Influence of hydrodynamics and probe response on oxygen mass transfer measurements in a high aspect ratio bubble column reactor: Effect of the coalescence behaviour of the liquid phase

The usual gassing-in and gassing-out method was applied in a high aspect ratio bubble column using both tap water and coalescence-inhibiting liquid mixtures that represent the coalescence behaviour of biological media. The oxygen concentration curves were analyzed on the basis of different hydrodynamic models available from the literature, taking into account several parameters such as sensor dynamics, mixing in both phases, oxygen depletion in the gas phase or the axial evolution of pressure in the column. The sensitivity of K(L)a(L). values to the assumptions of each model was examined first in relation to physical considerations, but also as a function of the sensitivity of the optimization procedure. The results showed that the conventional method consisting in cutting the beginning of the oxygen response curves along with an assumption of perfect mixed reactor had to be avoided both in tap water and coalescence-inhibiting mixtures, except at low superficial gas velocity. They showed also that this simple model could not be improved by taking sensor dynamics into account. Conversely, several characteristic times and a dimensionless pressure ratio were established in order to define general quantitative criteria able to determine the best model for reliable K(L)a(L) estimation from dynamic oxygen response curves with the aim to contribute to design and scale-up of bubble column bioreactors. (C) 2007 Elsevier B.V. All rights reserved.