| Cells per carrier | — |
| Cells per cm² | — |
| Cells per mL | — |
| Total cells to inoculate | — |
| Inoculum volume from stock | — |
| Attached after efficiency | — |
| Carriers receiving zero cells ? | — |
The green band is the workable window: fast enough to suspend the carriers, slow enough that the Kolmogorov microscale λ stays above half the bead diameter. Run at the bottom of it. Rafiq and colleagues found that operating at the minimum speed required to just suspend the carriers is what preserves the cells' critical quality attributes.
| Detachment agitation (5× culture speed) | — |
| λ during the burst | — |
| Detached cell diameter assumed | 15 µm |
| Detachment time | ~7 min |
| Expected harvest | — |
Multiply the carrier's specific surface area (cm² per gram dry weight) by the carrier mass, which is the load in g/L times the working volume in litres. Cytodex 1 is 4400 cm²/g and Cytodex 3 is 2700 cm²/g. At a typical 3 g/L load that gives 13.2 and 8.1 cm²/mL of volumetric growth area. A 2.5 L culture of Cytodex 1 at 3 g/L provides 33 000 cm², about 190 confluent T-175 flasks worth of surface.
Run at the minimum speed that just suspends the carriers, Njs. The Zwietering correlation estimates it, but the coefficient S is geometry-specific: published values are 4.8 for a 125 mL Bellco spinner and 9.2 for a DASbox with dual marine impellers. Calibrate S once by observing the speed at which your own carriers just clear the vessel base, then the correlation will scale correctly as you change load, bead type or volume. Operating above Njs adds shear without adding suspension.
Typically 3 to 6 cells per carrier for mesenchymal stem cells and 5 to 10 for continuous lines such as Vero or HEK293. The consequence worth watching is Poisson statistics. At a mean of 3 cells per carrier about 5% of beads receive none at all; at a mean of 5 that falls to roughly 0.7%. Empty carriers only become productive via bead-to-bead transfer, which published work describes as a nontrivial problem, so a low seeding ratio quietly costs you usable surface.
Damage is governed by the Kolmogorov microscale λ, the size of the smallest turbulent eddies. Once λ drops below about one half to two thirds of the bead diameter, eddies act across the carrier and strip attached cells. Because λ shrinks as agitation rises, every rpm past Njs moves you toward that limit. In one published HEK293T study on Cytodex 3, cells stayed attached at 250 rpm and were visibly sheared off at 300 rpm, which is where λ crosses half the bead diameter for that geometry.
Two steps: detach the cells, then separate them from the beads. A published scalable method raises agitation roughly fivefold above the culture speed with a dissociation reagent present, detaching cells in about seven minutes, then uses filtration to separate the single-cell suspension from the carriers. Harvest efficiency exceeded 95% and the cells stayed viable with their quality attributes intact, because a detached 15 µm cell is far below the Kolmogorov scale even at the elevated speed.
Divide the total carrier surface area by the flask area, 175 cm² for a T-175. Published work reported a single 5 L stirred-tank holding 2.5 L of culture producing as many hMSCs as about 65 fully confluent T-175 flasks in a robotic scale-out system. The ratio improves quickly with carrier load, because surface area scales linearly with g/L while the vessel footprint does not change.
At the cell densities and scales typical of current stem-cell microcarrier work, surface aeration through the medium is usually sufficient and sparging is avoided. Bubble rupture is more damaging than bulk shear, and the usual mitigation, Pluronic F-68, is harder to justify in a cell therapy process where the cell itself is the product rather than a secreted molecule. Raise carrier load and volume before you reach for a sparger, and if you must sparge, add it as a separate low-flow duty rather than increasing total gas throughput.