The Bitumen Upgrading of Nigerian Oil Sand by ... - ACS Publications

Jun 5, 2014 - Campus Esbjerg, Niels Bohr vej 8, Esbjerg 6700, Denmark. Ismaila Jimoh. Schlumberger, Kanalholmen 1, Hvidovre Copenhagen 2650, ...
1 downloads 0 Views 2MB Size
Article pubs.acs.org/EF

The Bitumen Upgrading of Nigerian Oil Sand by Supercritical Carbon Dioxide Modified with Alcohols Svetlana Rudyk* Oil and Gas Research Centre, Sultan Qaboos University, Muscat 123, Sultanate of Oman

Pavel Spirov Section of Chemical Engineering, Department of Biotechnology, Chemistry and Environmental Engineering, Aalborg University, Campus Esbjerg, Niels Bohr vej 8, Esbjerg 6700, Denmark

Ismaila Jimoh Schlumberger, Kanalholmen 1, Hvidovre Copenhagen 2650, Denmark

Gholamreza Vakili-Nezhaad Department of Petroleum and Chemical Engineering, Sultan Qaboos University, Muscat 123, Sultanate of Oman ABSTRACT: The experimental study of bitumen extraction from Nigerian oil sand by supercritical carbon dioxide was carried out using a high-pressure extractor. Fifty grams (50 g) of oil sand was placed into a 100-mL extractor at a temperature of 110 °C and pressures of 50, 60, or 65 MPa for the interaction with pure carbon dioxide or modified by the addition of 3 g of ethanol or 2-propanol in three runs to investigate the co-solvent effect on recovery. The liquid recovery of bitumen, using pure carbon dioxide, was determined to be equal to 16.3%, on average, after three runs over the entire pressure range, with a maximum of 18.5% at 50 MPa. The addition of ethanol improved liquid recovery, while the addition of 2-propanol worsened it by ∼5% (on average). The maximum liquid recovery of 24.5% after three runs, using ethanol-modified carbon dioxide, was observed at 60 MPa. The average outgassing losses obtained with the addition of co-solvents increased in the following order: pure carbon dioxide (16.1%), ethanol (19.5%), 2-propanol (24.93%). The obtained bitumen fractions were upgraded by rejection of asphaltenes and petcoke. Chromatographic analysis (gas chromatography−mass spectroscopy, coupled with total ion chromatography (GC-MS TIC)) has shown that the composition of the collected fractions after the first runs is lighter than that observed after the subsequent runs. The fuel-related physicochemical properties of the extracted bitumen fractions were calculated using Khan’s correlations based on the refractive indices. The obtained characteristics matched those observed in the published data of maltene of Athabasca bitumen. chemical properties similar to conventional crude oils.3,4 Similar to that observed with other fossil fuels, marine animals and vegetation deposited with sediments in the coastal waters in prehistoric times are major sources for all of the elements in oil sands.5 If the vast amount of Nigerian oil sand is exploited, it can serve as a feedstock for Nigerian refineries, because of the presence of aliphatic components, probably alkanes.6 However, the huge deposits of oil sand found in southwestern Nigeria remain untapped, because of concerns about the environmental impact. One of the efficient methods currently in use is surface mining, where excavated oil sand is washed with hot water and sodium hydroxide (NaOH) added to improve bitumen separation from the sands.7 This process uses large amounts of water and requires high energy consumption. Despite this

1.. INTRODUCTION The need for alternative sources of energy has become even more acute, in light of the widespread recognition of the dwindling conventional world oil reserves. The interest in exploring other avenues of complimenting and/or eventually replacing this resource is growing rapidly. Tar, oil, or bituminous sands, which are abundant and vastly unexplored, represent a ready alternative to conventional crude oil. They are mixtures of sand, clay, and bitumens that do not flow easily and, hence, are difficult to produce. The largest world oil sands deposits are in Canada, Venezuela, Madagascar, the United States, and Russia.1 In the future, when the crude oil prices rise, the interest in the development of oil sands will increase. Extensive oil sand deposits have been known to occur in southwestern Nigeria since the early 1900s. The deposits cover a 5−8 km belt stretching over a region 120 km in length.2 The oil sand is located on or near the surface and, thus, is suitable for surface mining, as performed in Alberta, Canada. Bitumen is very sticky, with high viscosity and density (1.0 kg/m3), having © 2014 American Chemical Society

Received: March 3, 2014 Revised: May 29, 2014 Published: June 5, 2014 4714

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

extract. In the second case, a co-solvent should be soluble with SC-CO 2 to enhance its density, viscosity, or specific intermolecular interactions. The other physical properties, such as molecular weight and melting and boiling points, as well as structural properties such as molecular shape and angularity of the compound lead to large variations in the solubility under specific process conditions.23,24 Besides, the hydrocarbon mixtures are complex compounds that are composed of various hydrocarbon groups, which are differently solubilized in the presence of co-solvents. Alcohols have been widely tested for the extraction of various substances by SC-CO2.25−29 The recovery increases as the density of the solvent increases, which is explained by the density effect or specific intermolecular interactions. The addition of co-solvents with higher molecular weight was recognized as a way to increase the overall density of the solvent.30 On the other hand, numerous studies31−33 showed that higher extraction and higher solubility of the solutes in SCCO2 have been achieved using ethanol, because of the proven ability of ethanol to participate in hydrogen bonding as a prominent hydrogen bonding acceptor. Some co-solvents, including 2-propanol/isopropanol, are known for their strong hydrogen donating properties that promote the hydrogenation of hydrocarbons, which can be an effective method to upgrade bitumen. This study deals with the extraction of bitumen of the Nigerian oil sand by pure carbon dioxide and modified with the addition of ethanol or 2-propanol in order to compare the effect of these two alcohols on the recovery and upgrading of bitumen. The extracted oil fractions are analyzed by chromatographic analysis (gas chromatography, coupled with total ion chromatography (GC-MS TIC)) and refractive indices, which allow the physicochemical characteristics to be calculated using existing correlations.

fact, bitumen production is economically feasible and applied to a vast area of tar sands in Canada. The problem with this method is that the large amounts of contaminated water are discharged directly into tailing ponds on the surface. Other consequences of such methods of processing tar sand, ranging from water pollution to the emission of greenhouse gases, especially in Canada, and the lack of water in many other countries, bring into sharp focus the urgent need for an alternative means of extracting oil from oil sand. A more effective and less environmentally damaging procedure could be the breakthrough needed to open a new chapter in the exploitation of tar sands. As a result, several heat/solvent hybrid processes have been introduced, which include the solvent and steam; superheated solvents; steam and CO2; H2S and CO2; methane;8,9 carbon dioxide, methane, and nitrogen;10 butane and propane;11,12 and supercritical water, nitrogen, and toluene.13 The existing methods oriented on deeper reservoirs include injecting steam (300−340 °C) and solvents (particularly cyclic steam stimulation (CSS)), in situ combustion (ISC), and steam-assisted gravity drainage (SAGD) in order to reduce bitumen viscosity, but all have certain disadvantages.14 With CSS, the maximum recovery rarely exceeds 20%. SAGD requires a large amount of heat, which can make it inefficient and not economical in low-porosity, highly fractured, and many other complex reservoirs.15 In the Vapex process,14 a mixture of propane and/or butane and a commercially available noncondensable gas (methane, natural gas) are injected into the reservoir to reduce the oil viscosity. In 2003, Talbi et al.15 showed that, because the solubility of CO2 in heavy oils is higher than that of methane, the addition of CO2 makes the Vapex process more cost-effective and environmentally friendly than the conventional methods. The fractionation of bitumen by pure and modified supercritical carbon dioxide was characterized experimentally and numerically in the works described in refs 16 and 17. In 1978, Williams et al. patented the extraction of hydrocarbons from oil shales and tar sands by using various solvents, including supercritical carbon dioxide, at a pressure of 10 MPa and temperatures up to 550 °C.18 Supercritical fluid extraction has been attracting a great deal of interest, because this technique can considerably reduce sample preparation time and can provide bitumen recovery from solid and semisolid samples that is equal to or better than that of the classical extraction techniques.19 In addition, supercritical fluids possess gaslike mass-transfer properties and the solvation characteristics of liquids. Their high diffusivity allows them to penetrate solid materials, and their liquidlike densities enable them to dissolve bitumen from a solid matrix.20 Supercritical fluids are compressible, and small pressure changes produce significant changes in their density and in their ability to solubilize compounds. Also, supercritical fluids have almost no surface tension;21 thus, they can penetrate lowporosity materials while their very low viscosity provides favorable flow characteristics. These properties enable supercritical fluids to provide excellent extraction efficiency and speed, which can be increased by adding small amounts of cosolvents.22 The addition of liquid co-solvents to supercritical carbon dioxide (SC-CO2) are chosen for their ability to interact efficiently with either the solute or the solvent. In the first case, a co-solvent is selected for its ability to dilute the extract, diminish its viscosity, and thereby enhance the flow of the

2. MATERIALS AND EXPERIMENTAL PROCEDURE 2.1. Materials. The sample of oil sand was taken from the Ofoso oil sand field located in the Nigerian oil sand belt, which lies on the onshore areas of the Eastern Dahomey (Benin) Basin. The obtained oil sand was in a solid compact form, with a slight bitumen odor. Carbon dioxide (99.9% pure) was supplied by Strandmollen A/S, Denmark. Pure ethanol was purchased from VWR Prolab. 2-Propanol (99.9% purity) was purchased from AppliChem, BioChemica GmbH, Germany. Tetrahydrofuran (THF) was obtained from BDH Company. 2.2. Solvent Extraction. Solvent extraction was carried out with tetrahydrofuran (THF) to determine the content of bitumen in the investigated oil sand sample. The mass of the oil sand sample used was 10 g. The oil sand was mixed with THF to form a solution with oil sand/THF (1 g of oil sand/6 mL THF), and then the mixture was stirred overnight and filtered. Apparently clean sand/clay was obtained by drying the residue on the filter paper, and the amount collected was weighed. The filtrate sample (the THF-soluble fraction) after extraction was evaporated using a rotary evaporator at 80 °C to remove the THF. After evaporation, the remaining semisolid was dried in the oven at 70 °C overnight until a constant weight and measurements were taken to obtain the bitumen weight fraction. 2.3. X-ray Diffraction Analysis. The oil sand sample was manually ground by an agate mortar and pestle for ∼10 min to achieve a fine particle size for X-ray diffraction (XRD) analysis. In this study, the XRD patterns of the sample powders was measured using a PANalytical X’Pert PRO X-ray diffractometer with Cu Kα radiation generated at 40 kV and 40 mA. To identify the phases present in the sample, the XRD pattern of the sample was compared with a calculated pattern in a database from the International Centre for Diffraction Data (ICDD-PDF-2). 4715

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

2.4. Elemental Analysis (EDS) by Scanning Electron Microscopy. To perform EDS analysis, the sample was dissolved in hexane to remove the oil, which will cause charging in the sample. The dark-colored grains were selected for EDS analysis. 2.5. Experimental Procedure. The sample of oil sand was weighed, put into a closed container, and placed into a Electro Helios oven for preliminary heating to 120 °C for a period of 2 h. The sample was measured after preheating in order to determine evaporation losses. The weight of the sample before and after preheating was 100 and 99.73 g, respectively, implying that the calculated sample loss was equal to 0.27%. The melted sample was mashed by a wooden stick and placed in the extraction cell to begin the experiment. A commercial high-pressure extractor (Model Spe-ed SFE) was used to carry out the extraction experiments. The layout of the process can be found in ref 34. The oven was heated to a temperature of 110 °C. The extraction cell containing the melted oil sand sample was placed into the oven. To start the experiment, the inlet valve of the system was opened, and CO2 was fed into the system from the storage tank via the pump for 1−2 min to attain the required pressure. The gas was injected into the extraction cell through the orifice in the bottom cap. The sample of oil sand and CO2 interact for 30 min in static mode, followed by the collection of extracts in dynamic mode; the inlet valve, outlet, and vent valves were opened, and additional gas was supplied to transport the extract along a steel pipe 0.3 mm in diameter into the collecting test tubes. The collecting test tubes were plugged with cotton in order to trap the hydrocarbons in gas phase. The bitumen was collected in a test tube during 10 min of collection time. The empty test tubes were weighed before and after the collection. Afterward, the inlet valve was closed and the outlet valves opened to release CO2. The procedure was repeated in three consecutive runs, implying that, following the collection of extracts after the first run, the gas was injected again at the same pressure, followed by the extract collection. After the third run, the extraction cell was demounted from the system, carefully opened, and the remaining oil sand was removed from the extraction cell. The oil sand remains and the test tubes with collected bitumen were weighed by using a balance. The first experiment was conducted with pure carbon dioxide. The other tests in the same pressure range were performed using SC-CO2 modified with the addition of ethanol or propanol as co-solvents. To bring the co-solvent into the system of oil and carbon dioxide, 3 g of propanol or ethanol was deposited onto a cotton ball and placed at the bottom of the reactor in the entry of CO2 flow. While CO2 was injected in the reactor, the co-solvent vaporized and interacted with the crude oil. 2.6. GC-MS TIC Chromatographic Analysis. The samples of crude bitumen and extracted by modified SC-CO2 with added ethanol or propanol were analyzed via GC-MS. For the analysis, 0.5 mL of bitumen was diluted with 1.5 mL of n-pentane. The system consists of a gas chromatograph (Model GC-CP-3800) coupled with a mass spectrometer (Model MS-ION TRAP 2000). Separations were performed on a Factor Four VF-5 ms capillary column (20 × 0.15 × 0.39) from Varian (Middelburg, The Netherlands). The initial temperature in the GC oven was 50 °C for 2 min, followed by increasing the temperature up to 300 °C. Helium was used as a carrier gas at a pressure of 20 psi. The MS was operated in electron ionization mode. The temperature was 300 °C. The mass range was from m/z 50 to m/z 600 in scan mode. The GC-MS data were collected using selected ion recording. The total run time was 45 min. The single number carbon (SCN) groups for C10−C30 were identified by comparison with mass spectra reported in the National Institute of Standards and Technology (NIST) library. 2.7. Refractive Index Measurements. To measure the refractive index (RI), a Mettler Toledo Liquiphysics Excellence refractometer (Model RM40) was used. It is a digital refractometer that measures the RI immediately and accurately. One or two drops of the sample are added to the cell, where a high-resolution optical sensor measures the total reflection of light from the light-emitting diode (LED) after it strikes the sample and the refractometer reports the results with an accuracy up to five decimal places. Each measurement required 2 s.

The cell was rinsed with distilled water and acetone and then dried before and after each sample measurement. 2.8. Numerical Calculations of Experimental Results. The results of extraction by pure or modified SC-CO2 are represented by the total recovery, liquid recovery, and outgassing losses. The liquid recovery of bitumen (R) after each run represents the amount of collected bitumen and is calculated as

R (%) =

Wc × 100 Wi

(1)

where Wc is the weight of bitumen collected in the test tubes (calculated as the difference of the weights of the test tubes after and before the collection), and Wi is the amount of bitumen contained in the piece of sand (estimated from the THF solvent extraction). The total recovery of bitumen is calculated as a difference between the weight of the initial tar sand sample and the weight of the tar sand sample removed after the experiment. Outgassing losses are calculated by subtracting the liquid recovery from the total recovery. The distribution of the extracted bitumen between phases (liquid and gas) was calculated as K ‐value =

y x

(2)

where y represents outgassing losses and x is the liquid recovery.

3. RESULTS AND DISCUSSIONS 3.1. Solvent Extraction. The extraction of bitumen from the oil sand sample by tetrahydrofuran (THF) was carried out in order to determine the content of bitumen in the sample. For the most well-known deposits, the bitumen content of oil sand ranges between 7% and 20%: ∼12%−17% in Canada, ∼11% in Trinidad.1 The fraction of the solid part after drying was ∼88.3 wt % and the bitumen part was 11.3 wt %. Approximately 0.4 wt % was lost, probably because of the volatile components. These measurements are consistent with the results of early studies cited by Fassasi,35 showing that the Nigerian oil sand is composed of 12% bitumen, 84% sand, 2% clay, and 4% water. The 12% bitumen content was taken as the typical value for the calculation of the bitumen amount contained in each investigated sample for the calculation of the recovery. For a sample with a mass of 50 g, the initial bitumen content (Wi) was calculated to be equal to 6 g. The oil sand washed with hexane was analyzed via electron microscopy. Most of the grains had a light color, with a few darker grains. According to the XRD analysis, the light grains were identified as quartz (according to reference code 01-070-8054). The darker grains were also similarly analyzed. The energydispersive spectroscopy (EDS) analysis has identified the presence of elemental Si, Al, Fe, S, and a trace of K, which indicates feldspar. 3.2. Upgrading of Bitumen. The examples of the bitumen sample extracted by pure and modified carbon dioxide is shown in Figure 1. The visual observations of all bitumen extracts revealed that they differed in color and viscosity: the bitumen color was honeylike when extracted by pure CO2 at pressures of

Figure 1. Photographs of some bitumen samples extracted by pure and modified supercritical carbon dioxide (SC-CO2). 4716

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

50 and 60 MPa, while it was black at a pressure of 65 MPa. All the samples that were extracted with the co-solvents were transparent orange. This evidence implies that fractions are upgraded when undergoing an extraction process via rejection of asphaltenes and petcokes, giving the black coloration to the bitumen. Al-Sabawi et al. upgraded bitumen via the application of various solvent mixtures at a pressure of 7.5 MPa and temperature of 200 °C to remove metals.36 The upgrading of bitumen extracted by toluene (37.5 mL) from the oil sand sample (12 g), assisted by the gas injection of CO2 (10 MPa) and H2 (6.2 MPa) at 100 °C in the presence of catalysts, is also described in ref 37. The removal of asphaltenes and petcokes before refining is desirable. The metals normally concentrated in asphaltenes cause problems in catalytic refinery processes, and they are detrimental in hydrocracking and hydrotreating. The petcokes are collected as residues after bitumen refining and are burned as a coal substitute in power plants, releasing up to 10% more CO2 into the atmosphere than coal. 3.3. Carbon Dioxide Extraction. The variations in CO2 density enable selective extraction, implying that the fractions of different compositions can be obtained from complex compounds such as bitumen by manipulating the pressure. In 2008, Akinlua et al. extracted hydrocarbons from Niger Delta sedimentary organic rock at a pressure of 34.5 MPa and various temperatures. 38 Our previous investigations of SC-CO 2 extraction of crude oil39,40 and heavy hydrocarbon mixture34 had shown that the maximum extraction recovery could be expected at the pressures of >45 MPa, which creates denser CO2. The experiments were carried out at pressures of 50, 60, and 65 MPa, using pure CO2 and modified with the addition of 3 g of ethanol or propanol to increase the solvating power of CO2.30 Most publications report the results of one extraction run. If the recovery is low, the method can be misinterpreted as inefficient. However, the amount of gas injected at the specific pressure is limited by the dimensions of the extraction cell, implying that the volume of the solvent (CO2) can be insufficient to interact with the entire volume of the solute (bitumen). Our previous experiments on oil extraction from sands showed that the recovery after the second runs can be even higher than after the first runs (unpublished data). To investigate the extraction progression, the experiments were carried out in three runs, to determine whether substantial amounts of additional bitumen could be obtained by multiple extractions. At higher temperatures, the extraction capabilities of carbon dioxide worsens;41 however, the oil sand should be melted for the extraction to occur, which, according to our observations,42 requires temperatures that are >100 °C. The first experiment was carried out at a temperature of 80 °C and a pressure of 65 MPa, using 50 g of melted oil sand preheated at 120 °C. At first run, the extraction was 7%, and the extraction after the second was 3%, totalling 10%. When the extraction cell was opened and the residues examined, it was found that the oil sand cooled and began to solidify, which explains the low recovery after the second run. To avoid bitumen solidification, the extraction operational temperature was set at 110 °C for all of the experiments. The results of bitumen extraction by pure CO2 at pressures of 50, 60 and 65 MPa and a temperature of 110 °C are shown in Figure 2. Because of the shortage of oil sand, the experiment was only duplicated at 60 MPa to verify the consistency of the extraction results (samples 1 and 2). Some other results can be

Figure 2. Liquid recovery of bitumen from the Nigerian oil sand sample by pure SC-CO2 at pressures of 50, 60 (in duplicate), and 65 MPa.

found in ref 33. The bitumen liquid recovery from two samples at this pressure differed by 1.3%, on average, after two runs and by 2.5% after the third run. After the first runs, the highest recovery of 9% was recorded at 50 MPa, which decreased after the pressure increased. The recovery after the second runs of 6.3%, on average, were similar at all pressures, with the exception of sample 2, which declined by 1%. Similarly, the average recovery of 4% after the third run was very close for the three samples, with the exception of sample 1, which had a recovery that was 2.5% higher. The recovery sharply decreased as the run number increased only at a pressure of 50 MPa. At pressures of 60 and 65 MPa, there was no strong dependency of the recovery on the number of runs. Regardless of the general decreasing trend of recovery from the first run to the third run, the conclusion can be made that the bitumen recovery of these four samples by pure CO2 is more dependent on the specifics of the porous matrix of the individual samples and the amount of bitumen extracted after the previous runs. The recovery negatively correlated with pressure only after the first run and showed no big difference in recovery after the second and third runs at various pressures, which indicates that additional bitumen could be extracted if one were to flash the samples with the dense CO2 more times. The highest liquid recovery of 12% with ethanol addition after first runs was achieved at a pressure of 60 MPa, followed by a sharp, almostlinear decrease after two subsequent runs, as can be seen in Figure 3. At pressures of 50 and 65 MPa, the liquid recoveries from the first runs to the third runs decreased gradually and were close in value. The liquid recovery with propanol addition after all three runs at a pressure of 60 and 65 MPa are very similar, with the maximum recovery being observed after the second run, as shown in Figure 4. The graph of liquid recovery with propanol at 50 MPa has a completely different shape, which could be due to the individual specifics of the sample. However, it again indicates that if a low amount of bitumen was recovered at previous runs, greater extraction can be expected after subsequent runs. The summaries of liquid recoveries resulting from all three runs are shown in Figure 5. The liquid recovery by pure CO2 at the pressures of 50 and 65 MPa was 18.5% and 15%, respectively, whereas for samples 1 and 2, it differed at 60 MPa, the recovery being 21% and 15%, respectively. When the co4717

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

worsened it by 5.4% (on average). The average liquid recovery of four samples at all pressures of all three runs was calculated to be 17.4% by pure CO2, 21.4% with ethanol addition, and 10.9% with propanol. The highest liquid recovery with ethanol and propanol was obtained at 60 MPa, although this was not a significant liquid recovery increment, compared to other pressures. The stable increase of liquid recovery with ethanol addition and decrease with propanol addition across the entire pressure range, compared to the pure SC-CO2 extraction, has shown that the co-solvents are not removed by SC-CO2 after the first runs. The similar observations were made in our experiments with the addition of water to oil sand,43 where the certain volumes of water were extracted after each run, exhibiting a minimal water recovery after the first runs. This demonstrated that the co-solvents were dissolved in the bitumen and their influence on the recovery could be traced in the next runs. Although the extraction from different samples can vary due to some difference in the specifics of the porous matrix, the liquid recovery improved in the following order: propanol, pure CO2, ethanol. Compared to the results of our investigations of the crude oil at operational temperatures lower than the boiling points of the co-solvents, the order of oil recovery increased in another sequence: pure CO2, propanol, ethanol.44 The unrecovered residues of bitumen and the total recovery including liquid recovery and outgassing losses are shown in Figure 6. Compared to the negligible outgassing losses of

Figure 3. Liquid recovery of bitumen from Nigerian oil sand sample by SC-CO2 with ethanol added.

Figure 4. Liquid recovery of bitumen from Nigerian oil sand sample by SC-CO2 with propanol added.

Figure 6. Liquid recovery and outgassing losses comprising the total recovery and unrecovered residues that remained after the extraction.

0.27%, which have occurred when the samples of bitumen were heated in the oven for 2 h at 120 °C, they are significant after the SC-CO2 extraction, resulting in average values of 16.1% by pure CO2, 19.5% with the addition of ethanol, and 24.93% with the addition of propanol. This demonstrates the strong capacity of SC-CO2 to vaporize bitumen components. Although the recovery of crude oil usually increases as the pressure increases, because of the increasing density of SCCO2, the total and liquid recovery of bitumen obtained by pure SC-CO2 substantially decrease with the rising pressure, which can indicate that the diffusion of too-dense CO2 into the

Figure 5. Summary of liquid recovery of bitumen obtained by pure and modified SC-CO2 after all three runs.

solvents were added, the highest liquid recoveries were achieved after the second runs. The addition of ethanol increased liquid recovery by 5.15%, while addition of propanol 4718

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

bitumen worsens. While the total recovery obtained with the addition of ethanol and propanol does not differ substantially, the liquid recovery obtained with the propanol is significantly lower and the outgassing losses are significantly greater than those obtained with the ethanol addition at all pressures. This shows that the distribution of the extracted bitumen between phases varies, depending on the type of the co-solvent used. It could be presumed that the vaporization is suppressed at high pressures, leading to the reduction of outgassing losses, as it can be observed using the example of the bitumen extraction by pure SC-CO2, where the outgassing losses substantially decreased at rising pressure. However, this is not always the case,39,43 because the presence of co-solvent interacting with solute, or solvent (or both) can change the phase behavior of the system. A good co-solvent displays high capacity, characterized by solubility and selectivity, determined by the ease of extraction.44 Two main factors are important for the solubility of compounds in SC-CO2: the solvent strength of SC-CO2, which is a function of density, and the volatility of compounds, which is a function of temperature.24 The volatility depends on the operational temperature and boiling temperatures (BP) of solute and co-solvent. At lower pressure, the volatility effect prevails while the density effect dominates at higher pressures. At lower pressures, the total recovery worsens if a co-solvent with a high boiling point is added but the vaporization increases with the addition of the co-solvents that have a low boiling temperature. Because propanol has a higher boiling point than ethanol, its greatest contribution in a gasification reaction is not expected. However, a similar result was also observed when investigating SC-CO2 extraction of crude oil.43 In our previous experiments on SC-CO2 extraction of crude oil at 60 °C with the addition of ethanol or distilled water,39 the outgassing losses increased with pressure, from 30 MPa to 60 MPa, compared to the decreasing tendency when acetone, propanol, or methanol43 were used as co-solvents. This indicates that the vaporization depends not only on the boiling point but also on the type of co-solvent. The interaction between pure or modified carbon dioxide and bitumen occurs via the transfer of components from the bitumen into the gas and from the gas to the bitumen.45,46 The hydrocarbons vaporize into the gas and the gas condenses and solubilizes bitumen. Bitumen-in-CO2 solubilities and bitumen partitioning coefficients, which allow for quantifying co-solvent selectivity and capacity in fractionating bitumen with supercritical CO2 at pressures below 16 MPa, were characterized using refs 16 and 17. Generally, selectivity toward phases increases with pressure and decreases with temperature,44 because of the conflicting effect of these parameters exhibiting a cross-over effect at higher pressure.47 Although the detailed composition of the gases was not measured due to the difficulties of the collecting of vapor phase, the distribution of extracted bitumen between phases in Figure 7 demonstrates that the graphs of K-values of pure SC-CO2 and ethanolmodified extractions follow each other. The graph of ethanol increased, reaching near equal distribution between liquid and vapor phases at pressures of 60−65 MPa. The gasification was much higher in the presence of propanol. Similarly, in the experiment with the addition of distilled water, the vaporization was 8% at a pressure of 30 MPa, then it abruptly increased at a pressure of 40 MPa, achieving 20%−25% at pressures up to 60 MPa.39 Thus, the increased vaporization due to the presence of some co-solvents cannot be explained by the closeness of the

Figure 7. Distribution of the extracted bitumen between phases.

boiling point to the operational temperature and, instead, should be attributed to the selectivity. Thus, the selectivity of the co-solventthat is, how the co-solvent interacts with the SC-CO2 and with the bitumenis perhaps of even greater significance than solvent capacity.44 The increase of propanol in the gas phase with pressure was also observed in refs 45 and 47, which could have occurred due to the transfer hydrogenation of bitumen from propanol. The hydrogenation of heavy oil and asphaltene at high temperatures, by supercritical water or in the presence of catalysts and solvents, is important for the bitumen upgrading and suppressing of coke formation.13,48 Greater bitumen recovery and lighter hydrocarbon composition than that achieved by pure carbon dioxide extraction were obtained by the similar extraction experiments with the water addition42 and may confirm the conclusion regarding the role of water. Morimoto et al.13 found that the effect of supercritical water and supercritical gas, such as nitrogen, on bitumen upgrading has a physical nature while the effect of a supercritical solvent such as toluene has a chemical nature. Whether the effect of alcohols added to SC-CO2 has a physical or chemical nature in the promotion of the bitumen gasification and the selectivity of alcohols between phases requires further study. Various methods and techniques of characterization of bitumen fractions obtained by different methods are available in the literature.13,47 In this study, the extracted fractions were analyzed by GC-MS chromatography and by RI measurements, allowing one to calculate different parameters using existing correlations.49−52 3.4. GC-MS TIC Chromatography Analysis. The chromatograms of the hydrocarbon mixtures of various originations have unique patterns that allows them to be used as fingerprints.53 The stages of maturity or biodegradation can be traced by observing the chromatograms.54 The fractions undergoing SC-CO2 extraction at various pressures have also unique fingerprints.40,42 The fingerprint of the crude bitumen extracted with tetrahydrofuran (THF) has an almost-triangular shape, linearly increasing from C13 to the domelike top in the range of 26−30 min matching to C26−C30 SCN groups, as shown in Figure 8. The peaks are not eluted, except for the two peaks at 28 and 30 min of retention time. The peak of C28 triterpane was identified at m/z = 177 as a biomarker C28H48, 28 Bisnor 17β (H)-hopane (28-bisnorhopane)55,56 by comparison with the spectra in the NIST library, as described in ref 42. The identification of the second peak requires further studies. Pristane or phytane peaks were not detected, because of the 4719

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

However, the uplifting of the baseline in the region of highboiling compounds signifies the presence of the heavier fractions in the samples. Such uplifting of the baselines with increasing pressure can be seen in the chromatograms with the propanol starting from 30 min of retention time. In addition, the shift of the chromatograms obtained after the second runs to the region of higher-boiling fractions can also be observed at pressures of 60 and 65 MPa, which implies that the fractions extracted after the first runs with propanol are lighter; the baselines of the chromatograms obtained at a pressure of 50 MPa after two runs almost coincide in the high-boiling-point region. The bitumen fractions extracted after the second runs with propanol are darker than those extracted after the first runs. Regardless the number of runs, the shape of the chromatograms extracted with ethanol addition is foldlike. The shift to the region of lower-boiling-point fractions can also be observed after the second runs, compared to the first runs at pressures of 50 and 60 MPa, while the chromatograms obtained at a pressure of 65 MPa after both runs coincide. The baselines in the regions of higher-boiling-point fractions almost coincide after both runs at all pressures, implying that the pressure effect is not pronounced with ethanol addition. The chromatogram of the first run at a pressure of 60 MPa covers the larger interval, matching to the lower-boiling-point compounds, compared to the second run, which explains the maximum recovery obtained in this case in the entire SC-CO2 extraction experiment. By visual examination, all the samples extracted with ethanol looked similar. Overlay of the fingerprints extracted by pure and modified SC-CO2 is shown at a pressure of 60 MPa after the first runs in Figure 10, as an example. The difference between fingerprints at other pressures is not that significant. The pressure effect of pure carbon dioxide on the extraction of heavier compounds is

Figure 8. Chromatogram of the crude bitumen sample.

severe biodegradation of the bitumen. To compare the fingerprints of the extracts after various runs at the same pressure, the normalization of the k-axis scale was fulfilled, based on the equality of the heights of the biomarkers. The overlay of the chromatograms obtained after the first and second runs with the addition of propanol or ethanol are shown in Figure 9; the fingerprints of the samples extracted by pure carbon dioxide are described in ref 42. The shape of the area of the unresolved complex mixture (UCM) hump mainly consisting of naphtenes (cycloalkanes) with propanol alternating from a fold-like appearance after the first runs to a triangular-like appearance after the second runs at all pressures. The high-boiling hydrocarbon fractions cannot be fully characterized by chromatographic analysis, because their boiling temperatures are much higher than the operating temperature.

Figure 9. Chromatogram of extracted bitumen at a pressure of 60 MPa with pure CO2, as well as with added ethanol and propanol. 4720

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

Figure 10. Chromatograms of bitumen samples extracted by carbon dioxide with the addition of ethanol or propanol after the first and second runs at various pressures.

Table 1. Khan’s Correlations Used in This Work for Characterization51 property H content, [H] (%) H/C ratio C(aro) H(aro) molecular weight, MW (g/mol) density, d (g/cm3) Conradson carbon residue, CCR (%)

definition

equation

[H] = 57.264 − 30.50n hydrogen-to-carbon ratio of liquid oil (atomic) H/C = 6.876 − 3.50n carbon aromaticity (aromatic carbon mass fraction in the liquid oil) C(aro) = 3.657n − 5.228 proton aromaticity (aromatic elemental hydrogen mass fraction in liquid oil) H(aro) = 2.102n − 3.103 mass of one mole liquid oil MW = 4383.2 − 2655n mass of unit volume of liquid oil d = 1.977n − 2.08 A destructive-distillation method for estimation of carbon residues in fuels and CCR = 37.75n − lubricating oils. 55.29 percent of hydrogen in the liquid oil (%)

evident from the observation that the UCM area in a higherboiling-point region becomes higher and broader as the pressure increases. To a lesser extent, it can be traced in the chromatograms of the samples extracted with the propanol addition. Thus, the kinetics of extraction process in the

observed level of significance 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0097

presence of ethanol is less sensitive to the pressure: the shape of the fingerprints and the involvement of the heavier fractions are not dependent on the number of runs. The visual inspection of the samples in Figure 1 confirms the conclusion of GC-MS: the darkness of the samples 4721

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

Table 2. Characterization of Bitumen Fractions Using Khan’s Correlations,51 Based on Their Refractive Indices Mass Fraction n

P (MPa)

co-solvent

RI

H (%)

H/C (wt/wt)

1 2 3 4 5 6 7 8 9

50 60 65 50 60 65 50 60 65

propanol propanol propanol ethanol ethanol ethanol pure CO2 pure CO2 pure CO2

1.5228 1.5227 1.5246 1.5169 1.5218 1.5174 1.5225 1.5217 1.5196

10.82 10.82 10.76 10.99 10.85 10.98 10.83 10.85 10.91

1.55 1.55 1.54 1.57 1.55 1.56 1.55 1.55 1.56

C(aro)

H(aro)

molecular weight, MW (g/mol)

density, d (g/cm3)

Conradson carbon residue, CCR (%)

0.341 0.340 0.347 0.319 0.337 0.321 0.339 0.337 0.329

0.0979 0.0977 0.1017 0.0855 0.0958 0.0865 0.0973 0.0956 0.0912

340.2 340.4 335.4 355.8 342.8 354.5 340.9 343.1 348.7

0.9306 0.9304 0.9341 0.9189 0.9286 0.9199 0.9299 0.9284 0.9242

2.26 2.19 2.26 1.97 2.16 1.99 2.18 2.15 2.07

hydrocarbon composition and better quality of the collected liquid fraction of bitumen was obtained with the addition of ethanol. The heaviest composition of the collected bitumen liquid fractions were obtained with the addition of propanol. The visual observations also confirm the above-mentioned conclusion, because, regardless the black color of the bitumen sample obtained by pure SC-CO2 at a pressure of 65 MPa, its viscosity is lower than of the sample extracted with the propanol at the same pressure.

corresponding to the heavier fractions decreases in the following order: pure SC-CO2, SC-CO2 + propanol, SC-CO2 + ethanol. 3.5. Characterization of the Bitumen Extracts Using Refractive Index Measurements. Characterization techniques based upon the liquid’s refractive index (RI) are used with petroleum distillates to predict fuel-related properties. Correlations based on the data obtained from relatively simple characterization techniques for predicting fuel-related physicochemical properties of such complex mixtures would facilitate the utilization of these liquids in various processes. Properties such as molecular weight and aromaticity of liquids influence their utilization behavior/properties. Measurements of molecular weight or aromaticity can be difficult, timeconsuming, and expensive; can require skilled operators; and are often beyond the resources of most small laboratories. The correlations between properties of petroleum liquids and their refractive indices are available in the literature.49−51 Khan52 has developed several empirical equations to correlate the RI values of the petroleum liquids with the liquid hydrogen content ([H], in weight percent) and the H/C ratio, aromaticity (carbon and proton), molecular weight (MW), density (d), and Conradson carbon residue (CCR), which are used in the present work for characterization of the bitumen fractions produced by the SCCO2 extraction method. A summary of these correlations with their observed level of significance is given in Table 1. The level of significance, in statistical terms, is defined as the probability of rejecting the null hypothesis (as the true hypothesis). According to the level of significance values given in Table 1, the linear relationship between various properties and RI has been justified. The results of the calculations based on these correlations are shown in Table 2. Comparison of the obtained characteristics with those available in the literature data for the bitumen fractions56 shows that they are very close. The average values of H/C atomic ratio is 1.55, and the hydrogen content is given as [H] = 10.85% of Nigerian bitumen extracts, which matches the reported data of the Athabasca maltene (H/C = 1.55 and [H] = 10.87%),56 which allows identifying the collected extracts as maltene. The average values of [H] and the H/C ratio have the following order: ethanol > pure CO2 > propanol. The large H/C atomic ratio indicates low density, a low content of aromatic hydrocarbons, and high cracking reactivity, resulting in high yields of light hydrocarbons and a low yield of coke. The average aromaticity of C (ar) and H (ar) decreases in the following order: propanol > pure CO2 > ethanol. The greater values of MW, d, and CCR indicate a heavier composition and changes as follows: ethanol > pure CO2 > propanol. The conclusion can be made that the lighter

4. CONCLUSION The described method can be applied to the recovery and upgrading of bitumen in the extractors produced by surface mining, although the ethanol addition also can be tested in a VAPEX process. The achieved bitumen recoveries from Nigerian oil sands after three runs, averaging to 16.3% and 21% by pure and ethanol-modified carbon dioxide, respectively, are encouraging. The addition of propanol leads to the greater recovery of the vapor phase and lower recovery of the liquid phase of bitumen, while the addition of ethanol facilitates extraction of lower-boiling-point fractions, compared to the higher-boiling-point fractions extracted with the propanol addition or by pure carbon dioxide. The upgrading of the bitumen through the rejection of sand, solids, petcokes, and asphaltenes is obtained during the process. Furthermore, the numerical correlations used to calculate various characteristics of bitumen, based on the refractive indices of the collected extracts, provide values very close to those reported in the literature and can be recommended for further use. There is an indication that the recovery can be improved by additional flushing, by preliminary soaking in ethanol, and by choosing more-efficient co-solvents. Compared to the hot water washing method, the supercritical carbon dioxide (SC-CO2) extraction is a more environmentally friendly method, because it does not require much water and does not create tailing ponds. If the gas recycling is applied, the SC-CO2 extraction can become costeffective. All of these imply that the SC-CO2 method deserves further consideration.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The experiments on bitumen supercritical extraction were performed at Aalborg University, Esbjerg Campus, Denmark. 4722

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

(21) Al-Abri, A.; Amin, R. Effect of Hydrocarbon and NonHydrocarbon Gas Injection on the Interfacial Tension of a Gas Condensate System. Chem. Eng. Technol. 2011, 34 (1), 127−133. (22) Dehghan, A. A.; Kharrat, R.; Ghazanfari, M. H. Visualization and Quantification of Asphaltinic-Heavy Oil Displacement by Co-Solvents at Different Wettability Conditions. Pet. Sci. Technol. 2010, 28 (2), 176−189. (23) Anitescu, G.; Tavlarides, L. L. Solubilities of solids in supercritical fluidsI. New quasistatic experimental method for polycyclic aromatic hydrocarbons (PAHs) + pure fluids. J. Supercrit. Fluids 1997, 10 (3), 175−189. (24) Modey, W. K.; Mulholland, D. A.; Raynor, M. W. Analytical Supercritical Fluid Extraction of Natural Products. Phytochem. Anal. 1996, 7 (1), 1−15. (25) Lee, F. M., Distillation: Extractive Distillation. In Handbook of Methods and Instrumentation in Separation Science; Elsevier: London, 2009; Vol. 2, pp 127−136. (26) Lucia, A.; Finger, E. J. Co-solvent selection and recovery. Adv. Environ. Res. 2004, 8 (2), 197−211. (27) Harper, R. G.; Harper, P. M.; Hostrup, M., Extraction: Solvent Based Separation In Handbook of Methods and Instrumentation in Separation Science, Elsevier: London, 2009; Vol. 2, pp 621−622. (28) Huang, Z.; Chiew, Y. C.; Lu, W. D.; Kawi, S. Solubility of aspirin in supercritical carbon dioxide/alcohol mixtures. Fluid Phase q. 2005, 237 (1), 9−15. (29) Yonker, C. R.; Smith, R. D. Solvatochromic behavior of binary supercritical fluids: the carbon dioxide/2-propanol system. J. Phys. Chem. 1988, 92 (8), 2374−2378. (30) Enick, R.; Olsen, D.; Ammer, J.; Schuller, W. Mobility and Conformance Control for CO2 EOR via Thickeners, Foams, and Gels-A Literature Review of 40 Years of Research and Pilot Tests; U.S. Department of Energy: 2012; p 267. (31) Rincón , J.; Camarillo, R.; Rodríguez, L.; Ancillo, V. Fractionation of used frying oil by supercritical CO2 and cosolvents. Ind. Eng. Chem. Res. 2010, 49 (5), 2410−2418. (32) Reiser, S.; McCann, N.; Horsch, M.; Hasse, H. Hydrogen bonding of ethanol in supercritical mixtures with CO2 by 1H NMR spectroscopy and molecular simulation. J. Supercrit. Fluids 2012, 68, 94−103. (33) Saharay, M.; Balasubramanian, S. Electron donor−acceptor interactions in ethanol−CO2 mixtures: An ab initio molecular dynamics study of supercritical carbon dioxide. J. Phys. Chemistry B 2006, 110 (8), 3782−3790. (34) Rudyk, S.; Spirov, P. Three-Dimensional Scheme of Supercritical Carbon Dioxide Extraction of Heavy Hydrocarbon Mixture in (Pressure; Temperature; Recovery) Coordinates. Energy Fuels 2013, 27 (10), 5996−6001. (35) Fasasi, M.; Oyawale, A.; Mokobia, C.; Tchokossa, P.; Ajayi, T.; Balogun, F. Natural radioactivity of the tar-sand deposits of Ondo State, Southwestern Nigeria. Nucl. Instrum. Methods Phys. Res., Sect. A 2003, 505 (1), 449−453. (36) Al-Sabawi, M.; Seth, D.; de Bruijn, T. Effect of modifiers in npentane on the supercritical extraction of Athabasca bitumen. Fuel Process. Technol. 2011, 92 (10), 1929−1938. (37) Brough, S. A.; Riley, S. H.; McGrady, G. S.; Tanhawiriyakul, S.; Romero-Zeron, L.; Willson, C. D. Low temperature extraction and upgrading of oil sands and bitumen in supercritical fluid mixtures. Chem. Commun. 2010, 46 (27), 4923−4925. (38) Akinlua, A.; Torto, N.; Ajayi, T. R. Supercritical fluid extraction of aliphatic hydrocarbons from Niger Delta sedimentary rock. J. Supercrit. Fluids 2008, 45 (1), 57−63. (39) Rudyk, S. Determination of saturation pressures using experimental data of modified SC-CO2 extraction of crude oil by consistency test. J. Supercrit. Fluids 2013, 82, 63−71. (40) Rudyk, S.; Hussain, S.; Spirov, P. Supercritical extraction of crude oil by methanol- and ethanol-modified carbon dioxide. J. Supercrit. Fluids 2013, 78, 63−69. (41) Danesh, A., PVT and Phase Behaviour of Petroleum Reservoir Fluids; Elsevier Science: Amsterdam, 1998; Vol. 47, p 388.

The authors thank Saif Amer Al-Mammari and Maissa Sassi (both from Sultan Qaboos University, Oman) for performing measurements, and Dr. O. A. Ehinola (Energy and Environmental Research Group (EERG), Department of Geology, University of Ibadan, Nigeria) for their assistance in obtaining the oil sand samples.



REFERENCES

(1) Chilingar, G. V.; Yen, T. F., Bitumens, Asphalts, and Tar Sands; Elsevier Science: Amsterdam, 1978; Vol. 7. (2) Enu, E. Textural characteristics of the Nigerian tar sands. Sediment. Geol. 1985, 44 (1), 65−81. (3) Obiajunwa, E.; Nwachukwu, J. Simultaneous PIXE and PIGME analysis of a Nigerian tar sand sample from a deep borehole. J. Radioanal. Nucl. Chem. 2000, 245 (3), 659−661. (4) Omole, O.; Olieh, M. N.; Osinowo, T. Thermal visbreaking of heavy oil from the Nigerian tar sand. Fuel 1999, 78 (12), 1489−1496. (5) Nadkarni, R. A. Modern Instrumental Methods of Elemental Analysis of Petroleum Products and Lubricants; ASTM International: West Conshohocken, PA, 1991; Vol. 1109. (6) Technical Overview: Nigeria’s Bitumen Belt and Development Potential; Ministry of Solid Minerals Development: Nigeria, 2006; p 27. (7) Miller, J.; Misra, M. Hot water process development for Utah tar sands. Fuel Process. Technol. 1982, 6 (1), 27−59. (8) Rezaei, N.; Chatzis, I., Incorporation of Heat in the VAPEX Process: Warm VAPEX. In Canadian International Petroleum Conference, Petroleum Society of Canada: June 12−14, Calgary, Alberta, 2007. (9) Peterson, J. A.; Riva, D.; Edmunds, N. R.; Solanki, S. C., The Application of Solvent−Additive SAGD Processes in Reservoirs with Associated Basal Water. Presented at the Canadian Unconventional Resources and International Petroleum Conference, Oct. 19−21, 2010; Society of Petroleum Engineers: Calgary, Alberta, Canada, 2010. (10) Al-Murayri, M.; Harding, T. G.; Maini, B. B. Solubility Of Methane, Nitrogen, and Carbon Dioxide in Bitumen and Water for SAGD Modelling. J. Can. Pet. Technol. 2011, 50 (7/8), 34−45. (11) Jiang, Q.; Yuan, J.-Y.; Russel-Houston, J.; Thornton, B.; Squires, A. Evaluation of Recovery Technologies for the Grosmont Carbonate Reservoirs. J. Can. Pet. Technol. 2010, 49 (5), 56−64. (12) Pathak, V.; Babadagli, T.; Edmunds, N. R. Experimental Investigation of Bitumen Recovery from Fractured Carbonates Using Hot-Solvents. J. Can. Pet. Technol. 2013, 52 (4), 289−295. (13) Morimoto, M.; Sugimoto, Y.; Saotome, Y.; Sato, S.; Takanohashi, T. Effect of supercritical water on upgrading reaction of oil sand bitumen. J. Supercrit. Fluids 2010, 55 (1), 223−231. (14) Das, S. K. Vapex: An Efficient Process for the Recovery of Heavy Oil and Bitumen. SPE J. 1998, 3 (3), 232−237. (15) Talbi, K.; Maini, B., Evaluation of CO2 based Vapex process for the recovery of bitumen from tar sand reservoirs. Presented at the SPE International Improved Oil Recovery Conference in Asia Pacific, Oct. 20− 21, 2003, Kuala Lumpur, Malaysia. (16) Yu, J. M.; Huang, S. H.; Radosz, M. Phase behavior of reservoir fluids: VI. Cosolvent effects on bitumen fractionation with supercritical CO2. Fluid Phase Equilib. 1994, 93, 353−362. (17) Huang, S. H.; Radosz, M. Phase behavior of reservoir fluids V: SAFT model of CO2 and bitumen systems. Fluid Phase Equilib. 1991, 70 (1), 33−54. (18) Williams, D. F.; Martin, T. G. Extraction of oil shales and tar sands, U.S. Patent 4,108,760, 1978. (19) Pang, T.; McLaughlin, E. Supercritical extraction of aromatic hydrocarbon solids and tar sand bitumens. Ind. Eng. Chem. Process Des. Dev. 1985, 24 (4), 1027−1032. (20) Demirbas, A. Recovery of asphaltenes from tar sand by supercritical fluid extraction. Pet. Sci. Technol. 2000, 18 (7−8), 771− 781. 4723

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724

Energy & Fuels

Article

(42) Rudyk, S.; Spirov, P. Upgrading and extraction of bitumen from Nigerian tar sand by supercritical carbon dioxide. Appl. Energy 2014, 113, 1397−1404. (43) Rudyk, S.; Spirov, P.; Hussain, S. Effect of co-solvents on SCCO2 extraction of crude oil by consistency test. J. Supercrit. Fluids 2014, 91, 15−23. (44) Chiu, B.-S.; Wilkinson, G. T. Ternary phase equilibria of the isopropanol+water+carbon dioxide system at high pressure. Korean J. Chem. Eng. 1999, 16 (2), 187−192. (45) Al-Wahaibi, Y. M.; Al-Hadrami, A. K. The influence of high permeability lenses on immiscible, first- and multi-contact miscible gas injection. J. Pet. Sci. Eng. 2011, 77 (3−4), 313−325. (46) Al-Wahaibi, Y. M. First-Contact-Miscible and MulticontactMiscible Gas Injection within a Channeling Heterogeneity System. Energy Fuels 2010, 24 (3), 1813−1821. (47) Yoon, S.; Bhatt, S.; Lee, W.; Lee, H.; Jeong, S.; Baeg, J.-O.; Lee, C. Separation and characterization of bitumen from Athabasca oil sand. Korean J. Chem. Eng. 2009, 26 (1), 64−71. (48) Sato, T.; Mori, S.; Watanabe, M.; Sasaki, M.; Itoh, N. Upgrading of bitumen with formic acid in supercritical water. J. Supercrit. Fluids 2010, 55 (1), 232−240. (49) Modarress, H.; Vakili-Nezhaad, G. R. A New Characterization Factor for Hydrocarbons and Petroleum Fluids Fractions. Oil Gas Sci. Technol.Rev. IFP 2002, 57 (2), 149−154. (50) Riazi, M. R.; Al-Sahhaf, T. A. Physical Properties of n-Alkanes and n-Alkylhydrocarbons: Application to Petroleum Mixtures. Ind. Eng. Chem. Res. 1995, 34 (11), 4145−4148. (51) White, C. M.; Perry, M. B.; Schmidt, C. E.; Douglas, L. J. Relationship between refractive indices and other properties of coal hydrogenation distillates. Energy Fuels 1987, 1 (1), 99−105. (52) Khan, M. R. Correlation between Refractive Indices and Other FuelRelated Physical/Chemical Properties of Pyrolysis Liquids Derived from Coal; Technical Report, U.S. Department of Energy, Morgantown Energy Technology Center: Morgantown, WV, USA, 2013; pp 223− 332. (53) Clark, H.; Jurs, P. Classification of crude oil gas chromatograms by pattern recognition techniques. Anal. Chem. 1979, 51 (6), 616− 623. (54) Christensen, J. H.; Tomasi, G. Practical aspects of chemometrics for oil spill fingerprinting. J. Chromatogr. A 2007, 1169 (1−2), 1−22. (55) Noble, R.; Alexander, R.; Kagi, R. I. The occurrence of bisnorhopane, trisnorhopane and 25-norhopanes as free hydrocarbons in some Australian shales. Org. Geochem. 1985, 8 (2), 171−176. (56) Brenner, J. Lithologic variation of the rare biomarker compound 28, 30-Bisnorhopane within the miocene monterey formation in the Texaco Anita-14 well, Santa Barbara County, California. AAPG Bull. 1994, 78 (4), 658.

4724

dx.doi.org/10.1021/ef500493n | Energy Fuels 2014, 28, 4714−4724