from the analogy with the g-l data, extend the region of the homogeneous regime
stability. Subsequently, the performance of multistage slurry reactors in the homogeneous
and the heterogeneous bubbling regimes should be compared, regarding the appropriate
hydrodynamic, mixing and transport characteristics. Attention should be also paid to
possible differences in the effect of solid particles on the values of such system
characteristics, reflecting a different extent of micro-scale turbulence under the
homogeneous and heterogeneous bubbling conditions.
In general, the flow pattern studies should be extended from the pseudohomogeneous
slurries to those with particles possessing non-negligible slip speeds, i.e. to the systems
exhibiting different flow patterns for the liquid and solid phases and/or solids
concentration profiles in individual stages and along the reactor height. Both small
particles with large density difference from the liquid phase and large particles with the
density close to ρL should be employed in experiments to simulate conditions in the two
most common industrial applications of g-l-s reactors—catalytic reactors and bioreactors.
Further attention should be paid to the effect of solids loading and particle size and
density on the rate of gas-liquid mass transfer in g-l-s systems and on the values of
respective parameters of three-phase beds (gas holdup, specific interfacial area, liquid-
side volumetric mass transfer coefficient, kLaL). Analysis of related literature on single-
stage bubble column reactors (Brück and Hammer, 1986; Pandit and Joshi, 1986;
Kratochvíl and Kaštánek, 1989; Khare and Joshi, 1990; Wilkinson et al., 1992) indicates
the extreme complexity of this issue, as well as sometimes contradictory findings
reported by various authors. Additional experiments are needed to allow reliable
prediction of the solid phase effect under complex conditions of simultaneous influence
of liquid phase properties and/or gas dispersion mode. While there seems to be a broad
consensus in the literature on the positive effect of fine particles with adsorbing capacity
(Alper et al., 1980; Pal et al., 1982; Pandit and Joshi, 1986; Kratochvíl and Kaštánek,
1989) and, on the other hand, of large particles with diameter above 1 mm (Ostergaard
and Fosbol, 1972; Sittig, 1977; Khare and Joshi, 1990), contradictory information found
in the literature on the effect of fine inert particles implies different possible mechanisms
of their influence on bubble coalescence. To clarify this issue, coalescence studies should
be performed in a coalescence cell (Zahradník et al., 1995), ensuring clearly defined
hydrodynamic conditions (contacting of two bubbles in a pseudoinfinite medium), in a
wide range of liquid and solid phase properties and solids concentrations. In addition,
experiments in multistage bubble columns should be aimed at linking the fundamental
knowledge of the gas-liquid mass transfer in microparticle slurry systems with the
specifics of sectionalised units (e.g. hindered establishing of equilibrium bubble size in
coalescing systems due to gas redispersion by the internal plates and low bed height in
individual stages).
Experimental programmes in the coalescence cell as well as in the slurry reactors
should include measurements in non-aqueous media (alcohols, esters, hydrocarbons), as
well as in viscous liquids (both Newtonian and non-Newtonian) and in aqueous solutions
of surface active substances (alcohols, electrolytes).
Data on liquid-solid mass transfer are needed to allow complex modelling of reactions
in g-l-s systems. The published information on particle—liquid mass transfer coefficient
(kSL) is, however, far from complete and additional experiments within a wider region of
particle sizes and solids loading are to be performed to verify general validity of
New methodologies for multiphase bioreactors 1 21