Water-immiscible solvents for the biological treatment of waste gases

In conventional biological systems for the treatment of waste gases, contaminants are transferred directly to the aqueous phase and then converted by the micro-organisms. When poorly water-soluble pollutants are to be removed, biological degradation is often limited by the slow transport from the gas to the aqueous phase. This transport limitation can be circumvented by contacting the gas directly with an intermediate water-immiscible organic solvent with a high affinity for these contaminants.The study described in this thesis evaluates at which conditions the use of an intermediate solvent is advantageous compared to systems featuring direct gas/water transfer. Aiming at this, different bioreactor configurations and different pollutants were considered and compared.In a review, the current state of the art concerning new biotechniques for a better removal of hydrophobic pollutants from waste gases is given. The transfer rate of these compounds is often improved either by reducing the water layer between the gas and the micro-organisms, or by augmenting the gas/water exchange area. An alternative strategy mentioned in literature involves the use of a water-immiscible solvent. However, limited information is given on the conditions at which the use of a solvent is beneficial.This research started with a preliminary theoretical study in which three solvent-containing systems were compared to bioreactor configurations with direct gas/water transfer. Because this study aimed at characterizing the systems in terms of mass transfer, the biological degradation was considered not rate- limiting. The systems with solvent consisted of a two-compartment system with a packed absorber for gas/solvent transfer and a bioreactor for solvent/water transfer and subsequent biological conversion. Three types of bioreactors were considered namely the liquid-impelled loop reactor (LLR), the packed-bed reactor and the stirred-tank reactor. From this study it was concluded that the use of a solvent as intermediate phase is advantageous in a bioreactor configuration featuring large solvent/water exchange areas and mass transfer coefficients such as in a stirred-tank reactor. This was therefore the type of bioreactor chosen to carry out the experiments.Initially, the effect of the solvent volume fraction and of the partition coefficients gas/water ( m gw ) and gas/solvent ( m gs ) on the mass transfer rate, was studied in a one-compartment system, namely the stirred-tank reactor. The gas containing the contaminant was sparged directly through the solventin-water dispersion. The effect of the solvent on the gas-to-water transfer rate was tested during steady-state experiments in the presence of biological conversion. This was done at increasing solvent volume fractions by measuring the change of the outlet gaseous concentration or of the linear cell growth rate. The model pollutants tested were toluene ( m gw22°C = 0.21) and oxygen ( m gw22°C = 32.7) and FC40 was the solvent used. An enhancement of the mass transfer rate of a factor up to 1.1 was found for toluene at 10 % (v/v) of FC40, while the oxygen transfer rate increased by a factor 2 at the same solvent volume fraction.To be able to predict theoretically at which solvent hold-ups and for which gaseous compounds an enhancement of the gas/water mass transfer rate is expected, a steady-state mathematical model was set-up. This was done for different system configurations and therefore different mechanisms for gas/water transfer through the solvent were considered. An overall gas/water mass transfer coefficient, characteristic of each mechanism, was derived. Since the driving force is the same for all the system configurations, namely the gas/water driving force, the performance of the different systems can be assessed on the basis of this overall gas/water mass transfer coefficient. For a particular configuration, this coefficient is a function of gas/water, gas/solvent and solvent/water exchange areas, mass transfer coefficients and partition coefficients. The model predictions were validated by carrying out steady-state experiments in a stirred-tank reactor containing a solvent-in-water dispersion. The overall gas/water mass transfer. coefficient was determined in the absence of biological conversion to avoid the influence of the cells or their products. Removal of the compound was achieved by passing a continuous flow of compound-free gas or water through the reactor. The solvent was kept in the vessel by means of a small settler. This study was carried out with toluene and oxygen. As found out in the aforementioned experiments with cells, the mass transfer enhancement found for oxygen was larger than the one for toluene at increasing solvent volume fractions. While for oxygen a 2.2-fold increase of the overall gas/water mass transfer coefficient was found at 15 % (v/v) FC40, an enhancement of circa 1.2 was observed for toluene at the same solvent amount. The enhancement factors predicted by the model agreed well with the experimentally determined values.The effect of the cells on the enhancement of the overall gas/water mass transfer coefficient at increasing solvent volume fractions was studied. This was done with ethene and the ethene-degrading bacterium Mycobacterium parafortuitum. At all solvent volume fractions tested, enhancement factors with cells were higher than enhancements without cells. In the presence of cells, a 1.8- fold increase of the k l a value was found at 26 % (v/v) FC40, whereas this value increased of a factor 1.2 in the absence of cells at approximately 19 % (v/v) FC40. The emulsifier effect of the microorganisms or their excretion products was suggested as a possible cause for this difference.A two-compartment system consisting of a packed absorber for gas/solvent transfer and a stirred-tank reactor for solvent/water transfer and microbial degradation was tested with ethene as model pollutant. The solvent was recycled between absorber and bioreactor. A 9 % and a 15 % elimination efficiency was obtained at solvent flows of 6 x 10 -8m 3/s and 11.3 x 10 -8m 3/s, respectively. Although the removal efficiencies obtained were low due to an inefficient use of the column, the feasibility of the system to remove ethene has been demonstrated. The systems's performance was described by a steady-state mathematical model. Simulated ethene removal efficiencies agreed well with the experimental results. It was concluded that with the system dimensions and operational conditions used during the experiments, the limiting step was the mass transfer in the absorber. Furthermore it has been shown theoretically that higher removal efficiencies can be obtained by using a solvent as intermediate phase than by recycling the aqueous suspension between the absorber and the bioreactor, if the contaminant has a higher solubility in the solvent than in water.