Multiphase Bioreactor Design

Finally, Murhammer and Goochee (1990a) showed that certain small molecular
weight Pluronics, like Pluronic L35 and F38 offer just as good protection as Pluronic
F68, but result in less foam formation.
In conclusion, the presence of a stable foam layer from which cells rapidly drain may
be beneficial to the process, since in such a situation oxygen transfer is enhanced in the
absence of cell death and physical loss of cells. Serum, PVP and PEG are not favoured as
protective additives, since they result in increased foaming and cell entrapment.

Bubble Coalescence and Break-up

It has been suggested in the literature that the coalescence and break-up of bubbles in
sparged bioreactors may also be detrimental to animal cells (Oh et al., 1989, 1992; Wang
et al., 1994; Yang and Wang, 1992). Until now the only studies available on these
phenomena are done in sparged stirred-tank bioreactors (Gardner et al., 1989; Kunas and
Papoutsakis, 1990b; Michaels et al., 1995a; Oh et al., 1989, 1990; Wang et al., 1994;
Yang and Wang, 1992). In these reactors different situations may occur depending on the
sparging rate, sparger position and the Froude number (Oh et al., 1992):

(29)

where N (s−^1 ) is the agitation speed, Di (m) is the impeller diameter and g (m·s−^2 ) is the
gravitational constant. The Froude number represents the balance between centrifugal
inertial forces and buoyancy forces. In essence three situations may be discerned: (i) at
high Froude numbers, i.e. at low agitation rates and low aeration rates, air is entrained in
the trailing vortices of the impeller, where stable cavities are formed. Bubbles
continuously coalesce with the cavity and break-up from the cavity; (ii) at low Froude
numbers bubbles will be entrained in the vortex region but no stable cavities are formed.
The bubbles are accelerated and may break-up; (iii) the agitation does not interfere with
the bubbles.
As stated before, Wang et al. (1994) developed a generalised bubble break-up and
coalescence model, that views bubble escape at the surface as a special case of
coalescence. In this model the first-order death rate is proportional to the specific surface
area. They used the data of Oh et al. (1989) to verify their theory. Oh et al. (1989, 1992)
studied the reduction in growth rate of different cells as a function of gas flow, agitation
speed and sparger position. They found that at a fixed gas-flow rate cell death increased
with the agitation speed. However, since higher agitation speeds also gave rise to smaller
and, consequently, more bubbles, it remained unclear whether bubble coalescence and
break-up contributed to cell death. They concluded that in the absence of bubbles
agitation rates can be much higher than previously assumed. This was confirmed by the
study of Kunas and Papoutsakis (1990), who showed that in the absence of bubbles the
agitation rate could be as high as 800 rpm. Above this rate the eddy size became so small
that cells were damaged. Moreover, they showed that even in the presence of 5000 small
(50–300μm) bubbles per ml, but in the absence of a gas headspace, growth was not
affected up to agitation rates of 600 rpm. This suggests that bubble coalescence and
break-up does not significantly contribute to cell death. Using a bioreactor set-up
specially designed to measure effects of bubble coalescence and break-up in the bulk,

Multiphase bioreactor design 482