimental and CFD). In contrast to stirred systems, where the power input can be assumed to be constant, the power input oscillates in 1 DOF wave-mixed bioreactors (Bai et al, 2019a;Svay et al, 2020). The same phenomenon was observed in the CELL-tainer bioreactor.…”

“…As expected, wavemixed systems lie in between. Svay et al (2020) posit that the highest energy dissipation values occur at the gas/liquid interface. In their investigations, the same behavior was observed, and hydrodynamic heterogeneities of Φ = 8.8 -32 were determined.…”

“…Similar to the specific power input, the k L a-value is also one of the most important key parameters in process characterization and scale-up. The determination of the k L avalue by CFD was carried out analogously to the method employed by Svay et al (2020) for a 1 DOF wave-mixed bioreactor. The specific interface a was determined using a contour surface (α f = 0.5).…”

“…Oxygen transfer is increased more effectively by increasing the rocking rate and the rocking angle than by increasing the aeration rate. Studies presented in the literature on the bioengineering characterization of wave-mixed bioreactors with 1 DOF and a maximum working volume of up to 500 L have contributed to their extensive use (Eibl and Eibl, 2006;Eibl et al, 2009a,b;Kalmbach et al, 2011;Löffelholz et al, 2013b;Marsh et al, 2017;Bai et al, 2019a,b;Svay et al, 2020;Bartczak et al, 2022). In addition, these bioreactors (e.g., HyPerforma Rocker Bioreactor, ReadyToProcess Wave Bioreactor, Biostat RM Bioreactor) are universally used.…”

Formulation Considerations for Autologous T Cell Drug Products

“…Mixing efficiencies of buffer components requires attention since the excipient concentrations must remain within the formulation design space defined during the pharmaceutical development studies. Computational fluid dynamics could be used to relate mixing efficiency to wave bag/chamber dimensions, fill volumes and rocking angle/stirring speed, building on recent simulations for fluid flow in wave bags and stirred tank bioreactors [ 27 , 28 ]. In practice, confirmation of the target DMSO concentration may be inferred by proxy through comparison of osmolality and T cell quality/quantity before and after the formulation step and post-thaw.…”