Productivity in a CHO perfusion culture reactor was maximized when pCO
(2) was maintained in the range of 30-76 mm Hg. Higher levels of pCO(2
) (>105 mm Hg) resulted in CHO cell growth inhibition and dramatic red
uction in productivity. We measured the oxygen utilization and CO2 pro
duction rates for CHO cells in perfusion culture at 5.55 x 10(-17) mol
cell(-1) sec(-1) and 5.36 x 10(-17) mel cell(-1) sec(-1) respectively
. A simple method to directly measure the mass transfer coefficients f
or oxygen and carbon dioxide was also developed. For a 500 L bioreacto
r using pure oxygen sparge at 0.002 VVM from a microporous frit sparge
r, the overall apparent transfer rates (k(L)a+k(A)A) for oxygen and ca
rbon dioxide were 0.07264 min(-1) and 0.002962 min(-1) respectively. T
hus, while a very low flow rate of pure oxygen microbubbles would be a
dequate to meet oxygen supply requirements for up to 2.1 x 10(7) cells
/mL, the low CO2 removal efficiency would limit culture density to onl
y 2.4 x 10(6) cells/mL. An additional model was developed to predict t
he effect of bubble size on oxygen and CO2 transfer rates. If pure oxy
gen is used in both the headspace and sparge, then the sparging rate c
an be minimized by the use of bubbles in the size range of 2-3 mm. For
bubbles in this size range, the ratio of oxygen supply to carbon diox
ide removal rates is matched to the ratio of metabolic oxygen utilizat
ion and carbon dioxide generation rates. Using this strategy in the 50
0 L reactor, we predict that dissolved oxygen and CO2 levels can be ma
intained in the range to support maximum productivity (40% DO, 76 mm H
g pCO(2)) for a culture at 10(7) cells/mL, and with a minimum sparge r
ate of 0.006 vessel volumes per minute.