by 75%, meaning that about 900 cells were killed per bubble burst. This is somewhat
lower than the value of 1150 cells found previously, which may be caused by the fact that
not all liquid is collected and not all the cells that are affected by the bubble burst end up
in the upward jet. For instance, the downward jet may also kill part of the cells. However,
the variation in counts for both types of experiments was quite high and observed
differences may not be significant. The cell concentration in the upward jet was about 2
times the bulk concentration indicating adsorption of cells to the bubble surface. For a
total of 1400 cells this would mean adsorption of 700 cells per bubble.
Wen and Tan (1999) calculated the amount of CHO cells adsorbed per unit bubble
surface using a foam-flotation technique. The amount of cells transferred to the foam will
depend on the amount of liquid transferred to the foam, the cell concentration, the
drainage of cells from the foam, and the adsorption of cells to the bubbles. They showed
that the amount of cells transferred to the foam per unit bubble surface depended on the
height of the column and the cell concentration. In serum-free medium without
surfactants and at a cell concentration of 10^6 cells.mlâ^1 bubbles were saturated with cells
at a height of 0.1 m. Assuming all cells present in the foam had adsorbed to the bubble,
one can calculate that for a bubble with a diameter of 3.5 mm about 700 cells had
adsorbed at saturation. Considering the high standard deviations and the differences in
cell types and column height this agrees remarkably well with the value found by Trinh et
al. (1994).
Using the theory of Meier et al. (1999) one can calculate the amount of cells that
adsorb to the bubble during rise for the situation of Trinh et al. (1994) and Wen and Tan
(1999). This results in a value of about 20 and 40 cells per bubble, respectively, which is
substantially less than the experimental value of 700 cells per bubble. One explanation
for this difference is that the model of Meier et al. (1999) only describes adsorption
during bubble rise. Additional adsorption of cells may occur during bubble formation and
break-up. Michaels et al. (1995a) showed that for serum-free medium liquid drains from
the bubble film leaving the cells entrapped. With 300 cells present in the film (Trinh et al.
1994) this can at least in part explain the observed discrepancy. In addition, as
mentioned, during bubble rise the flow patterns may not always resemble potential or
creeping flow and cells may be captured in the wake of a bubble without having been in
contact with the bubble. Finally, Dey et al. (1997) proposed on the basis of numerical
simulation of the cavity collapse that the liquid flows occurring during the collapse might
result in concentrating the cells near the axis of symmetry, where the highest rates of
strain occur. In the presence of surfactant this concentration effect as well as the
magnitude of the strain rates was much less resulting in less cells being subjected to less
damaging flows.
Trinh et al. (1994) also studied the number of cells killed per bubble rupture and the
number of cells in the upward jet in the presence of 0.1% Pluronic F68. The number of
cells killed per bubble in their sparging experiments reduced to zero, while the
concentration of cells in the upward jet was on average slightly lower than the bulk
concentration. The reduced cell concentration in the upward jet is probably due to the
prevention of cell attachment to bubbles, the rapid draining of cells from the bubble film
in the presence of 0.1% Pluronic (Michaels et al., 1995b) and reduction of the
concentration effect as proposed by Dey et al. (1997). The reduced cell concentration is
also in accordance with the decrease in adsorbed cells found by Wen and Tan in the
Multiphase bioreactor design 478