Integrated experimental and numerical study of a bubble column bioreactor with immobilized <i>S. cerevisiae</i> for ethanol production in non-Newtonian fermentation broth

The design and scale-up of bubble column bioreactors (BCBs) for ethanol fermentation are constrained by non-Newtonian fluid behavior and mass transfer limitations, particularly when immobilized cells are employed. This study integrates experimental and numerical approaches to evaluate ethanol production using <i>Saccharomyces cerevisiae</i> immobilized in calcium alginate beads within a 500 mL lab-scale BCB. A multiscale simulation framework was developed, combining Monod kinetics (SciPy), glucose diffusion in spherical beads (FiPy), and shear-thinning hydrodynamics using a power-law model (FEniCS, <i>n</i> = 0.55). Model predictions were calibrated with triplicate batch fermentations over 48 h, achieving strong agreement between simulations and experiments (R² &gt; 0.93; RMSE &lt; 1.0 g/L). The experimental ethanol yield was 0.44 g/g glucose (8.8 ± 0.3 g/L ethanol), closely matching the simulated value of 0.45 g/g. Diffusion modeling revealed substrate depletion in bead cores, explaining the ethanol plateau observed experimentally, as internal glucose dropped below 0.2 g/L despite bulk concentrations &gt;1.0 g/L at 24 h. Predictive simulations indicated that reducing bead diameter from 4 mm to 2 mm improved core glucose availability (0.1 → 0.8 g/L at 24 h) and increased yield from 0.45 to 0.52 g/g. Scale-up simulations from 500 mL to 2 L predicted higher circulation velocity (0.012 → 0.025 m/s), but stagnant zones increased disproportionately (8% → 15% reactor volume) due to shear-thinning effects. These results demonstrate the value of integrating experimental data with multiscale modeling to enable mechanistic interpretation, predictive scenario testing, and rational scale-up of immobilized-cell BCBs under non-Newtonian conditions.