Multiphase Bioreactor Design

aggregates or, more likely, as monomers. This leads to the presence of surfactant in the
products’ stream and makes the separation and purification of products extremely
difficult, especially when they are to be used in foods, pharmaceuticals, cosmetics or
other health areas.
An attempt to solve or minimise the surfactant contamination problem was made by
Hakoda et al. (1996) and Nakamura and Hakoda (1992). The authors used the electro-
ultrafiltration (EUF) method to decrease the gel formation in the membrane surface,
improving the filtration flux, and achieving the separation of the AOT reverse micelles.
The rejection of AOT was naturally influenced by the direction of applied voltage. The
best solution occurred when the cathode (with negative charge) was installed in the
permeate side—this can be explained by the repulsion of the AOT negative charge. The
permeation flux increased with the electrical field strength. The rejection of AOT
increased, but did not exceed 15%, while the rejection of water reached 30% (Nakamura
and Hakoda, 1992). A drawback of the EUF was the deactivation of lipase caused by the
application of voltage.
Other problems of the membrane reactors are the leakage of enzyme, adsorption and
concentration polarisation leading to membrane fouling and enzyme deactivation
promoted by shear forces.
One interesting work performed in a reversed micellar reactor was reported by Chiang
and Tsai (1992). The reaction under study was the hydrolysis of olive oil by Candida
rugosa lipase microencapsulated in reverse micelles of AOT/isooctane. The main portion
of the reactor was a recycle dialysis stirred cell used to integrate the reaction and product
recovery. The resistance of the dialysis membrane to reversed micelles was controlled by
the water content and the rejection value after 24 hours was 95.9%. UV absorption was
used to detect surfactant and the rejection coefficient was 97.3% after 24 hours (Chiang
and Tsai, 1992). Despite the good results in micelle retention, further evaluation would be
necessary to analyse the continuous operation of this type of reactor over time.

Characterisation of Micellar Membrane Bioreactors

Membrane properties and characterisation

The choice of a membrane for a specific enzyme process is determined by its nominal
molecular weight cut-off (NMWCO), which depends on the average pore size and pore
size distribution. Other properties of the membrane such as chemical resistance to organic
solvents, temperature, pH and pressure should also be considered with respect to the
operating conditions and the cleaning/sterilisation process (Hildebrandt, 1991).
Ultrafiltration ceramic membranes, despite their cost, are highly resistant to chemicals
and extreme operating conditions.
The membrane characterisation comprises the membrane porosity/morphology and
also its surface properties. Recently, atomic force microscopy (AFM) has developed as a
powerful tool for the characterisation of membranes (Bowen et al., 1996).
MBR must be characterised in terms of filtration properties. The filtration flux rate, J,
through a membrane is defined as:

(15)

Reversed micellar bioreaction systems 211