Fractionation of Peace River Bitumen Using Supercritical Ethane

Ind. Eng. Chem. Res. 2000, 39, 3875-3883

3875

Fractionation of Peace River Bitumen Using Supercritical Ethane and Carbon Dioxide J. L. Rose, W. Y. Svrcek,* and W. D. Monnery Department of Chemical and Petroleum Engineering, University of Calgary, 2500 University Drive N.W., Calgary, Alberta T2N 1N4, Canada

K. Chong Core Laboratories Canada Ltd., 2810 12th Street N.E., Calgary, Alberta T2E 7P7, Canada

The purpose of this work was to provide valuable data for the use of supercritical fluids in both surface and in situ bitumen and heavy oil recovery operations. Experiments were conducted using a semibatch extractor packed with a mixture of bitumen and sand, and the experiments were performed such that thermodynamic equilibrium governed the extraction process. Experimental temperatures and pressures were varied in order to observe their impact on the ability of supercritical ethane and carbon dioxide to recover oil from the bitumen-sand mixture. The results showed that ethane consistently outperformed carbon dioxide in terms of its ability to extract oil from the bitumen-sand mixture. In all cases, supercritical ethane produced yields that were much greater than those obtained with supercritical carbon dioxide. Analysis of the extracted material revealed that the bitumen extracted using ethane was much heavier and contained a wider range of hydrocarbons. Introduction Supercritical fluids are ideal candidates for solvents in extraction processes because they combine properties of both gases and liquids, which greatly improves their solvent capabilities.1 Near the critical point, the properties of supercritical fluids are strong functions of temperature and pressure.2 The benefits of these characteristics have been well documented and include the following: increased solubilities; lower solvent-to-feed ratios; easier solvent recovery; reduced mass-transfer resistance; and lower operating temperatures, resulting in lower energy requirements. Finally, supercritical fluids also permit the processing of temperature-sensitive products.2-5 Resource companies are relying on alternative sources as means of satisfying the world’s demand for petroleum and petrochemical products because of the continual consumption of conventional crude supplies. New sources of hydrocarbons such as bitumens, heavy oils, oil shales, and coal have properties that are significantly different from those of traditional crude oils.6 Therefore, to recover and process these new fluids in a cost-efficient manner that is also environmentally responsible, new technologies must be developed. Processes that exploit the enhanced solvent capabilities of supercritical fluids represent one area being explored in an attempt to address these concerns. The benefits of employing supercritical fluids for processing hydrocarbon mixtures is evident in the success of the ROSE process and propane deasphalting.2 These processes rely on the solvent behavior of fluids near their critical points to remove asphaltenes from * Author to whom correspondence should be addressed. Phone: (403) 220-5751. Fax: (403) 284-4852. E-mail: [email protected].

streams such as lube oil feeds. Research is being conducted in an attempt to apply supercritical fluids to other areas of the petroleum industry. Such areas include the following: processing and liquification of coal;7,8 extraction of oil from oil shales;9-11 fractionation of crude oils, bitumens, and wax-bearing residue;12-15 characterization of bitumens and heavy hydrocarbon mixtures;16 and in situ production of heavy oils.17 Supercritical fluid technology is still not widely used on an industrial scale, even though the benefits of using supercritical fluids are known and examples of their successful use have been reported.2 One of the reasons for this discrepancy is the dearth of data for supercritical fluid systems.18 Without the experimental data, the development of models and equations that can be used to design and evaluate these new technologies is limited. Therefore, the purpose of this project was to fill a portion of this void by providing data for systems composed of supercritical fluids and bitumens. An apparatus was designed to be capable of producing data for the extraction of bitumens using different supercritical fluids. Two supercritical fluids, ethane and carbon dioxide, were used to extract the soluble components of a bitumen from the Peace River region located in northern Alberta, Canada. Ethane was selected as one of the solvents by virtue of the facts that the critical temperature for ethane (32.2 °C) is near ambient and the critical pressure is relatively low (4.88 MPa). Carbon dioxide was selected because it is a common fluid used in enhanced oil recovery methods19 and a popular solvent for commercial supercritical fluidextraction processes.2 Advantages of using carbon dioxide include the low solvent cost, nontoxicity, low critical temperature, and ready availability.1,20 Experiments were conducted to compare the capabilities of supercritical ethane and carbon dioxide as solvents for the extraction of oil from bitumen.

10.1021/ie000320r CCC: $19.00 © 2000 American Chemical Society Published on Web 09/09/2000

3876

Ind. Eng. Chem. Res., Vol. 39, No. 10, 2000

Figure 1. Schematic of experimental apparatus used in this work.

Experimental Apparatus and Procedure The apparatus used in this work incorporated a semibatch extractor in order to perform dynamic extraction studies of a bitumen-sand mixture using supercritical fluids. The apparatus was designed such that the solvent exiting the extractor was saturated with the soluble components of the bitumen. The experimental apparatus allowed for samples of the extracted oil to be collected at different time intervals and analyzed to determine how different operating temperatures and pressures affected the extraction process and to observe how the sample properties changed as the extraction proceeded. The experimental apparatus used in this study is shown in Figure 1. For the bitumen extractions, the extractor was packed with 940 g of the bitumen-sand mixture. The sand and bitumen were combined in the lab such that the final mixture was 15 wt % bitumen, a bitumen-to-sand ratio typical of the reserves found in areas such as Peace River and the Athabasca region in northern Alberta.6 A positive displacement pump was used to bring the system up to pressure and then supply solvent at a set flow rate. The pump had two cylinders, each with a volume of 0.5 L and a quick-change transmission that allowed the flow of the fluid to be fixed to one of 28 possible rates by adjustment of the gears. The flow rates ranged from 0.005 to 1.12 L/h per cylinder at the pump operating pressure and temperature, and the maximum pressure that the pump could achieve was 27.5 MPa. The extractor and the majority of the tubing were placed inside a Blue M DC-166F oven, which maintained the temperature of the extractor and preheated the solvent to the desired operating temperature. The temperature was accurately controlled, with the average standard deviation in the operating temperature for all of the experiments being (0.45 °C. The solvent was preheated in the oven; it entered the bottom of the extractor and flowed upward through the extractor, removing the soluble portion of the bitumen. The resulting light phase exited the extractor and passed

through a micrometer valve, CV1 in Figure 1, that was used to manually control pressure. The micrometer valve was also used to decrease the pressure from the operating pressure to atmospheric pressure, causing the extracted bitumen to precipitate. The extracted samples were collected in two phase separators that were exchanged at 20-min intervals over the course of the experiment. The time interval was chosen such that there was sufficient oil in the separators for analysis. The separators were weighed before and after to determine the mass of oil extracted. The bitumen-free ethane exited through the top of the separators and passed through a wet test meter before being vented. At the end of a run, the system was depressurized, and the bitumen-sand mixture remaining in the extractor was collected and weighed, thus closing the mass balance. The mass balances for the experiments closed to within an average of -1.5 wt % of the initial feed, with the maximum loss being only -2.2 wt %. Three runs were performed at each temperature and pressure, with an average absolute relative deviation of in the amount of bitumen extracted being only (2.1%. The key piece of equipment was the semibatch extractor, which was the contact vessel for the supercritical solvent with the bitumen-sand feed. The extractor was constructed at the University of Calgary of 2-in. o.d. Schedule 160 pipe made of 316 stainless steel. The extractor was rated up to a maximum pressure of 21.8 MPa at a temperature of 204 °C. The extractor was connected to the other parts of the apparatus using 1/4in. NPT/Swagelock fittings, and a port was added to the side of the extractor to facilitate the attachment of a thermocouple used to monitor the temperature inside the extractor. Gas distribution in the extractor was aided by the presence of a 0.3-cm-thick distributor preceded by a 4.0-cm-deep wind-box section. The distributor was constructed of porous stainless steel with an average pore size distribution of 2 µm. The porous steel was welded to the walls of the extractor to ensure

Ind. Eng. Chem. Res., Vol. 39, No. 10, 2000 3877

that the solvent could not bypass the distributor by flowing around the edges of the distributor plate. A residence time distribution study was performed, verifying that the solvent flowed through the packed extractor in a plug-flow manner.21 The operating pressure was controlled manually using a micrometer valve (CV1 in Figure 1). During the experiments, the pressure was dropped across the control valve from the operating pressure down to atmospheric pressure. This caused the temperature at the valve to decrease, thus requiring that the valve be heated using heat tape to prevent plugging in the casing of the valve. Pressure was controlled accurately for all experiments, with the average standard deviation in pressure being (0.04 MPa. Two pressure transducers attached to the computer data acquisition system were used to monitor and record the system pressures. A Rosemount Alphaline Model 1151 GP pressure transducer was used to record the pressure at the pump (P1), and a second Rosemount transducer (P2) was used to measure the extractor pressure. The temperatures in the system were monitored and recorded using T-type special thermocouples attached to a computer data acquisition system. Separators were attached to the system downstream of the micrometer valve with glass wool at the top to prevent any entrained oil from leaving the separator. The separators were placed in an ice bath during operation to reduce the temperature and enhance the precipitation of the extract. The performance of the separators was verified by analyzing the exit gas before it entered the wet test meter to verify that all of the bitumen extracted was collected by the separators.21 A binary system was used to evaluate the experimental apparatus. Solubility of n-hexadecane in CO2 has been reported in the literature12,22 and used by other researchers to verify the operation of their apparatuses.12 Literature data for the CO2/hexadecane system was published at two temperatures, 32 and 38 °C, and pressures ranging from 7.6 to 17.2 MPa. Experiments were performed over the above conditions, and the resulting solubilities of hexadecane in CO2 are shown in Figure 2. The results obtained were in excellent agreement with published data, having an average absolute deviation of 3.7% in the n-hexane solubility. On the basis of the results from the pure component study, it was concluded that the apparatus would provide reliable solubility data. The residence time distribution study demonstrated that the solvent was distributed evenly across the extractor and contacted all of the bitumen. Therefore, there was a high degree of confidence that the apparatus could provide equilibrium data that was accurate for the Peace River bitumen-supercritical ethane system. Materials The ethane used in the extraction experiments was purchased from Praxair Distribution Inc. and had a purity greater than 99.0 mol %. The hexadecane was obtained from Aldrich Chemical Co. and was 99% pure. No further purification of the n-hexadecane was performed. The carbon dioxide was purchased from Praxair and had a purity of 99.5 mol % with a maximum moisture content of 34 ppm. The sand was purchased from Target Products Ltd. of Calgary and was produced at a plant in Morinville, Alberta. The sand is predominantly silica (93 wt %),

Figure 2. Solubility of n-hexadecane in CO2 as a function of pressure at temperatures of 32 and 38 °C. Table 1. Properties of the Peace River Bitumen Used in This Work density at 15 °C initial boiling point final boiling point percent residue total sulfur pour point Conradson carbon residue C5 asphaltenes viscosities saturates aromatics resins asphaltenes

998.3 kg/m3 10.2° API 160 °C 584 °C 49.52% 5.5 wt % +3 °C 8.9 wt % 18.4 wt % 10270 cSt at 20 °C 3718 cSt at 40 °C 1506 cSt at 40 °C 26.96 wt % 42.65 wt % 15.03 wt % 15.36 wt %

with a small amount of alumina (4 wt %) and trace amounts of oxides. The sand was classified as 20-40 mesh with an average diameter of 0.663 mm. The density of the sand was 1640 kg/m3, and the porosity was 0.4. The properties of the sand were typical of the sand found in northern Alberta and sand used by other researchers when conducting heavy oil recovery experiments.23 The bitumen feedstock for this study was Peace River bitumen obtained from the Peace River facility located in northern Alberta, operated by Shell Canada Limited. The bitumen properties are summarized in Table 1. Results and Discussion The first step in the experimental process was to verify that the bitumen extracted was controlled by thermodynamic equilibrium and not limited by masstransfer resistances. A series of experiments were conducted at 47 °C and 10.5 MPa during which different masses of Peace River bitumen were put into contact with supercritical ethane at different flow rates. The

3878

Ind. Eng. Chem. Res., Vol. 39, No. 10, 2000 Table 2. Experimental Temperatures and Resulting Solvent Densities at 10.5 MPa temperature, °C

density, kg/m3

13.90 47.78 63.27 76.71 92.10

360.57 331.60 281.76 234.04 187.98

Table 3. Experimental Pressures and Resulting Solvent Densities at 47 °C pressure, MPa

density, kg/m3

7.3 10.5 12.2 15.0

263.68 331.53 348.42 366.50

Figure 3. Extraction curves for experiments performed at 47 °C and 10.5 MPa using different amounts of bitumen and different solvent flow rates.

solvent flow rate was constant for each extraction and set at rates ranging from 0.2 to over 2.24 L/h, and the masses of bitumen used were 35 and 140 g. Figure 3 shows the extraction curves from these experiments. The data are plotted as the cumulative percent of bitumen extracted as a function of cumulative volume of ethane measured by the wet test meter, corrected to standard conditions, and divided by the mass of bitumen in the extractor. The changes in the solvent flow rate and mass of bitumen caused the contact time between the supercritical ethane and the bitumen to vary. The contact times were varied from a low of approximately 1 min to conditions under which the contact time approached 42 min, an increase of over 40 times. Figure 3 shows that the amount of bitumen recovered was not dependent upon either the solvent flow rate or the amount of bitumen in the extractor. Therefore, the mass extracted over this range of contact times was independent of the solvent-bitumen contact time, thus indicating that the data collected were controlled by thermodynamic equilibrium and not mass-transfer resistances. If the system had been in the mass-transfer-dominated regime, then the amount of bitumen extracted would have been timedependent and thus a function of the solvent flow rate and/or the amount of bitumen in the extractor. For example, as the flow rate increased, the time available for the solvent to contact the bitumen would decrease, resulting in less oil being recovered from the extractor. Because this was not the case for this system, the results indicated that, when the solvent exited the extractor, it was saturated with the soluble fraction of bitumen and thus in thermodynamic equilibrium. Ethane Extraction Results. A series of experiments were conducted to examine how the extraction of Peace River bitumen using supercritical ethane responded to variations in operating temperature and pressure. Both operating variables have a strong affect on the proper-

Figure 4. Weight percent of extracted bitumen as a function of cumulative volume of ethane measured for extractions performed at 10.5 MPa and temperatures ranging from 37 to 92 °C.

ties of supercritical ethane, especially near the critical point.1,2 Five different temperatures were investigated in this study, ranging from 36.9 to 92.1 °C, at a constant pressure of 10.5 MPa and at four pressures ranging from 7.3 to 15.0 MPa at 47 °C. These conditions, along with the corresponding solvent densities are summarized in Tables 2 and 3. The effect of temperature on the cumulative quantity of oil recovered from the bitumen as a function of the cumulative volume of ethane measured by the wet test meter is shown in Figure 4. This figure shows that, as the temperature decreased, the amount of bitumen extracted increased. A comparison of the cumulative weight percent extracted after approximately 175 std L of ethane had been injected showed that the weight percent of bitumen extracted increased from 4.5 wt % at 92 °C to 30.7 wt % at 36.9 °C, an approximate 6-fold increase. These results were consistent with the data for binary systems consisting of supercritical ethane and a heavy hydrocarbon. For example, Schmitt and Reid collected solubility data for acridine, 1,4-naphthoqui-


Figure 5. Weight percent of extracted bitumen as a function of cumulative volume of ethane measured for extractions performed at 47 °C and pressures ranging from 7.3 to 15.0 MPa.

none, and naphthalene in a number of supercritical fluids, including ethane, over a wide range of temperatures and pressures and for isobars above but less than approximately twice the critical pressure. They observed that, as the temperature increased, the solubilities decreased.24 Similar results were presented by McHugh et al. for naphthalene and supercritical carbon dioxide.2 The data shown in Figure 4 for Peace River bitumen and supercritical ethane exhibited trends similar to those seen for the binary data at pressures such as 10 MPa, approximately twice the critical pressure of ethane. The effect of pressure on the amount of bitumen dissolved in supercritical ethane at a constant temperature of 47 °C is shown in Figure 5. This figure clearly illustrates that, as the operating pressure increased, the amount of bitumen extracted also increased. For example, at 7.3 MPa, approximately 5 wt % of the bitumen was extracted after the bitumen contacted 50 std L of ethane. When the pressure was increased to 15.0 MPa, the amount of bitumen recovered with the same amount of ethane increased to 20 wt %, or by approximately four times. Again, the increase in the amount of oil extracted was expected on the basis of the results of the numerous binary systems that have been discussed in the literature.22,24,25 The published data revealed that, for a given temperature, the solubility of heavy hydrocarbons in supercritical fluids increases as the experimental pressure increases, as did the solubility of Peace River bitumen in supercritical ethane, as shown in Figure 5. The changes in the amount of extractable oil can be explained by examining the shift in the nature of the system as changes were made to the operating temperatures and pressure. At lower temperatures and higher pressures, the density of ethane approached the standard liquid density of ethane, which is 356 kg/m3. Under these conditions, the system behaved similar to a liquid-liquid extraction process. Liquid solvent systems

Figure 6. Cumulative yield for 130 std L of solvent versus solvent reduced density.

have higher solubilities because of the increased attractive forces compared to vapor solvents;24 therefore, the solubility of the bitumen fractions was much higher, and more bitumen was recovered. Conversely, at higher temperatures and lower pressures, the solvent density decreased to values much lower than the liquid density. With such a low solvent density, the system now acted similar to a vapor-liquid extraction process, and as a result, the extraction capacity of the solvent decreased, resulting in less bitumen being recovered. Also of importance is the question of whether the amount of extracted bitumen changed solely because of changes in the solvent density or whether variations in the temperature and/or pressure caused other changes in the behavior of the system. Deo et al. reported that temperature had additional effects on the yields produced for bitumen-supercritical solvent systems.12 They reported that supercritical propane exhibited enhanced solvent properties when extractions were conducted at temperatures near the critical temperature of propane. It was concluded that the increased solubilities of bitumen could be attributed to the proximity of the operating temperature to the solvent critical temperature as well as solvent density. This contradicted the conclusions drawn by Subramanian et al. who found that, for the bitumen-supercritical propane systems studied, the extractions were dominated chiefly by the pure solvent density.13 The data from this work have been presented in terms of reduced density to determine which of the previous conclusions pertained to the Peace River bitumensupercritical ethane system. In Figure 6, the weight percent of bitumen extracted with 130 std L of ethane solvent has been plotted as a function of the reduced solvent density. The extraction yields increased with increases in density for all cases, demonstrating that the extraction yields were a strong function of the pure solvent density and that there were no enhanced solvent

3880


Figure 7. Boiling point curves for selected samples extracted at 10.5 MPa and temperatures of 37 °C (P) and 63 °C (A).

properties when the extractions were performed near the critical temperature, as had been observed by Deo et al. Therefore, on the basis of the results from this work and that of Subramanian et al., it can be concluded that the extraction of bitumen was controlled predominantly by the density of the pure supercritical solvent. The effects of temperature and pressure on the composition of the extracted bitumen were shown by the boiling point curves for the extracted samples. Figure 7 presents the boiling point curves for samples taken during the first, third, and fifth extraction windows at a constant pressure of 10.5 MPa and temperatures of 37 and 63 °C. This plot shows that the composition of the extracted bitumen changed with changes in the operating conditions. The oil produced when the temperature was 37 °C was consistently heavier than the oil produced at 63 °C, as shown by the higher boiling point curves for the 37 °C samples. This conclusion can be drawn because of the fact that heavier molecular weight hydrocarbons generally have higher boiling points. Similar trends were observed when the operating temperature was kept constant and the pressure changed.21 At lower pressures, the bitumen extracted was lighter, and it became progressively heavier as the operating pressure increased. As the operating temperature was increased or the pressure decreased, the composition of successive extracted oil samples became more uniform. This was evident in the initial and final boiling points for the samples shown in Figure 7. For example, samples A1 and P1 were obtained after the bitumen-sand mixture contacted the same amount of ethane but at temperatures of 63 and 37 °C, respectively. Whereas both samples have the same initial boiling point, the sample produced at 63 °C (A1) has an end point 120 °C lower than that of the sample P1, which was extracted at 37 °C. The large difference between the initial and final boiling points for P1 implies that this sample contained a wider range of hydrocarbons. Therefore, not only was

more oil recovered as the temperature was lowered or the pressure increased, but the range of components recovered was also greater under these conditions.21 The composition of the extracted samples varied during the course of every extraction performed. For example, in Figure 7, the sample A1 has a lower boiling point curve than sample A3, which in turn is lower than that of A5; therefore, as the extraction proceeded, the oil produced became progressively heavier. The magnitude of these changes was related to the conditions under which the extraction was performed. A similar trend was observed during the experiments performed at different temperatures and pressures.21 The results of the composition analysis show one of the benefits of extracting oil from bitumen using supercritical fluids. As the pressure was increased or the temperature decreased, the oil extracted became increasingly heavier. Hence, by manipulating the extraction temperature or pressure, the characteristics of the oil extracted can be controlled. For example, a low initial operating pressure would produce oil that was relatively light, and by a gradual increase in the pressure, the product could be made heavier. At a given operating condition, the composition of the extract changed in a very specific manner; that is, the bitumen could be separated from the sand mixture and fractionated all in one step. Carbon Dioxide Extraction Results. Oil was also extracted from Peace River bitumen using supercritical carbon dioxide over a range of temperatures and pressures. The operating conditions for the experiments were chosen in order to make meaningful comparisons between the performance of ethane and carbon dioxide. Conducting experiments at different temperatures and pressures would show if carbon dioxide behaved similarly to ethane with respect to changes in the operating conditions. Temperatures and pressures were selected in order to compare the solvents under conditions such that the actual and reduced densities were of the same magnitude. Figure 8 presents the extraction curves for experiments conducted using carbon dioxide as the solvent at constant a temperature of 34 °C and pressures of 10.0 and 15.0 MPa and at a constant pressure of 12.2 MPa and temperatures ranging from 34 to 55 °C. This plot shows that the yields increase as the operating pressure increases and as the operating temperature decreases. The increase in the oil extracted was consistent with the solubility behavior of pure hydrocarbons in supercritical carbon dioxide, as per the work of Schmitt et al. and Johnston et al., who revealed that the solubility of a number of different hydrocarbons increased as the extraction pressure was increased and also increased as the extraction temperature decreased.24,25 Therefore, the data obtained for the solubility of Peace River bitumen in the supercritical carbon dioxide were consistent with results for pure hydrocarbon-carbon dioxide systems. Compositional variations of the extracted oil produced with supercritical carbon dioxide were also observed using in boiling point curves for the extracted oil. These data indicated that the composition of the extracted material changed marginally over the range of experimental temperatures and pressures.21 For example, when the operating pressure was increased, the samples became slightly heavier, as indicated by the increase in the boiling point curves for the extracted material


Figure 8. Cumulative weight percent of bitumen extracted using supercritical carbon dioxide at temperatures ranging from 34 to 55 °C and pressures from 10 to 15.0 MPa.

Figure 9. Boiling point curves for selected samples extracted at 34 °C and pressures of 10.0 MPa (C) and 15.0 MPa (J).

obtained at 15.0 MPa (J) as compared to those obtained at 7.3 MPa (C), as shown in Figure 9. The compositions of successive samples obtained during the course of an extraction performed using carbon dioxide were relatively uniform. This is evident from a comparison of the boiling point curves in Figure 9. The first samples have boiling point curves that are almost identical to the boiling point curves for the third samples, thus revealing that the compositions of the samples were indeed

Figure 10. Comparison of extraction yields using supercritical ethane and carbon dioxide at 47 °C and 12.2 MPa.

similar and did not change appreciably over the course of the extraction. The implication of the compositional results was that the choice of solvent had a significant impact on the composition of bitumen that could be extracted. The range of hydrocarbons extracted with supercritical carbon dioxide was smaller than that obtained when ethane was used. The fact that changes in the operating conditions had a negligible impact on the composition of the bitumen extracted meant that the composition could not be manipulated as successfully with supercritical carbon dioxide as it could when when supercritical ethane was used as the solvent. Comparison at the Same Operating Conditions. The experimental conditions under which the carbon dioxide experiments were performed were selected to provide a realistic comparison of the two solvents. Experiments were performed using both solvents at the same temperature and pressure. Data were generated that were used to compare the solvents at conditions such that their reduced and actual densities were the same. Figure 10 shows the two extraction curves produced when supercritical ethane and carbon dioxide were used as the solvent contacting the bitumen at 47 °C and 12.2 MPa. The supercritical ethane extracted almost six times more oil than carbon dioxide for the same volume of solvent. Even after more than 1.5 times more carbon dioxide than ethane was injected, the cumulative weight percent extracted was only 7 wt % for the carbon dioxide compared to approximately 30 wt % for ethane. This was not overly surprising, as studies of binary systems showed that hydrocarbons had higher solubilities in supercritical ethane than in carbon dioxide for the same temperature and pressure increase.24,25 Figure 11 shows the boiling point curves for selected samples produced using supercritical ethane (Q) and carbon dioxide (G). The material extracted with carbon

3882


Figure 11. Boiling point curves for samples obtained using ethane (Q) and carbon dioxide (G) at 47 °C and 12.2 MPa.

dioxide was much lighter, as indicated by the lower boiling point curves for the carbon-dioxide-produced samples. The bitumen extracted with carbon dioxide was also more uniform in composition than the bitumen obtained using ethane. Hence, the boiling point curves for the samples produced using ethane cover a wider range of temperatures, indicating a wider range of hydrocarbon constituents in these samples. The composition of the samples obtained using ethane changed more over the duration of the extractions than did the composition of samples obtained when carbon dioxide was the solvent. This fact is attested in Figure 11, where the difference between the boiling point curves for the ethane-produced samples is greater than that for the samples extracted using carbon dioxide. These results suggest that a greater degree of fractionation occurs during the extractions conducted using supercritical ethane. Again, this implies that the flexibility to control the composition of the extracted bitumen was greater when ethane, as opposed to carbon dioxide, was used as the solvent because the change in the composition of the extracted material was greater during the course of the ethane extractions. Comparison at the Same Reduced Density. The next phase in the comparison of the two solvents was a comparison of the results obtained when the operating conditions were such that the densities of the two solvents were the same. The maximum ethane density used was 366.8 kg/m3. Using the NIST database,26 it was possible to calculate a set of temperatures and pressure that would result in the density of supercritical carbon dioxide being 366.8 kg/m3. Unfortunately, experiments performed using carbon dioxide at this density resulted in negligible amounts of oil being extracted. Therefore, it can be concluded that bitumen was not significantly soluble in supercritical carbon dioxide when the solvent density was similar to that of supercritical ethane.

Figure 12. Comparison of yields between supercritical ethane and carbon dioxide at the same reduced densities of 1.37 and 1.76.

The lowest carbon dioxide density used in these experiments was 520.3 kg/m3. At these conditions, only 3 wt % of the bitumen was extracted. To achieve such a high density for ethane, the temperature would have to be as low as possible and the pressure as high as possible. The temperature could not be reduced below the critical temperature if ethane were to remain supercritical. At 32 °C, the critical temperature of ethane, the operating pressure required to achieve the desired density would be well over 70 MPa. This was above the pressure rating for the apparatus, and thus these experiments could not be performed. It was possible to produce extraction data for carbon dioxide and ethane at the same reduced densities. Data from these experiments are presented in Figure 12 for two cases for which the reduced densities of both solvents were the same. The data show the extraction curves for reduced solvent densities of 1.37 and 1.76. Ethane consistently extracted more bitumen than carbon dioxide, even though the reduced densities were equal. Approximately three times more oil was produced using supercritical ethane than using the same amount of carbon dioxide. Boiling point curves for selected samples produced under conditions such that the reduced densities of the solvents were both 1.37 were measured.21 Again, the results indicate that the bitumen extracted using ethane was heavier and contained a wider range of constituents than that produced using carbon dioxide. In addition, there was a greater degree of fractionation when ethane was used as the solvent, as evidenced by the greater difference in the two samples produced using ethane. Conclusions A review of the published literature on supercritical fluids indicated that supercritical fluid technology could address a number of process issues; however, their


application to industrial-scale processes remains quite limited. This is primarily the result of a lack of both experimental data and accurate models capable of describing the behavior of supercritical systems, especially those involving complex, undefined hydrocarbon mixtures such as bitumens and heavy oils. An apparatus capable of providing accurate equilibrium data for the extraction of bitumen from a bitumen-sand mixture using supercritical solvents was designed and constructed. The experimental apparatus provided excellent pressure and temperature control, exhibited excellent repeatability, and was able to produce solubility data that were in good agreement with published values. A novel method was developed for verifying the assumption that the process was controlled by thermodynamic equilibrium rather than masstransfer resistances. New data for the systems comprised of Peace River bitumen and two solvents, ethane and carbon dioxide, were collected. Experimental results indicated that the extraction process was dependent on the operating conditions. Extraction yields could be increased from 15 wt % recovered to 33 wt % by increasing the extraction pressure and from 5 wt % to 31 wt % by decreasing the operating temperature. As temperature was decreased and pressure increased, the extracted oil became heavier, as reflected by the changes in the boiling point curves of the extracted bitumen. These trends were also observed over the course of a single extraction run in which the samples obtained later in the extraction were the heaviest. This indicated that one of the benefits to using supercritical extraction was that not only was the bitumen being separated from the sand, but also a degree of fractionation occurred during the extraction process. Extractions were also conducted using supercritical carbon dioxide as the solvent, and the new data were used to compare the solvent ability of supercritical ethane with another common supercritical fluid. Although both solvents exhibited the same trends with respect to changes in pressure and temperature, the recoveries for ethane were more that six times greater, indicating that supercritical ethane would be a much better solvent that carbon dioxide for the processing of complex hydrocarbon mixtures such as bitumens and heavy oils. Acknowledgment The authors thank NSERC for the financial support required to complete this work. A special thanks is also given to Core Laboratories Canada Inc. in Calgary, Alberta, for their assistance with the analysis of the bitumen and the extracted samples. Literature Cited (1) Taylor, L. T. Supercritical Extraction; John Wiley & Sons: New York, 1996. (2) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction: Principles and Practice, 2nd ed.; Butterworth: Stoneham, MA, 1994. (3) Parkinson, G.; Johnson, E. Supercritical Processes Win CPI Acceptance. Chem. Eng. 1989, July, 37. (4) Hoyer, G. G. Extraction with Supercritical Fluids: Why, How, and so What? CHEMTECH 1985, July, 440.

(5) Williams, D. F. Extraction with Supercritical Gases. Chem. Eng. Sci. 1981, 36, 1769. (6) Speight, J. G. The Chemistry and Technology of Petroleum, 2nd ed.; Marcel Dekker: New York, 1991. (7) Sunol, A.; Beyer, H. Mechanism of Supercritical Extraction of Coal. Ind. Eng. Chem. Res. 1990, 29, 842. (8) Kershaw, J. R.; Bagnell, L. J. Extraction of Australian Coals with Supercritical Aqueous Solvents. In Supercritical Fluids: Chemical and Engineering Principles and Applications; Squires, T. G., Paulaitis, M. E., Eds.; American Chemical Society: Washington, D.C., 1987. (9) Kesavan, S. K.; Ghosh, A.; Polasky, M. E.; Parameswaran, V.; Lee, S. Supercritical Extraction of Stuart Oil Shale. Fuel Sci. Technol. 1988, 6, 505. (10) Triday, J.; Smith, J. M. Dynamic Behaviour of Supercritical Extraction of Kerogen from Shale. AIChE J. 1988, 34, 658. (11) Bartle, K. D.; Erdem-Senetalar, A.; Kadioglu, E.; Tolay, M. Sub-Critical and Super-Critical Solvent Extraction of the Avgamasya Asphaltite of South Eastern Turkey. In Characterisation of Heavy Crude Oils and Petroleum Residues: Symposium International. Lyon, France, June, 1984, p 217. (12) Deo, M. D.; Hwang, J.; Hanson, F. V. Supercritical Fluid Extraction of a Crude Oil, Bitumen-Derived Liquid and Bitumen by Carbon Dioxide and Propane. Fuel 1992, 71, 1519. (13) Subramanian, M.; Hanson, F. V. Supercritical Fluid Extraction of Bitumens from Utah Oil Sands. Fuel Process. Technol. 1998, 55, 35. (14) Oschmann, H. J.; Prahl, U.; Severin, D. Separation of Paraffin From Crude Oil by Supercritical Fluid Extraction. Pet. Sci. Technol. 1998, 16, 133. (15) Cheng, J.; Fan, Y.; Zhan, Y. Supercritical Propane Fractionation of Wax-Bearing Residue. Sep. Sci. Technol. 1994, 29, 1779. (16) Chung, K. H.; Xu, C.; Hu, Y.; Wang, R. Super-critical Fluid Extraction Cuts Deep into the Bottom of the Barrel. 46th Canadian Chemical Engineering Conference, Kingston, Canada, Sept 1996. (17) Lim, G. B.; Kry, R. P.; Harker, B. C.; Jha, K. N. Cyclic Stimulation of Cold Lake Oil Sand With Supercritical Ethane. International Heavy Oil Symposium, Society of Petroleum Engineers, Calgary, Canada, June 1995. (18) Brennecke, J. F.; Eckert, C. A. Phase Equilibria for Supercritical Fluid Process Design. AIChE J. 1989, 35, 1409. (19) Donaldson, E. C., Chilingarian, G. V., Yen, T. F., Eds. Enhanced Oil Recovery, II: Processes and Operations; Elsevier Science Publishers: New York, 1989. (20) Taylor, L. T. Supercritical Fluid Extraction; John Wiley & Sons: New York, 1996. (21) Rose, J. L. Extraction of Peace River Bitumen Using Supercritical Ethane. Ph.D. Dissertation, University of Calgary, Calgary, Alberta, Canada, 1999. (22) Larson, L. L.; Silva, M. K.; Taylor, M. A.; Orr, F. M., Jr. Temperature Dependence of L1/L2/V Behaviour in CO2/Hydrocarbon Systems. SPE Reservoir Eval. Eng. 1989, February, 105. (23) Butler, R. M.; Mokrys, I. J. Recovery of Heavy Oils Using Vaporized Hydrocarbon Solvents: Further Development of the Vapex Process. J. Can. Pet. Technol. 1993, 32, 6, 56. (24) Schmitt, W. J.; Reid, R. C. The Influence of the Solvent Gas on Solubility and Selectivity in Supercritical Extraction. In Supercritical Fluid Technology; Penninger, J., Radosz, M., McHugh, M., Krukonis, V., Eds.; Elsevier Science Publishers: Amsterdam, The Netherlands, 1985. (25) Johnston, K. P.; Ziger, D. H.; Eckert, C. A. Solubilities of Hydrocarbon Solids in Supercritical Fluids. The Augmented van der Waals Treatment. Ind. Eng. Chem. Fundam. 1982, 21 (3), 191. (26) NIST Thermophysical Properties of Pure Fluids Database, Version 3.1; National Institute of Standards and Technology: Gaithersburg, MD, 1986.

Received for review March 13, 2000 Revised manuscript received June 30, 2000 Accepted July 10, 2000 IE000320R

Fractionation of Peace River Bitumen Using Supercritical Ethane

Recommend Documents