The enzymatic hydrolysis of lipids in a hydrophilic membrane bioreactor

The production of fatty acids from lipids (fats and oils) currently takes place in a physical chemical process at a high temperature and pressure. Fatty acids are applied in numerous products such as soaps, detergents and chemicals for pharmaceutical, household and industrial applications. For certain applications the conventional fatty acid production process is not suitable because of side reactions or because of the impurity of the end product. For that kind of applications an enzymatic process, taking place under atmospheric pressure and moderate temperatures, can be appropriate. Another advantage of the enzymatic process is the relatively low energy input.A potential drawback of the enzymatic hydrolysis is however the high enzyme cost as compared to the added value of the end product. This invokes the necessity to re-use the enzyme (lipase). Re-use can be carried out readily when the enzyme is immobilized. Usually enzymes are immobilized to an inert carrier. This method however is not appropriate for lipases since the hydrolysis reaction takes place at the interface of a lipid and a water phase; the spontaneous adsorption of the free enzyme to the Interface will not take place when it is bonded to a carrier. A membrane however provides the opportunity to Immobilize the enzyme just at the interface of lipid and water. For the hydrophilic membrane used in this thesis, the lipase was immobilized on the lipid-side of hollow fiber membranes (chapter 2). The water phase can diffuse in the membrane, while the membrane is impermeable for the lipid phase. This results in an interface at the lipid-side of the membrane where the enzyme is present. The aim of this study was to further develop the knowledge on the membrane reactor and to optimize the system. This was done by investigating fundamental aspects such as reactor parameters, hydrolysis kinetics and mass transfer, as well as practical aspects such as operation mode and system design.The occurrence of possible mass transfer limitations was studied in chapter 3. The diffusion of glycerol/water in the membrane wall as well as the diffusion of lipid components in the core of the hollow fiber membrane was studied using mass transfer models. The rate of both diffusion phenomena appeared to be an order of magnitude higher than the conversion rates reached. Therefore, diffusion was concluded not to limit the conversion rate in the membrane reactor.In chapter 2 the concept of the hydrophilic membrane bioreactor was described and the results of batch hydrolysis experiments were presented. From these results a power law model for the activity as a function of the fatty acid concentration was derived. This model was further worked out in chapter 4; the value of the power was determined more accurately and the influence of the glycerol concentration was incorporated in the model. The kinetics of the free and immobilized enzyme showed no substantial differences. In chapter 4 also a more fundamental model was proposed based on the three consecutive hydrolysis steps. This model could predict the concentrations of fatty acids, mono-, di- and triglycerides in batch experiments at different glycerol concentrations.The stability was studied in chapter 2 in continuous-flow experiments. The half life time of lipase in the membrane reactor was determined as 43 days. In chapter 5 it was shown that the stability of the system can be increased by more than a factor 10 by the addition of CaCl 2 to the water phase. Under comparable conditions, the stability of the membrane-immobilized lipase was shown to be a factor 4 more stable than the free enzyme. For both systems, Arrhenius kinetics were valid for the enzyme inactivation. Also the hydrolysis reaction compiles with Arrhenius kinetics, as proven for the free enzyme. Models for the prediction of the effect of temperature on activity and productivity showed that on a long term the lowest possible temperature Is favoured. Further, in chapter 5 the influences of pH and other additions to the water phase were investigated.The immobilization of lipase on the membrane was carried out by ultrafiltration of a lipase solution. In a period of 100 h after Immobilization, wash-out of enzyme took place (chapter 2 and 6). Thereafter, no free enzyme could be detected any more in the reactor. Ultrafiltration results in a gel layer; There are strong indications that finally a monolayer of lipase remains on the membrane. This would imply that only the lipase-membrane Interactions and not the inter-lipase Interactions are strong enough to withstand the shear force.The interfacial behaviour of the free enzyme was studied in a stirred cell with a controlled interface. Under saturation conditions, the free enzyme adsorbed to the lipid/water interface in a multi-molecular layer (corresponding to ca. 11 lipase layers). Only apart of this layer appeared to be active (corresponding to ca. 3 lipase layers). The activity related to the amount of protein below saturation corresponded with the activity related to the amount of protein immobilized in the membrane reactor. Therefore, it can be concluded that immobilization to the membrane does not result in a loss of activity.An alternative hydrolysis reactor is presented in chapter 7. It consists of an emulsion and two circuits, in which ultrafiltration is carried out with a hydrophilic and a hydrophobic membrane. Thus the water/glycerol product flow is removed by the hydrophlic membrane and the fatty acid/glyceride product flow is removed by the hydrophobic membrane. In contrary to the hydrophilic membrane reactor, the reactive area in this hybrid reactor is not restricted by the membrane area. This implies that in the hybrid reactor the ratio of conversionrate and membrane area can be higher than in the hydrophilic membrane reactor. Indeed, the initial activity of the hybrid reactor, as related to the membrane area, was proved to be higher than the activity of the hydrophilic membrane reactor. A further development of the hybrid reactor system is needed in order to prevent membrane fouling and enzyme inactivation and to optimize the fluxes of the membranes.It can be concluded that the hybrid has high potentials, but that further research is needed. The hydrophilic membrane reactor can be concluded to be a stable system yielding pure products glycerol/water and fatty acid/lipid. By adjusting the net flow rates of the phases, the final glycerol and fatty acid concentrations in the product flows can be regulated. The effects of temperature and product concentrations on the production rate in time can be predicted using the models presented. Thus the results can serve to evaluate the performance of this type of reactor for industrial lipid hydrolysis. The selection of an optimum hydrolysis reactor system for a certain application will require a similar evaluation for all alternative systems.