Physical chemistry and process engineering of an emulsion - membrane bioreactor

Fatty acids (and glycerol) are produced by hydrolysis of fats and oils in counter-current fat- splitting columns which operate at a temperature of 200-240 °C and a pressure of 50-60 bar. Undesired side-products are formed during the process. These have to be removed in order to obtain an acceptable product. The side-reactions do not take place if the fatty acids are produced enzymatically at 30 °C.

The enzyme lipase catalyses the hydrolysis reaction at the oil/water interface. Therefore, a large oil/water interfacial area has to be available to the enzyme for a high volumetric reactor activity. A stirred vessel is used in which an emulsion is formed by thorough mixing of oil, water and enzyme. The products (fatty acid and glycerol) are separated from the emulsion in the stirred vessel by means of two membrane separation steps. A (modified) hydrophobic membrane is used to selectively separate the fatty acids from the protein-rich emulsion. With a hydrophilic membrane the water phase, which also contains the glycerol, is removed from the vessel. This reactor concept, the emulsion/membrane bioreactor, is the subject of this thesis.

The hydrophobic membrane can become permeable for water during the experiment. This effect is caused by enzyme adsorption at the membrane. The enzyme forms a hydrophilic layer on the membrane, in which case it can be preferentially be wetted by the water phase and eventually the water phase will permeate through the membrane (chapter 3).

For proper reactor operation, the hydrophobic membrane has to remain selectively wetted and sufficiently permeable for the fatty acids. Therefore, lipase adsorption has to be prevented. A membrane pre-treatment method with block copolymer F108 as shown in chapter 2 is an effective method to prevent lipase adsorption (see chapter 5), and, therewith, water permeation (see chapter 3). Block copolymer F108 consists of two hydrophilic poly(ethylene oxide) blocks at both ends of the molecule and one hydrophobic poly(propylene oxide) block in the middle. If F108 is contacted with a hydrophobic surface then the middle block will adsorb at the surface and both hydrophilic groups will extend from the surface thus forming a so called "brush" configuration. The "brush" hinders enzyme molecules that approach the surface in such a way that they can not adsorb. The effectivity of hindrance at a hydrophobic surface is a function of the length of the poly(ethylene oxide) groups and the number of pre-adsorbed block copolymer molecules. It is concluded that steric repulsion is the mechanism behind prevention of protein adsorption (chapter 5).

If the block copolymer is adsorbed at a hydrophilic surface then the molecule will adsorb relatively flat onto the surface (pancake configuration). The block copolymer can not prevent enzyme adsorption in this configuration (chapter 5).

Pre-adsorbed block copolymer F108 prevents protein adsorption. However, the membrane hydrophilicity should not be changed (too much) by the presence of the block copolymers. Otherwise the membrane might become water permeable because of the presence of the block copolymers. In chapter 4 it is shown theoretically and experimentally that hydrophobic surfaces with pre-adsorbed block copolymers remain oil-wetted. The surface properties are hardly influenced by the presence of F108 and the surface can be wetted by a large variety of "oils".

The F108-modified membrane can be used for the continuous hydrolysis of oil in the emulsion/membrane bioreactor. From literature a model for the production of fatty acid and glycerol in a membrane reactor is adapted and extended with a model for enzyme inactivation in an emulsion. The model predictions are in agreement with the experimental data (chapter 6).

With the model an economic evaluation of the emulsion/membrane bioreactor is made in chapter 7. It is found that the production per gram enzyme, the concentrations of fatty acid and glycerol in the product streams and the production per m ²membrane area is within the limits for an economically feasible processes. If the reaction is carried out in a co- current series of reactors then also the volume of the stirred vessels is economically feasible.

A promising spin-off of the modification research is the use of the block copolymer for filtration of aqueous protein solutions (see chapter 7). It is shown that the flux of this membrane remains high during long-term operation.