Scale-up: why bigger vessels are different physics

Doubling a bioreactor’s volume is not doubling anything that matters to a cell. The cell’s needs — oxygen per gram of biomass, removal of metabolic heat, a uniform chemical environment — stay the same per unit volume, but the processes that deliver them scale with the vessel’s geometry, and they scale differently from each other. Surface area per volume falls as the vessel grows, so a jacket that easily removes heat from a 10 L bench reactor struggles at 10 m³. Stirring that blends a bench vessel in seconds needs a minute or more in a large tank. The bench process was never “the process”; it was one point on a continuum of transport regimes, and scale-up moves you to a different point.

Gradients replace the environment

In a small vessel, mixing is fast enough that every cell sees the same concentrations — one environment. In a large tank, blending times approach minutes, and that is longer than the biology: cells consume glucose and oxygen faster than the tank averages them. The result is standing gradients — sugar-rich here, depleted there, oxygenated near the sparger, starved in the corners — and cells circulate through them on a loop. Each cell now experiences a fluctuating environment: feast and famine, oxygen excess and deficit, high CO₂ near the bottom where hydrostatic pressure pushes gas solubility up. Metabolic behaviour that never appeared at bench scale is not a mystery; it is the signature of inhomogeneity the bench vessel could not produce.

The oxygen budget rarely closes quietly

Dissolved oxygen in warm aqueous media is scarce — of the order of 0.2 millimoles per litre at air saturation — while a dense culture consumes it in seconds. Oxygen must therefore be transferred continuously from sparged gas, at a rate given by the volumetric transfer coefficient kLa times the gap between saturation and current concentration. kLa is bought with agitation and gas flow, and both have side effects: more power per volume means more shear and more heat to remove, and driving it harder strips CO₂ or foams the medium. At every scale the engineer is balancing a budget with three variables that all push on each other; scale-up changes the exchange rates between them, which is why a recipe that closed the budget at 10 L fails at 1000 L without redesign.

Shear, and what the intermediate regime is for

Animal cells have no cell wall; they are damaged by the tip speeds of large impellers and by the bursting of bubbles at the surface they ride up to. Shear tolerance therefore caps how hard the oxygen budget can be pushed, closing the circle of constraints. None of this can be computed with confidence from bench data alone, because the governing physics — turbulence, bubble behaviour, blending — changes regime with scale and is measured, not derived. That is the actual function of pilot and demonstration facilities: real hardware in the intermediate regime, where the gradients and the budget conflicts actually appear, so a process can be characterised as a design space — which conditions the product tolerates at which scale — rather than as a single recipe. The transfer package that leaves such a facility is a map of that space, and it is what makes the next vessel a calculation instead of a gamble.