process considerations in the biodesulfurization of crude oil

ORNLJCP-100055

t'

"The submitted manuscript has been authored by a contractor of the US. Government under contract DE-AC05-960R22464. Accordingly, the US. Government retains a

nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or allow others to do so, for US. Government purposes. "

-

i

Process Considerations in the Biodesulfurization of

Crude Oil

Abhijeet P. B o d e * and Eric N. Kaufman

Bioprocessing Research and Development Center Chemical Technology Division

Oak Ridge National Laboratory? Oak Ridge, Tennessee, USA 3783 1-6226

For Presentation at the

5" Annual International Petroleum Environmental Conference Albuquerque, NM

October 20-23, 1998

* Corresponding author, e-mail: h r o t eaaornt. gov

DE-ACO5-96OR.22464 with Lockheed Martin Energy Research Corp. Research supported by the Office of Oil and Gas Processing, U.S. Department of Energy under contract

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, make any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply i ts endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

Process Considerations in Biodesulfirization, Borole and Kaufian. 1

Process Considerations in the Biodesulfurization of

Crude Oil

Abhijeet P. Borole* and Eric N. Kaufman

Bioprocessing Research and Development Center Chemical Technology Division

Oak Ridge National Laboratory? Oak Ridge, Tennessee, USA 3 783 1-6226

"The submitted manuscript has been authored by a contractor of the US. Government under contract DE-AC05-960R22464. According&, the US. Government retains a

nonexclusive, royalv-fiee license to publish or reproduce the published form of this contribution, or allow others to do so, for US. Governmentpurposes. "

ABSTRACT ?-

Biodesulfurization offers an attractive alternative to conventional hydrodesulfurization due to the mild operating conditions and reaction specificity afforded by the biocatalyst. The enzymatic pathway existing in Rhodococcus has been demonstrated to oxidatively desulfurize the organic sulfur occurring in dibenzothiophene while leaving the hydrocarbon intact. In order for biodesulfurization to realize commercial success, a variety of process considerations must be addressed including reaction rate, emulsion formation and breakage, biocatalyst recovery, and both gas and liquid mass transport. This study compares batch stirred to electro-spray bioreactors in the biodesulfurization of both model organics and actual crudes in terms of their operating costs, ability to make and break emulsions, ability to effect efficient reaction rates and enhance mass transport. Further, sulfur speciation in crude oil is assessed and compared to the sulfur specificity of currently available biocatalysts. -

KEY WORDS: crude oil desulfurization, Rhodococcus, electrostatic spraying, dibenzothiophene, biodesulhization

* Corresponding author, e-mail: borolea@ornl .sov Research supported by the Office of Oil and Gas Processing, U.S. Department of Energy under contract DE-AC05-960R22464 with Lockheed Martin Energy Research Corp.

Process Considerations in Biodesu1ji.o-ization, Borole and KauJinan. 2

INTRODUCTION

Biological processing of fossil fuel feedstocks offers an attractive alternative to conventional thermochemical treatment due to the mild operating conditions and greater reaction specificity afforded by the nature of biocatalysis. Efforts in microbial screening and development have identified microorganisms capable of petroleum desulfurization (see for example (1- 12)), denitrification (6), demetalization (6), cracking (6, 13) and dewaxing. Further investigation and manipulation of enzymatic pathways responsible for these reactions (4, 14-17) has led to processes which are approaching commercial application, particularly in the area of bioiogicai desulfurization (7, 18). Biological desulfurization of petroleum may occur either oxidatively (see for example (3, 5, 8, 9, 12, 17, 19-21), or reductively (2,22-24). In the oxidative approach, organic sulfur is converted to sulfate and may be removed in process water. This route is attractive due to the fact that it would not require further processing of the sulfur and may be amenable for use at the well head where process water may then be reinjected. In the reductive desulfurization scheme, organic sulfur is converted into hydrogen sulfide which may then be catalytically converted into elemental sulfur, an approach of utility at the refinery. A sampling of desulfurization rates achieved with oxidative and reductive microorganisms have been summarized in (25). Regardless of the mode of biodesulfurization, key factors affecting the economic viability of such processes are biocatalyst activity and cost, differential in product selling price, sale or disposal of co-products or wastes from the treatment process, and the capital and operating costs of unit operations in the treatment scheme.

The selection of the petroleum feedstock in biodesulfurization will play a large role in the overall economic viability of the process. Biodesulfurization may be utilized as a pretreatment to crude oil before entering pipelines, may be applied as an alternative to hydrotreating the crude at the refinery, or may be applied in the polishing of refinery products such as diesel or gasoline. The particular application will determine the extent of desulfurization necessary and hence the treatment cost per barrel. At the wellhead, a biodesulfurization unit may be used to treat marginally sour crudes (0.6 - 0.7% S) converting them to sweet crudes (<OS% S) and claiming the price differential in sweet versus sour crude in segregated pipeline systems (currently, the premium for sweet crude is -$I/Bbl). For this application, the extent of desired desulfurization is quite low and this may serve as an attractive initial niche for biodesulfurization. When utilized for refinery applications, the biodesulfurization process must compete with conventional hydrotreating. Here the economic viability of biorefining will be based upon its competitiveness relative to the catalyst replacement, hydrogen, and octane penalty costs associated with hydrotreating. -

Dibenzothiophenes and substituted dibenzothiophenes are some of the most recalcitrant organic sulfur compounds relative to hydrodesulfurization of crude oil and refined transportation fuels (7). DBT has been used as a sulfur model compound for over a decade of research. DBT desulfurization by the “4s” pathway as performed by Rhodococcus IGTS8 (19, 26) conserves hydrocarbon value as opposed to desulfurization via mineralization (see Figure 1). In this mechanism, sulfur oxidized and removed from the petroleum as sulfate which partitions into the biocatalyst containing aqueous phase. IGTSS and related biocatalyst have been shown capable of desulfurization of DBT, substituted DBT’s, and other sulfur species in model hydrocarbon systems (3, 5, 21,27, 28). There have been published reports of diesel (28) and whole crude oil desulfurization (4, 6, 24, 29, 30). However, most reports of crude biodesulfurization investigate

Process Considerations in Biodesulfurization, Boi-ole and Kaufinan. 3

only the total sulfur content of the crude before and after biodesulfurization, and do not reveal sulfur speciation data which may guide further biocatalyst development.

While significant activity is progressing in the engineering and chemical modification of enzymes so that they may hnction in purely organic solutions (31-33), inherent to all of the current bioprocessing of fossil feedstocks schemes is the need to contact a biocatalyst containing aqueous phase with an immiscible or partially miscible organic substrate. Factors such as liquid / liquid and gas / liquid mass transport, amenability for continuous operation and high throughput, capital and operating costs, as well as ability for biocatalyst recovery and emulsion breaking are significant issues in the selection of a reactor for aqueous / organic contacting. Traditionally, impeller-based stirred reactors are utilized for such mixing due to their ease of operation and wide acceptance in the chemical and biological processing industries. Such mechanically stirred reactors contact the aqueous and organic phases by imparting energy to the entire bulk solution, i.e. the impeller must move the contents of the reactor. Energy input in the stirred reactor is a function of the phase ratio, oil viscosity, density, reactor size, impeller speed, etc. (34). An analysis of the energy requirements for stirred reactors is included in this study.

Alternative processing schemes (18) propose the use of “motionless mixers”, in which the two phases are pumped over a reversing helical coil which creates turbulent eddies. While this method reduces the number of moving parts in the reactor, it does not reduce power requirements since costs are transferred from the impeller to the pumps required to move the liquids past the coil. Liquid velocities greater than 4 m/s are required to form emulsions (-10 pm) and one can not form emulsions in coil tubes greater than 3 mm in diameter (35). For tubes greater than 10 cm in diameter one can not form droplets smaller than 1 mm. Like the stirred reactor, no capacity exists for emulsion breaking within the motionless mixer.

Recent advances in the area of contactors for solvent extraction have lead to the development of electrically driven emulsion phase contactors (EPC-) for eficient contact of immiscible phases (36-41). In this concept, the differing electrical conductivity between the aqueous and organic phases causes electrical forces to be focused at the liquid / liquid interface, creating tremendous shear force (see for example (42)). This shear causes the conductive phase to be dispersed (5 pm droplet size - (43)) into the non-conductive phase, but does so with decreased energy requirements relative to mechanical agitators due to the fact that energy is imparted only at the liquid / liquid interface and not the entire bulk solution. Electrostatic crude oil desalters have been operated for several years (44). More recently, devices based upon the EPCW have been used commercially to water wash methyl tert-butyl ether feedstocks (with greater than 1 0-fold reduction in stage height (45)) and for organic extraction from aqueous analytical samples (46). Energy consumption on a volumetric basis has been measured to be 2.4 W/L for a 30% tributyl phosphate / 70% dodecane / distilled water system, four orders of magnitude less than- mechanical systems creating as fine an emulsion (40,43).

The configuration of the EPCTM developed at the Oak Ridge National Laboratory is shown in Figure 2 where the contactor serves to disperse a liquid with a greater density than the continuous phase. The reactor employs two different types of electrode regions in order to increase liquid throughput. The frrst, termed the ‘‘nozzle region”, provides a high capacity droplet dispersion by providing an electric field with a significant vertical component. This vertical field creates the dispersion at the nozzle entrance and accelerates it into the continuous phase. A second region termed the “operating channel” employs parallel plates carrying a modulated dc offset with high voltage spikes. This signal creates an oscillating horizontal

Process Considerations in Biodesulfurization, Borole and KauJinan. 4

electrical field which controls the residence time of the dispersed phase and serves to continuously coalesce and redisperse the droplets as they progress in a serpentine manner through the reactor. At the base of the reactor, an electrical field exists between the electrified central plate and the grounded aqueous phase, which accelerates the aqueous droplets to the organic / aqueous interface. In this manner, droplet coalescence and hence separation on the interface is enhanced. The EPCm creates droplets of water containing biocatalyst -5 pm in diameter within an organic phase.

With the success the EPCm has exhibited in the area of solvent extraction, it was proposed that it could be an efficient reactor system for aqueous / oil contacting in biorefining (47). Here, we compare the performance of the EPCm to that of a batch stirred reactor (BSR), investigate the required level of biocatalyst activity before the surface area afforded by the EPCTM becomes a factor in reactor performance, and characterize the emulsion formed by both reactors in the presence of bacteria. We have investigated the emulsion quality formed in the EPC, evaluated the power requirements and anaIyzed the mass transfer issues in comparison to stirred reactors. Results on biodesulfurization of actual crude oil by wild type Rhodococcus IGTS8 are also included. Finally, we assess the sulfur specificity of available biocatalysts with respect to sulfur compounds present in crude oils.

MATERIALS AND METHODS

The experimental procedures used for studying biodesulfurization in model systems have been discussed in detail in previous publications (25,48,49). A brief description is provided here to aid in the understanding of the results from model system experiments. A detailed description of oil experiments is provided here.

Model system experiments

BiocataIyst and soIvent systems

The model solvent system included dibenzothiophene (DBT) dissolved in hexadecane and was used to investigate reactor design and performance in an easily tractable chemical system. Rhodococcus sp. wild strain IGTS8 (ATCC 53968) was provided by Energy BioSystems Corp. and served as the biocatalyst. The sequence of DBT oxidation by IGTSS is shown in Figure 1,- (Kilbane, 1989) and detailed enzymatic steps in the pathway are discussed in (17). A frozen cell paste was used which had a cell dry weight of 0.28 g/g of original frozen material. The aqueous phase in all experiments consisted of 0.156 M, pH 7.5 potassium phosphate buffer.

.

Batch stirred reactor experiments

Experiments conducted in batch stirred reactors (BSR) typically utilized 50 g of frozen Rhodococcus sp. wild type strain IGTS8 (ATCC 53968) cell paste which were brought up to 750 mL with 0.156M (pH 7.5) phosphate buffer (1X cell density) and added to 250 mL of 0.6wt%

Process Considerations in Biodesulfirization, Borole and Kaufian. 5

DBT in n-hexadecane. The experiment was conducted at 30°C, with agitation at 800 RPM, and aeration with either room air or pure oxygen at a rates of either 0.2 or 1 .O SLPM.

Mass transport experiments were conducted in the BSR to evaluate the liquid-liquid (1-1) or gas- liquid (g-I) mass transfer limitations. These experiments were organic continuous with 250 mL of aqueous phase dispersed into 750 mL of hexadecane. The cell density was varied over a 20- fold factor to study the mass transfer limitations in the biodesulfurization process. In order to separate the gas-liquid from liquid-liquid mass transfer issues, experiments were conducted at a fixed cell density and varying the oxygen tension or flow rate.

E P P experiments

The details of the experimental set up and instrumentation used for the EPCTM experiments are given in (25). In EPCW experiments, the organic liquid served as the continuous phase into which an aqueous biocatalyst was dispersed. The temperature of the organic phase was controlled at 30°C by pumping the liquid through a constant temperature bath. Typically, a 100 mL aqueous slurry of biocatalyst was contacted with 2,400 mL of organic phase by recirculating through the reactor at 5.0 mumin. The cell density with respect to the amount of aqueous phase used in these EPCm experiments corresponded to a 4X concentration. A higher cell density was used in the EPCW as compared to the BSR experiments due to the lower aqueous:organic phase ratio in the EPCW and to enable DBT conversion significant enough to be observed by the analytical procedure employed. Aeration of the aqueous phase was achieved by passing air through a diffuser located in an external reservoir containing the aqueous phase.

In experiments to determine possible mass transport limitations in the EPCTM, 66.7 g of frozen cell paste was brought to 100 mL with potassium phosphate buffer (1OX cell density) and used as the aqueous phase. In order to evaluate if oxygen mass transfer was a limiting factor, the aqueous reservoir used for recirculating the aqueous phase was sparged with pure oxygen at a flow rate of 50 mumin. This experiment was also conducted at a 1OX cell density.

Biodesulfurization of Van Texas Crude oil

Biodesulfurization of Van Texas crude oil was studied in batch stirred reactors to evaluate the substrate specificity of the biocatalyst. The experiment was conducted over a treatment period of 6 days. The crude had an API specific gravity of 310, and a sulfur content of 0.96 M.%. The crude oil did not contain volatiles due to production at elevated temperature (-99OC). Experiments were performed in batch stirred reactors utilizing 50 g of frozen Rhodococcus sp. wild type strain IGTS8 (ATCC 53968) cell paste which were brought up to 750 mL with 0.156M- (pH 7.5) phosphate buffer. The cells were suspended in the phosphate buffer prior to addition to the reactor. The reactor vessel used was a 1-L VirTis Omni-Culture fermentor (model 178657, Gardiner, NY), utilizing a 6-bladed Rushton-type impeller with 2 baffles. The reactor was kept at 3OoC, agitated at 800 RPM, and aerated with room air at a rate of 0.2 standard liters per minute (SLPM). A water condenser was used on each reactor to capture volatiles which were expected to be minimal or non-existent considering the fact that the operating temperature was much less than that of the oil reservoir. The experiment was conducted with 250 mL of crude oil, treated with 750 mL of the aqueous phase. Samples (30 mL from the top of the organic phase) taken during the course of biological treatment were collected after ceasing the agitation and aeration for 5 min to allow the aqueous and organic phases to separate. The reactor contents were

.


emptied at the end of the run and centrifuged at 6000 rpm in a Beckman Model TJ-6 centrifuge to obtain a sample of treated crude oil. Closed samples were boiled in a closed container for 30 min to halt biological activity.

Analytical

Model system experiments

In the experiments reported here, DBT and 2-HBP concentrations in the aqueous phase were below our levels of detection. DBT and 2-hydroxybiphenyl (2-HBP) concentrations in n- hexadecane were measured by gas chromatography using a Hewlett Packard 5890 gas chromatograph equipped with a flame ionization detector.

Crude oil

A GC-SCD method was used to determine the sulfur content of the aromatic fraction of the oil. In conventional gas chromatographic analysis of whole oil, only a portion of the sample elutes from the column. To allow facilitated observation of sulfur in the treated oil, whole oil samples were fractionated according to ASTM method D2007. An extended ASTM D2887 procedure was used for chromatographic separation of the aromatic fraction of the crude oil. The extended ASTM D2887 procedure is a gas chromatographic simulated distillation method to obtain comparable data to the ASTM D2892 physical distillation for crude oil. An AC (Analytical Controls Inc., 3448 Progress Drive, Bensalem, PA 19020) software on Hewlett-Packard Chemstation was used to control a HP 5890 GC and produce the simulated distillation reports. Sulfur analysis was performed by modifying the ASTh4 D2887 procedure by adding a Sievers Chemiluminesence sulfur specific detector after the flame ionization detector. The dual detectors produce a chromatogram for the hydrocarbon simulated distillation and a chromatogram for the sulfur distribution in the sample. The GC method employed a 10m x 0.53 mm AC Sim Dist methyl silicone column with a 0.9pm film thickness. Helium was used as a carrier gas. The GC oven program began at 400 C with a ramp of 150 /min to 3800 C followed by a holding time of 4.5 min. The injector port was temperature programmed to begin at 1250 C and ramp up to 3800 C at 400 /min. The ASTM D2887 calls for carbon disulfide as the diluent for the oil samples, but in this procedure, triflouro-trichloroethane was used since sulfur determination was one of the analytical objectives. In the case of fractionated oil analysis, total sulfur was calculated from GC-SCD data by integrating the chromatograms and comparing their areas to those of known standards resulting in an accuracy of 0.01% S. This integration could not be performed on whole oil GC-SCD data since the baseline failed to return to zero and the resulting area was indeterminant.

-

Process Considerations in Biodesulhrization, Borole and Kaufian. 7

RESULTS AND DISCUSSION

Rate of biodesulfurization

The specific rate of DBT desulfurization by Rhodococcus sp. was typically between 1 and 5 mg 2-HBP produced per dry g of biocatalyst per hour. Specific rates of 2-HBP production in the batch stirred reactor and the EPCTM reactor systems were within experimental variance and no appreciable difference in desulfurization rates were seen between the two reactors. Due to the high surface area reported in the EPCm (25), higher rates were expected in the EPCTM, however, similar performance was observed in both reactors. The reaction rate obtained here, was without any supplemental carbon or energy source. Note that the only available carbon and energy source for the biocatalyst other than what may be carried over in the frozen cell paste, was hexadecane and DBT. However, DBT was not used as the carbon source by the biocatalyst, since the end product of DBT conversion was 2-HBP (thus preserving the carbon number and fuel value). Other studies (outlined in (25)) have utilized additional external carbon and energy sources and have reported higher activities with Rhodococcus sp. A commercial scale biodesulfurization process may require a higher cell density to achieve maximum conversion in a minimum time, provided it does not affect yield with respect to biocatalyst usage. In order to study effect of cell density, experiments were conducted in the BSR at different cell loadings. This experiment also lead to determination of the cell density at which point the BSR becomes mass transfer limited. Another experiment in the EPCTM at this cell density was expected to evaluate the benefits of high surface area afforded by the EPCTM.

Mass transport issues

Results of the DBT desulfurization experiments conducted at varying cell densities in BSR’s are given in Figure 3. The rate of desulfurization, when normalized with respect to cell mass, was found to decrease with increasing cell density indicating that mass transfer resistance was the controlling process in desulfurization. A statistical analysis of the data indicated mass transfer limitation between 5X and 1 OX cell density in the BSR. The mass transfer limitation may be due to gas-liquid or liquid-liquid mass transport resistance.

The results of experiments conducted in the BSR at 1OX cell density indicated no gas-liquid mass transfer limitation (Figure 4). Increasing the rate of air supply or increasing the oxygen tension in the reactor through the use of pure oxygen rather than air was not seen to affect HBP production. This suggests that the system may be limited by liquid-liquid mass transfer. Since the EPCTM reportedly provides larger liquid-liquid interfacial area, the BSR was compared with the EPCm for desulfurization activity at equal cell density.

Comparison of the EPCTM and BSR at 1OX cell density is given in Figure 5. As shown, no difference was observed in the desulfurization rates between the two reactors. Thus, either the system is not truly mass transport limited or the EPCTM does not provide a larger surface area for reaction under the present conditions. A detailed characterization of the emulsions formed in the BSR and EPCTM in the presence and absence of biocatalyst was conducted and is reported below.


Emulsion quality in BSR and EPCM

A detailed drop size analysis of the two-phase emulsion formed in the BSR has been reported previously (25). Characterization of the emulsion quality in BSR in the absence of biocatalyst has revealed 100-200 micron droplets under the conditions of experiments conducted here. The droplets formed in the EPC; however, are in the 1-10 micron range. The ability to form fine emulsions in the EPC- without increasing energy utilization (see energy utilization section below) could have tremendous impact upon processing costs assuming that the biocatalyst utilized is active enough to be mass transport limited.

Emulsion quality in the presence of biocatalyst

Due to the opaqueness rendered by presence of biocatalyst, observations could not be made in situ during reactor operation. To determine the emulsion quality formed in the EPCW and BSR, and to determine whether the EPCW offers larger surface area than BSR, samples were collected from the reactors and observed under a microscope using a lOOx oil emersion objective. Microscopic examination of samples showed formation of a very fine emulsion in both reactors with droplet sizes ranging from 1 to 10 pm. Formation of such an emulsion in the BSR may be presumed due to production of biosurfactants by the biocatalyst IGTS8. Average droplet size for EPCTM and BSR samples were 2.54 f 2.40 pm and 3.08 k 1.78 p, respectively (n>300). Figure 6 shows a micrograph of the emulsion obtained in an EPC. Further, a significant amount of the biocatalyst was extracted and existed in the organic phase. Thus, a very fine emulsion is formed in the EPCW as well as the BSR, and it appears that it is for this reason that an augmentation in desulfurization rate is not seen in the EPCTM relative to the BSR. A couple of process issues warrant consideration here. Firstly, due to the formation of a very fine emulsion, downstream separation of the multiphase mixture to obtain clean organic fuel may require substantial efforts in terms of additional separation processes and can affect economics adversely. Secondly, the Rhodococcus biocatalyst was used here, was extracted into the organic phase. Both these characteristics are common to the gram-positive type of bacteria. On the other hand, traits of certain gram-negative bacteria, for e.g., hydrophilic cell wall structure, can result in retention of bacteria in the aqueous phase (vs. extraction into organic phase) and avoid costly downstream separation processes. Recent improvements in the field of genetic engineering can help construct the desulfurization genes into gram-negative bacteria producing useful biodesulfbrization biocatalysts.

9

Energy utilization by E P F M -

Typically, stirred reactors or impeller based reactors are capable of achieving water or oil droplet sizes of 100 -300 pm in diameter under the conditions used in this study when surfactants are not present. Tin order to create such droplet size distribution, the energy required is on the order of 1-6 W L (based upon empirical correlation’s (34)). It is estimated that if impeller based systems were capable of producing 5 pm droplets, it would require -25 kW/L (43) if surfactants are not present. The EPC- creates droplets of water containing biocatalyst -5 pm in diameter within an organic phase, and does so with a power requirement of 3 W/L (25). Thus, if a high activity biocatalyst is available, which is actually limited by mass transport, the EPCTM could result in tremendous savings over the batch stirred reactor. For instance, on a 1 L basis, a BSR using a

Process Considerations in Biodesulfirization, Borole and KauJinan. 9

3:l water to oil ratio and producing oil droplets of 150 mm in diameter creates 1 x 105 cm2 of interfacial surface area. On the same volume basis, an E P P creating 5 pm diameter aqueous droplets and having a 5% aqueous hold-up creates 6 x 105 of interfacial area at 1/15th the aqueous volume to do so. In a mass transport system, the rate of desulfurization would thus be expected to be six times as large using 93% less biocatalyst. An additional important point which needs to be noted here is that the fine emulsion formed in the EPCm is an unstable emulsion i.e., the emulsion breaks easily upon removal of the electric fields giving easy separation of the organic and aqueous streams.

Crude oil biodesulfurization

Oil samples collected from the BSR were analyzed by GC-SCD to obtain the distribution of organosulfur compounds in the crude oil. Analysis was done on the aromatic fraction of the oil and not the whole oil, since the baseline did not return to its initial value in case of the whole oil. The aromatic portion of the oil allowed quantification of sulfur removal with unambiguous chromatogram baseline accounting. This fraction of the Van Texas oil accounted for 22% of the oil’s original volume. As shown in Figure 7, the total sulfur content of this aromatic fraction was reduced from 3.8 to 3.2% in 6 days of treatment with IGTS8. Note that GC analysis does not report information for species with boiling points exceeding -565OC since these compounds do not elute fkom the column. In the chromatograms shown in Figure 7, these high boiling point compounds accounted for 16-1 8% of the hydrocarbon present.

The results indicate removal of comparatively low molecular aromatic sulfur compounds; however, a large portion of the organosulfur fraction does not seem to be affected by the biodesulfurization process. These results are the first that report sulfur speciation data in crude oil that has been treated with a publicly available biocatalyst. While it appears that this biocatalyst is capable of desulfurizing the majority of sulfur species present in FCC gasoline and diesel (DBT and substituted DBT compounds) and that only improvements in the rate of desulhization are needed for the commercialization of these processes, a great deai of research is needed for oil biodesulfurization to be realized. The sulfur specific oxidation of DBT by Rhodococcas resulted fiom over 15 years of research using DBT as the model organic sulfur compound in coal and oil. Detailed sulfur speciation studies and biocatalyst development is needed to achieve desulfurization of the broad spectrum of organic sulfur species present in crude oil and to realize the promises of petroleum biodesulfurization.

4

CONCLUSIONS

A variety of process considerations in the biodesulfurization of petroleum feedstocks were addressed in this study including reaction rate, emulsion formation and breakage, biocatalyst recovery, and both gas and liquid mass transport. Comparison of batch stirred reactor to EPCm revealed formation of high surface area in the EPCm in the absence of surface-active agents. Presence of biocatalysts capable of producing biosurfactants results in fine emulsions in both reactors; however, poses a problem of downstream separation of the multiphase mixture. The use of EPCm as a biodesulfurizaton reactor can result in up to several orders of magnitude energy savings over BSR, if highly active biocatalysts with useful traits such as hydrophilic cell wall characteristics are made

Process Considerations in Biodesulfurization, Borole and Kaujinan. 10

available. Gas-liquid mass transfer was not a limiting factor in biodesulhxization studies with model systems. Further, biodesulfurization experiments with actual crude oil showed that presently available biocatalysts such as Rhodococcus sp. IGTSS are capable of removing essentially DBT and substituted DBT type compounds. Thus, there is a need for further development in biocatalysts capable of desulfurization of higher molecular weight non-DBT type sulfur compounds present in crude oil.

ACKNOWLEDGMENTS This work was supported by the Office of Oil & Gas Processing, U.S. Department of Energy under contract DE-AC05-960R22464 with Lockheed Martin Energy Research Corp. The authors wish to thank Dr. Robert Shong of Texaco for GC-SCD analysis of Van Texas crude oil. The authors acknowledge the material contributions of Energy BioSystems Corp., and the technical assistance of Drs. Punjai Selvaraj and Costas Tsouris at the Oak Ridge National Laboratory.

REFERENCES 1.

2.

3.

4.

5.

Constanti, M., J. Giralt, and A. Bordons, "Desulfurization of dibenzothiophene by bacteria". World Journal of Microbiology and Biotechnology, 1994. 10: p. 5 10- 516. Finnerty, W.R., "Organic sulfur desulfurization in non-aqueous media". Fuel,

Kayser, K.J., et al., "Utilization of organosulfur compounds by axenic and mixed cultures of Rhodococcus rhodochrous IGTSS". Journal of General Microbiology,

Kim, T.S., H.Y. Kim, and B.H. Kim, "Petroleum desulfurization by desulfovibrio desulfuricans m6 using electrochemically supplied reducing equivalent". Biotechnology Letters, 1990. 12( 10): p. 757-760. Lee, M.K., J.D. Senius, and M.J. Grossman, "Sulfur-specific microbial desulfurization of sterically hindered analogs of dibenzothiophene". Applied and Environmental Microbiology, 1995.61( 12): p. 4362-4366. Lin, M.S., et al., "Biochemical processing of heavy oils and residuum". AppZied Biochemistry and Biotechnology, 1996.57158: p. 659-664. -

. 7. Monticello, D. J., "Biocatalytic desulfurization". Hydrocarbon Processing, 1994.:

8. Ohshiro, T., Y. Hine, and Y. Izumi, "Enzymatic desulfurization of dibenzothiophene by a cell free system of Rhodococcus erythropolis D-1". FEMS Microbiology Letters, 1994. 118: p. 341-344.

9. Ohshiro, T., T. Hirata, and Y. Izumi, "Microbial desulfurization of dibenzothiophene in the presence of hydrocarbon". Appl. Microbiol. Biotechnol.,

1993. 72(12): p. 1631-1634.

1993.139: p. 3123-3129.

6.

p. 39-45.

1995. 44: p. 249-252.

Process Considerations in Biodesulfurization, Borole and Kaufian. 11

10.

11 .

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Shennan, J.L., "Microbial attack on sulfur-containing hydrocarbons: Implications for the desulfurization of oils and coals.". J: Chem. Tech. Biotechnol., 1996. 67:

Sorokin, D.Y., "Biological oxidation of sulfur atoms in C1-sulfur and other organosulfur compounds". Microbiology m, 1993.62(6): p. 575-581. Wang, P. and S . Krawiec, "Desulfurization of dibenzothiophene to 2- hydroxybiphenyl by some newly isolated bacterial strains". Arch. Microbiol.,

Premuzic, E.T., et al., Biochemical Alteration of Crude Oils in Microbial Enhanced Oil Recovery, in Biohydrometallurgical Technologies, A.E. Torma, M.L. Apel, and C.L. Briarley, Editors. 1993, The Minerals, Metals, and Materials Society. p. 401-413. Denome, S.A., et al., "Characterization of the desulfwization genes from Rhodococcus sp. strain IGTSS". Journal of Bacteriology, 1994.176(21): p. 6707- 6716. Ohshiro, T., et al., "Involvement of flavin coenzyme in dibenzothiophene degrading enzyme system from Rhodococcus erythropolis D- 1 'I. Biosci. Biotech. Biochem., 1995.59(7): p. 1349-1351. Vazquez-Duhalt, R., D.W.S. Westlake, and P.M. Fedorak, "Cytochrome c as a biocatalyst for the oxidation of thiophenes and organosulfides". Enzyme Microb. Technol., 1993.15(6): p. 494-499. Gray, K.A., et al., "Molecular mechanisms of biocatalytic desulfurization of fossil fuels". Nature Biotechnology, 1996. 14: p. 1705-1 709. Dounias, G.A. and K.D. Stavropoulos, "Economic Feasibility of Biochemically Upgrading Heavy Crudes," . 1995, Brookhaven National Laboratory. Kilbane, J.J. and B.A. Bielaga, "Toward sulfur-free fuels". CHEMTECH, 1990.:

Olson, E.S., D.C. Stanley, and J.R. Gallagher, "Characterization of intermediates in the microbial desulfhization of dibenzothiophene". Energy and Fuels, 1993. 7:

Omori, T., et al., "Desulfurization of akyl and aromatic sulfides and sulfonates by dibenzothiophene desulfurizing Rhodococcus sp. strain SY 1 'I. Biosci. Biotech. Biochem., 1995.59(7): p. 1195-1 198. Kim, H.Y., T.S. Kim, and B.H. Kim, "Degradation of organic sulfur compounds

Biotechnology Letters, 1990.12( 10): p. 761 -764. L i m a , H.M., L.A. Wilkins, and T.C. Scott, "Dibenzothiophene sulfur can serve as the sole electron acceptor during growth by sulfate reducing bacteria". Biotechnology Letters, 1995.17(1): p. 113-1 16. Kim, B.H., et al., "Selectivity of desulfurization activity of Desulfovibrio desulfuricans M6 on different petroleum products". Fuel Processing Technology,

Kaufman, E.N., et al., "Development of an electro-spray bioreactor for crude oil processing". Fuel Processing Technology, 1997.52: p. 127- 144.

p. 109-123.

1994.161: p. 266-271.

p. 747-751,

p. 159-164.

and the reduction of dibenzothiophene to biphenyl and hydrogen sulfide". -

1995.43: p. 87-94.


26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

. 40.

41.

42.

43.

Gallagher, J.R., E.S. Olson, and D.C. Stanley, "Microbial desulfurization of dibenzothiophene: a sulfur specific pathway". FEMS Microbiology Letters, 1993.

Ohshiro, T., T. Hirata, and Y. Izumi, "Desulfurization of dibenzothiophene derivatives by whole cells of Rhodococcus erythropolis H-2". FEMS Microbiology Letters, 1996. 142: p. 65-70. Rhee, S.K., et al. , "Desulfurization of dibenzothiophene and diesel oils by a newly isolated Gordona strain CYKS 1 'I. Applied and Environmental Microbiology,

Setti, L., et al. , "The effect of n-alkanes in the desulfurization of dibenzothiophene and of organic sulfur compounds in heavy oil by Pseudomonas SP.". Biotechnology Letters, 1992.14(6): p. 5 15-520. Premuzic, E.T. and M. Lin, "Process for producing modified microorganisms for oil treatment at high temperatures, pressures and salinity," . 1996: USA. p. 33. Almarsson, 0. and A.M. Klibanov, "Remarkable activation of enzymes in nonaqueous media by denaturing organic solvents". Biotechnology and Bioengineering, 1996.49: p. 87-92. Scott, C.D., et al. , "Bioprocessing in Nonaqueous Media Critical Needs and Opportunities," . 1995, Oak Ridge National Laboratory. Woodward, C.A. and E.N. Kaufinan, "Enzymatic catalysis in organic solvents: polyethylene glycol modified hydrogenase retains its sulfhydrogenase activity in toluene". Biotechnology and Bioengineering, 1996.52(3): p. 423-428. Perry, R.H., D.W. Green, and J.O. Maloney, eds. Perry's Chemical Engineering Handbook. 6 ed. . 1984, McGraw-Hill: New York. 2336. Chemineer, "Kenics Static Mixers KTEK Series," . 1988: North Andover, MA. p. 32. Byers, C.H. and A. Ammi, "Understand the potential of electro-separations". Chemical Engineering Progress, 1995. 91(2): p. 63. Kowalski, W. and 2. Ziolkowski, "Increase in rate of mass transfer in extraction columns by means of an electric field". Int. Chem, Eng. , 198 1.21 : p. 323. Martin, L., et al., "Electrical field contactor for solvent extraction". Sep. Sci. Technol., 1983.18: p. 1455. Scott, T.C. and R.M. Wham, "An electrically driven multistage countercurrent solvent extraction device: the emulsion phase contactor". Ind. Eng. Chem. Res., 1989.28: p. 94. Scott, T.C., D.W. DePaoli, and W.G. Sisson, "Further development of the electrically driven emulsion phase contactor". Ind. Eng. Chern. Res. , 1994. 33: p.

Thorton, J.D., T h e application of electrical energy to chemical and physical rate processes". Rev. Pure and Appl. Chem. , 1968. 18: p. 197. Ptasinski, K.J. and P.J. Kerkhof, "Electric field driven separations: phenomena and applications". Sep. Sci. Technol., 1992.27(8 & 9): p. 995. Scott, T.C. and W.G. Sisson, "Droplet size characteristics and energy input requirements of emulsions formed using high intensity pulsed electric fields". Separation Science and Technology, 1988.23: p. 1541-1550.

107: p. 31-36.

1998.64(6): p. 2327-2331.

-

1237-1244.

44.

45.

46.

47.

48.

49.

Process Considerations in Biodesulfuization, Borole and Kaufian. 13

Warren, K.W. and F.L. Prestridge. "Crude oil desalting by countefflow electrostatic mixing". in National Petroleum Refiners Association Annual Meeting. 1988. San Antonio, TX: National Petroleum Refiners Association. Schwartz, E., et al. "An improved process for MTBE unit feed pretreatment". in First SeDarations Division Topical Conference on SeDarations technologies: New develouments and Omortunities. 1992: American Institute of Chemical Engineers. Wood, T.S., et al. "Rapid, automated liquid / liquid extraction of contaminants fiom aqueous samples containing suspended solids". in First SeDarations Division Touical Conference on SeDarations technologies: New develonments and Otmortunities. 1992: American Institute of Chemical Engineers.. L i m a , H.M., T.C. Scott, and C.D. Scott, "Apparatus and Method for the Desulfurization of Petroleum by Bacteria,'' . 1995, United States Patent Number 5,458,752. Kaufinan, E.N., et al. "Biodesulfurization of dibenzothiophene and crude oil using electrospray reactors". in 3rd International Petroleum Environmental Conference. 1996. Albuquerque, NM. Kaufinan, E.N., J.B. Harkins, and A.P. Borole. "Biodesulfurization of Dibenzothiophene and Crude Oils Using Batch Stirred and Electro-spray Reactors". in International Petroleum Environmental Conference. 1997. San Antonio, TX.

Process Considerations in Biodesulfuization, Borole and KauJsnan. 14

Figure Captions

Figure 1: Biochemical pathway showing the sequence of DBT oxidation by IGTSI.

Figure 2: Schematic of the emulsion phase contactor and ancillary equipment.

Figure 3: Effect of cell density on 2-HBP production rate. The 0.5X - 15X refers to cell density and 1X corresponds to 16.7 g cell paste in 250 mL aqueous phase. As observed from this plot, mass transfer may be a limiting factor in the DBT desuikization process.

Figure 4: Effect of aeration rate and oxygen tension on 2-HJ3P production in batch stirred reactor. The experiment was conducted by dispersing 250 mL aqueous phase containing 166.7 g cell paste (IOX cell density) in 750 mL hexadecane containing 0.6 wt. % DBT.

Figure 5: Comparison of 2-HBP production rate in EPCTM and BSR. The BSR experiment was conducted by contacting 250 mL aqueous phase containing 166.7 g cell paste (1OX cell density) in 750 mL hexadecane containing 0.6 wt. DBT. In the EPCW experiment, 107 mL aqueous phase containing 66.7 g cell paste (1OX) was contacted with 2400 mL organic phase.

Figure 6: Micrograph of an EPCm sample collected at 4 hours after beginning of reactor operation. The sample was collected from the central region between the electrodes using a Teflon tubing connected to a syringe and observed under a microscope as described earlier. The average droplet size was measured to be 2.54 k 2.40 pm (n=300).

Figure 7. Analysis of the aromatic fraction of Van Texas crude oil by GC-SCD. The biotreatment results in removal of low molecular weight DBT-type compounds; however, the higher molecular weight compounds are not affected.


ORNL OWG 97C-263

2Sfydroxybiphenyl (2-H 5P’)

Figure 1: Biochemical pathway showing the sequence of DBT oxidation by IGTSS.

P

Process Considerations in Biodesulfirization, Borole and Kaujhan. 16

ZWmi Jmin 300C.Or

Figure 2: Schematic of the emulsion phase contactor and ancillary equipment.

ZOmUmin., Air

Process Considerations in Biodesulfirization, Borole and Kaufinan. 17

0

cell density

5 10 15 20 25 nmc, h

Figure 3: Effect of cell density on 2-HBP production rate. The 0.5X - 15X refers to cell density and 1X corresponds to 16.7 g cell paste in 250 mL aqueous phase. As observed from this plot, mass transfer may be a limiting factor in the DBT desulfurization process.


14

12

10

a 6

4

2

- /z/ -0.2 s l p air, r30

-1.Oslpmair. r34 ,l.OSlpn oxygen, r34

Figure 4: Effect of aeration rate and oxygen tension on 2-HBP production in batch stirred reactor. The experiment was conducted by dispersing 250 mL aqueous phase containing 166.7 g cell paste (IOX cell density) in 750 mL hexadecane containing 0.6 wt. % DBT.

Process Cohsiderations in Biodesul&rization, Borole and Kaufinan. 19

I

0 1 2 3 4 5 0 7 8 Tlme, h

Figure 5: Comparison of 2-HBP production rate in EPCn and BSR. The BSR experiment was conducted by contacting 250 mL aqueous phase containing 166.7 g cell paste (1OX cell density) in 750 mL hexadecane containing 0.6 wt. DBT. In the E X T M experiment, 107 mL aqueous phase containing 66.7 g cell paste (1OX) was contacted with 2400 mL organic phase. !-

Process Considerations in Biodesulfirization, Borole and KauJinan. 20

Figure 6: Micrograph of an EPCfM sample collected at 4 hours after beginning of reactor operation. The sample was collected from the central region between the electrodes using a Teflon tubing connected to a syringe and observed under a microscope as described earlier. The average droplet size was measured to be 2.54 f 2.40 pm (n=300).

Process Considerations in Biodesulfurization, Borole and Kaufkzan.

3 B

d e u n v)

8.Oe.5

7.oe5

6.0eS

s.oe5

4.0eS

3.oes

2.M

21

0 5 10 IS 20 25

Time (min.)

Figure 7. Analysis of the aromatic fraction of Van Texas crude oil by GC-SCD. The biotreatment results in removal of low molecular weight DBT-type compounds; however, the higher molecular weight compounds are not affected.

process considerations in the biodesulfurization of crude oil

Documents