Anaerobic biodesulfurization of thiophenes

Distillates from crude oil such as diesel and fuel oil may contain significant amounts of dibenzothiophenes and their alkylated derivatives, containing organically bound sulfur. Combustion of those fossil fuels leads to the release of polluting sulfur dioxide into the atmosphere, where it causes 'acid rain'. Due to stricter environmental legislation and depletion of crude oil reserves with low organic sulfur contents, effective desulfurization processes are becoming increasingly important. For instance: beginning in 2005 the maximal allowable sulfur content in gas oil in the European Community will be 0.005 wt.%. Currently, the refining industry applies the energy intensive physico-chemical hydrodesulfurization (HDS) process in order to reduce the sulfur content. Due to the high costs and inherent chemical limitations associated with HDS, biodesulfurization of hydrocarbon streams might represent an attractive complementary method to obtain sufficiently low sulfur levels. Bacteria require relatively mild process conditions (pressure and temperature) and bacterial enzymes are very selective in converting target molecules.The objective of this thesis was to develop an anaerobic biodesulfurization process. The thesis is build up around the reductive desulfurization reaction presented below.Dibenzothiophene (DBT) is converted under anaerobic conditions to biphenyl and sulfide with the concomitant conversion of reduction equivalents. The caloric value of the fuel molecule is retained and the sulfur is removed specifically.Chapter 1 presents a general introduction on physico-chemical and microbiological methods to desulfurize organic sulfur compounds.In Chapter 2 the DBT mass transfer rate within hydrocarbon droplets is compared to aerobic DBT conversion rates. The apolar DBT must diffuse to the hydrocarbon/water interface where bacteria prevail. The calculated values for the DBT mass transfer rate were compared to those found for aerobic DBT conversion rates, as reported in the literature. Temperature dependent data (ranging from 20 up to 60°C) of viscosity, density, and interfacial tension of various hydrocarbon distillates were incorporated in the model. The model simulated the DBT diffusion in hydrocarbon droplets as obtained in a stirred tank reactor. Based on these calculations, we estimated that the mass transfer rate of DBT within the hydrocarbon droplet to the hydrocarbon/water interface is at least a factor 10 to 10 4higher than the specific DBT conversion rates. However, the presence of a high specific surface area is essential to enhance the surface contact between bacteria and the hydrocarbon phase.The availability of a suitable biomass is crucial to develop this new bioprocess. In Chapter 3 a screening method is described to enrich biomass from mixed bacterial populations obtained from oil-polluted environments. The enriched cultures were able to grow in the presence of thiophenes as the sole electron acceptor. A proof of principle was obtained; the formation of sulfide and biphenyl from dibenzothiophene was shown conclusively. Also thiophene and benzothiophene depletion with concomitant sulfide formation was observed. However, apart from sulfide no thiophene nor benzothiophene desulfurization products could be demonstrated. The main problem during consecutive enrichments was the loss of biological activity after transferring the desulfurizing biomass. A mixed population was present and the active desulfurizing biomass was easily overgrown by acetogenic bacteria. Therefore, it was attempted to isolate the desulfurizing bacteria. The isolation procedure resulted in the availability of highly enriched cultures able to desulfurize thiophenes when cultivated using a selective medium with H 2 as electron donor and limiting amounts of bicarbonate and acetate (1 mM each).Based on process considerations H 2 gas is the most suitable electron donor for the reductive desulfurization process. In Chapters 4 and 5 attention is paid to the mass transfer rate of H 2 in a gas/water/hydrocarbon three-phase system using n -dodecane as model solvent. Because vigorous foam formation occurs when H 2 gas is directly added to a n -dodecane in water dispersion, it was proposed to saturate the n -dodecane with H 2 gas prior to disperse it into the water phase. Experiments to determine the H 2 mass transfer coefficients involved using physical methods are described in Chapter 4. The H 2 mass transfer coefficients between the gas and the n -dodecane phase ( k d ) and between the gas and the water phase ( k w ) were determined using a dynamic method by following the pressure decline in time, whilst the overall H 2 mass transfer coefficient between n -dodecane and water ( k dw ) was determined using a steady state method. The value for k dw was assessed using tritium-hydride (T-H instead of H-H) as the tracer. The effects of the temperature (30, 40 and 50 oC) and salt concentrations (0-250 mM) were studied. The value for k w [(9.7 ± 0.2) x 10 -5ms -1at 30ºC] was found to be a factor 3.3 higher than for k d [(2.89 ± 0.12) x 10 -5ms -1at 30ºC] because of the lower viscosity of water. No effect was found for the presence of salts (up to 250 mM NaCl) on the k w -value. The k dw -value determined in the steady state experiments at 30ºC was (5 ± 0.6) x 10 -6ms -1which is 19.4 times smaller than the above-mentioned k w -value. The considerable smaller value for k dw must be attributed to the additional mass transfer resistance introduced by the second liquid phase. Calculations of the maximal attainable H 2 flux revealed values of 0.016 x 10 -3mol/m 2s and 3.9 x 10 -3mol/m 2s for a n -dodecane/water and gas/water system, respectively. Therefore, the specific surface area between n -dodecane and water is the determining parameter for sufficient H 2 mass transfer. In Chapter 5, the H 2 mass transfer is described further using a bioreactor equipped with a nozzle to create very fine n -dodecane droplets. The specific surface area is dependent on the maximum attainable hold-up of n -dodecane and the diameter of the droplets. These parameters were studied in a model system consisting of n -dodecane and water supplemented with NaCl. The use of the nozzle resulted in droplets with a Sauter mean diameter of only 10.3±0.9mm. The droplet size was found to be independent of the applied pressure drop over the nozzle. The hold-up of n -dodecane in the aqueous medium is clearly dependent on the sodium ion concentration. The hold-up decreases rapidly (from 0.14 to 0.04) with increasing sodium ion concentrations due to coagulation; from 94 mM onwards the hold-up becomes 0.04. The application of n -dodecane droplets as carrier phase for H 2 mass transfer was demonstrated in batch tests for biological sulfate reduction. During operation of the bioreactor, biomass attached to the rising n -dodecane droplets and eventually flotated from the system.In addition biological steady state experiments were performed with hydrogenotrophic sulfate reducing bacteria to determine the H 2 mass transfer coefficient for a n -dodecane/water system ( k dw ). A value of (4.0±0.24) x 10 -6ms -1was found, which is close to the values found in the experiments using tritium hydride. Final calculations showed that the volumetric H 2 mass transfer rate (mol/m 3s) from n -dodecane to water can be comparable to values found for gas lift reactors, thus the high specific surface area that can be created by applying a nozzle can overcome the lower value of the H 2 flux (mol/m 2s) to a large extent.Chapter 6 addresses the role of sulfide on anaerobic biodesulfurization. The presence of increased sulfide concentrations is undesirable because it is expected that sulfide will inhibit the DBT conversion. Therefore, insight in the partitioning of gaseous hydrogen sulfide (H 2 S) over a three-phase gas/water/hydrocarbon system is required. The partitioning of H 2 S over a gas/water/ n -dodecane system is described. Experimental results matched well with the model predictions. The effect of the presence of an extra hydrocarbon phase ( n -dodecane) notably decreased the total sulfide in the water phase and the H 2 S fraction in the gas phase. The hydrocarbon phase serves as a sink for H 2 S molecules and by scrubbing the H 2 S in a separate process step ( e.g. during H 2 saturation) the sulfide concentration can be lowered to favor the anaerobic biodesulfurization.