It was shown e.g that the maximal local energy
dissipation density ε is proportional to (T/D)3, i.e. becomes larger
with smaller impeller diameters, however, practical experience
tells that the available kLa data does not depict a dependency on
D/T e.g. At the first glance this fact is at least astonishing.
However, it should be taken into account that the task of impellers is not only to produce shear but also to provide some
pumping action to support mixing. Thus the energy input by the
impellers is channeled by part into establishing shear rates and by
part into flow rates. At the same total energy input the impeller that
makes more shear will make less flow and this contributes to kLa
as well, for instance through gas recirculation and a corresponding
increase in the gas holdup e.g. (Ezimora and Lübbert, 1997).
This finally leads to a balancing effect so that it becomes difficult
to distinguish the various impellers with respect the mass transfer
performance.
In short, the tasks of the impellers in aerated stirred tank bioreactors
is not only to provide enough shear to disperse the gas, they
also must induce flows that are necessary to distribute locally supplied
substrates and correctives around the entire culture. This also
applies for the bubbles generated at the impellers. Mass transfer
and mixing are often investigated independently of each other. This
can be made theoretically, practically they are intimately related
to each other and cannot really be separated. However, what effect
outbalances the other cannot be predicted a priori. The influences
of shear and flow on kLa do not necessarily compensate each other;
hence,in order to get a better insight,the measurement information
must be improved.
The free parameters of the correlations are known to be mainly
dependent of the composition of the liquid phase. Even in the early
work on theses correlations it was shown that the coalescence
behavior of the bubbles lead to a clear change in the exponents and
this is well known to depend of the composition of the liquid. Hard
physical models that relate the parameters to medium properties
are not yet available.
Many different measurement techniques have been established
to measure the kLa in gas–liquid reactors. Practically all measurements
of the oxygen mass transfer coefficients, however, have
been performed in model reactors with simple model media under
conditions that do not reflect the normal operation in living cultures.
As measurements in model media already showed, the liquid
composition, particularly the concentrations of surfactants may
have a dramatic influence on the mass transfer results. This led
to extreme variances in the published mass transfer rate data. Real
cultures contain various ionic and/or surface active substances that
do change the coalescence/redispersion behavior of the two-phase
gas-liquid flow considerably during a cultivation run. In a practical
fermentation run,the sensitivity to changes in the medium composition
may be larger than the influence of power draws what can be
seen by inappropriate additionof antifoam agents.Hence,measurements
are required in the original culture particularly since these
cultures may considerably changing their rheological properties
during a cultivation run.