Batch Solvent Extraction of Bitumen from Oil Sand ... - ACS Publications

Apr 3, 2017 - An extraction cell of solvent-diluted bitumen from oil sand with a data logger was developed in this work. Centrifugation was performed ...
2 downloads 9 Views 2MB Size
Subscriber access provided by TRENT UNIV

Article

BATCH SOLVENT EXTRACTION OF BITUMEN FROM OIL SAND: 2. EXPERIMENTAL DEVELOPMENT AND MODELING Merouane Khammar, and Yuming Xu Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.6b03253 • Publication Date (Web): 03 Apr 2017 Downloaded from http://pubs.acs.org on April 8, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

BATCH SOLVENT EXTRACTION OF BITUMEN FROM

OIL

SAND:

2.

EXPERIMENTAL

DEVELOPMENT AND MODELING Merouane Khammar (*) and Yuming Xu AUTHOR ADDRESS. Natural Resources Canada, CanmetENERGY, One Oil Patch Drive, Devon, Alberta, Canada T9G 1A8 KEYWORDS. Centrifugation, filtration, solvent extraction, bitumen, oil sands, non-aqueous.

ABSTRACT. A new experimental technique is developed for the determination of saturated and unsaturated hydraulic properties of solvent-diluted bitumen flow through oil sands cake. This technique is based on continuous measurement of pressure and temperature during centrifugal extraction of solvent-diluted bitumen. Time evolution of extracted diluted bitumen volume and flow rate are obtained from pressure and temperature measurements. A criterion for the determination of the transition time from filtration to desaturation is applied. Experimental filtration data are fitted with an analytical model to obtain saturated hydraulic conductivity of the oil sand cake, and experimental desaturation data are fitted to obtain unsaturated hydraulic properties. The model is used to predict the evolution of bitumen extraction efficiency at

ACS Paragon Plus Environment

1

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 37

different centrifugal forces. The feasibility of continuous commercial extraction of bitumen from oil sands by centrifugal filtration is discussed.

1. INTRODUCTION Centrifugal filtration of solvent-diluted bitumen from oil sand is a promising technology for non-aqueous bitumen extraction on an industrial scale. Wu et al. [1] conducted batch centrifugal filtration of solvent-diluted bitumen from oil sand. They investigated effects of solvent quality, solvent-to-bitumen ratio, centrifugal force, and number of extraction stages. It was found that high bitumen recovery can be achieved after 10 minutes of centrifugal filtration. Bitumen extraction efficiency depends on extraction time, oil sand cake properties, diluted bitumen properties, and centrifugal force. The problem of maximizing diluted bitumen recovery trapped in mineral particles was studied by Faradonbeh et al.[2]. Gravity drainage and pressure filtration were investigated and a mathematical model was developed to predict the evolution of liquid saturation profiles along the sand pack. Pulati et al. [3] examined extraction of bitumen from oil sands using deep eutectic ionic liquid analogues. In this process, ionic liquid is mixed with solvent and oil sands followed by two separation processes: solid-liquid separation in a decanter centrifuge to separate sand from the liquid phase, and liquid-liquid separation to separate the analogue ionic liquid from solvent diluted bitumen. In the scale-up operation, a decanter centrifuge was used for solid-liquid separation. Although hydrocarbon content of separated sand particles was below the detection limit of infrared spectroscopy, analogue ionic liquid content of the separated sand varied from 5% to 10%. Problems related to clogged lines and excessive torque on centrifuge scroll were reported. Decanting centrifuges rely on

ACS Paragon Plus Environment

2

Page 3 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

centrifugal sedimentation and separation of solid particles from the liquid phase, and unlike filtering centrifuges, they do not desaturate the cake from the liquid phase. Our companion publication [4] presented the theoretical basis of batch centrifugal filtration and a numerical method for prediction of liquid fraction extracted from solvent-diluted oil sand bitumen. For complete bitumen dissolution in solvent, bitumen extraction efficiency corresponds to liquid extraction efficiency from the oil sand cake. A method was developed for determination of the transition time from filtration to desaturation and a fitting procedure was proposed for determination of cake hydraulic properties. These properties were to be used for prediction of liquid extraction efficiency under different centrifugal forces. The fitting procedure requires an experimental technique for continuous measurement of extracted liquid volume during batch centrifugation. Batch centrifugation has been used in petroleum and environmental engineering as an accelerated experimental technique for determination of hydraulic properties of liquid flow through porous samples [5-7]. Capillary pressure curves were obtained from equilibrium between capillary and adsorptive forces and centrifugal forces in unsaturated samples. Steady and transient flow experiments were developed. In the steady flow mode, the inflow and outflow rates are constant and equal. Experimental data were used in combination with model predictions to obtain capillary and hydraulic conductivity functions [8]. In the transient flow mode, time evolution of average saturation or filtrate volume was measured during centrifugation. The solution of the Richards equation in a centrifugal field was used to obtain unsaturated hydraulic parameters by fitting experimental data [5]. Different experimental techniques for online measurements during centrifugation are reported in the literature [5, 9, 10]: filtrate volume

ACS Paragon Plus Environment

3

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 37

measured by a high-speed camera [11], cake water content measured by electrical conductivity sensors [5, 9], and pressure head measured by head tensiometers [8]. Nimmo et al. [9] developed an experimental technique based on conductivity measurements for monitoring online water content in a porous sample during centrifugation. The finite difference method was used to solve the Richards equation in a centrifugal field and model predictions were successfully compared with experimental data. Simunek and Nimmo [5] used the same experimental technique as that presented by Nimmo [9], to measure time variations of soil water contents during a sequence of several rotation speeds. They presented a model based on a massconservative solution of the Richards equation in a centrifugal field. Experimental data were fitted with model predictions to obtain soil hydraulic properties. McCartney et al. [10] developed a custom-built centrifuge for the determination of soil hydraulic properties in steady-state flow mode. The discharge velocity was monitored using a pressure transducer that measured the evolution with time of filtrate volume. Calibration between outflow volume and transducer measurements was obtained from linear trends between liquid volume and transducer output voltage for different rotation speeds. Pressure transducer measurements were used to identify the onset of steady-state by comparing inflow and outflow rates. It was reported that pressure transducer measurements can be affected by air pressure variations inside the cell and the centrifugal force deflection on the pressure sensor. A time domain reflectometry waveguide and an array of three tensiometers embedded in the cell side wall were used to measure the average volumetric water content and the matric suction profile across the thickness of the porous clay sample.

ACS Paragon Plus Environment

4

Page 5 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

The use of electrical conductivity sensors for measurement of oil sand cake average saturation is not possible because, unlike water, bitumen diluted with organic solvent is not electrically conductive. Centrifuges equipped with high-speed cameras for online measurement of filtrate volume during centrifugation are used for the experimental determination of hydraulic properties for flow through core samples [11]. The cost of these centrifuges is prohibitive and most centrifuge cells are not chemically compatible with organic solvents when used for extended periods of time. In this work, a new extraction cell is developed where the filtrate volume is monitored continuously during centrifugation by a pressure-temperature data logger. Pressure-temperature measurements are converted into extracted liquid volume evolution with time. Liquid volume data are used in combination with model predictions to obtain hydraulic properties of solvent– diluted bitumen flow through oil sand cake. These properties are necessary for the prediction of the separation performance of continuous filtering centrifuges. Batch extraction efficiency is predicted for different rotation speeds. Model predictions are compared with thermogravimetric analysis measurements (TGA) performed on cake samples before and after extraction. 2. EXPERIMENTAL An extraction cell of solvent-diluted bitumen from oil sand with data logger was developed in this work. Centrifugation was performed in a commercial Beckman Avanti-31 centrifuge without modifications. A schematic and a photograph of the extraction cell are presented in Figure 1. A PRTemp140 data logger from Madgetech with a measurement rate of 1 Hz was used to measure pressure and temperature continuously during centrifugation. The pressure data logger is connected through a Swagelok fitting and a 1/4-inch stainless steel tube to the filtrate reservoir. It is positioned below the filter paper, facing downward and parallel to the filtrate reservoir, such

ACS Paragon Plus Environment

5

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 37

that the liquid surface in the reservoir is parallel to the surface of the pressure sensor. A stainless steel hose clamp was used to prevent bending of the tube connecting the pressure data logger to the reservoir during centrifugation. Two smaller hose clamps were used to clamp electrical wires on the data logger surface. The electrical wires allow connection to the docking station to enable data downloading without cell disassembly, as illustrated schematically in Figure 2. The stainless steel reservoir was wrapped with electrical tape to prevent a possible electrical short circuit with the data logger. A VWR grade 417 filter paper with retention of particles down to 40 micrometers was placed on top of a perforated stainless steel plate in the cell. A Viton o-ring and an aluminum ring were used to seal any micro-gaps between the filter paper edges and the cell wall. Pressure measured by the data logger reflects the centrifugal pressure exerted by the liquid layer height between the transducer sensor surface and the liquid surface. The centrifugal force also pushes the pressure sensor outward causing measurement deflection that depends on rotation speed. The conversion of pressure data to reservoir liquid volume data is described in the following section.

ACS Paragon Plus Environment

6

Page 7 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Sealing screw

Filter paper

Perforated plate

Pressure and temperature data logger

Cell base

Lid

Cylindrical cell

Filtrate reservoir

Electrical wires for data logger

(a)

(b) Figure 1 – (a) Photograph and (b) schematic view of the new extraction cell;    is the water height in the reservoir and       .

ACS Paragon Plus Environment

7

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 37

Extraction cell

Electrical wires

Computer

Docking station

Figure 2 – Experimental setup with extraction cell, electrical wires, docking station, and computer for logger setup and downloading data.

2.1. CONVERSION OF PRESSURE MEASUREMENTS TO RESERVOIR LIQUID VOLUMES A mass of water was used to fill the volume below the transducer, the fitting, the stainless steel pipe, and part of the reservoir. Negative pressure was then generated in the reservoir by using a 30-mL syringe for over 1 min to pull out any air bubbles that may have been trapped in the pipe or the fitting. A syringe with a long needle was used to empty the reservoir while leaving the pipe and the fitting filled with water. The required mass of water was then added to the reservoir.

ACS Paragon Plus Environment

8

Page 9 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

The cell was closed and sealed with a sealing screw. The contents were then centrifuged for 5 min at different rotation speeds ranging from 500 rpm to 3000 rpm. Recorded pressure and temperature during centrifugation for a reservoir water mass of 5.54 g are presented in Figures 3 and 4, respectively. There are slight temperature variations with time for the different water contents. These variations would have slightly affected pressure measurements because the cell is sealed. Figure 5 shows averaged pressures over approximately 5 min of spinning for different rotation speeds and different reservoir water contents ranging from 5.54 g to 19.29 g. These averaged pressures were obtained after correcting for the effect of temperature variation with time.

Figure 3 – Pressure recorded for a water mass of 5.54 g at different rotation speeds. (Dotted line): 500 rpm, (dash-dotted line): 1000 rpm, (dashed line): 1500 rpm, (continuous line): 2000 rpm, (thick dashed line): 2500 rpm, (thick continuous line): 3000 rpm.

ACS Paragon Plus Environment

9

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 37

Figure 4 – Temperature recorded for a water mass of 5.54 g at different rotation speeds. (Dotted line): 500 rpm, (dash-dotted line): 1000 rpm, (dashed line): 1500 rpm, (continuous line): 2000 rpm, (thick dashed line): 2500 rpm, (thick continuous line): 3000 rpm.

Figure 5 – Average pressure recorded over 5 min at different rotation speeds and water masses in the reservoir. (□): 5.54 g, (+): 8.16 g, (○):10.63 g, (∆): 13.27 g, (): 17.25 g, (*): 19.29 g.

ACS Paragon Plus Environment

10

Page 11 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

The radial pressure gradient generated in a liquid layer by a centrifugal field is given by:  

  ,

(1)

where P, is the pressure, r, is the radial position from the axis of rotation,  is the liquid density, and  is the rotation speed. Equation (1) is integrated between the radial positions of the water level surface and the transducer to obtain the following equation: 

  −    

  !



  ,

(2)

where  is the air pressure inside the sealed cell. This pressure can be affected by temperature changes as follows: "   

#$%&  ' #$%&

,

(3)

" " and  are air temperature and pressure in the cell before sealing. where (

 ! ,



are the radial positions of the transducer and the water level, respectively.

The radial position of the water level,

 ,

can be obtained from the water height in the

reservoir and the radial position at the bottom of the reservoir,

 +



**

**

−  ,

, as follows: (4)

where  is the reservoir water height, which can be easily calculated from the water volume in the reservoir,

,$-.& /,$-.&

, and the reservoir diameter, 0 ! +* .

ACS Paragon Plus Environment

11

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 37

Equation (2) is used to calculate the centrifugal pressure applied by water in the reservoir on the pressure transducer. Table 1 summarizes measured radial positions of reservoir bottom surface, transducer sensor surface, and water levels for different reservoir water volumes used in the calculation of the centrifugal pressure. These radial positions were obtained by adding and subtracting different caliper measurements on the cell. Table 1 – Radial positions of the reservoir bottom, transducer, and reservoir water levels used in pressure measurements and calculations Water level index 1 Water mass (g) 5.54 Water reservoir height (mm) 13.5 129.4 ** (mm) 115.6  + (mm) 125.3  ! (mm)

2 8.16 19.9 129.4 109.3 125.3

3 10.63 25.9 129.4 103.2 125.3

4 13.27 32.3 129.4 96.8 125.3

5 15.23 37.1 129.4 92.0 125.3

6 17.25 42.0 129.4 87.1 125.3

7 19.29 47.0 129.4 82.1 125.3

The difference function, noted as 12  , between calculated and measured variations of the centrifugal pressure with rotation speed for different reservoir water contents is presented in Figure 6. This difference function appears to show a trend that depends only on rotation speed and not on reservoir water volume. The largest deviation between the different water content plots is 0.2 bar, at 3000 rpm. The difference function, 12  , is thought to be caused by the effect of centrifugal force on the pressure sensor. Deflection on the pressure transducer by centrifugal force may require frequent calibrations. Effect of transducer deflection can be minimized if the data logger was placed such that the surface of the sensor is parallel to the centrifugal force. However, this configuration was not possible due to the size of the data logger.

ACS Paragon Plus Environment

12

Page 13 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 6 – Difference between calculated and measured pressures during centrifugation , 12  , for different rotation speeds and reservoir water contents. (□): 5.54 g, (+): 8.16 g, (○):10.63 g, (∆): 13.27 g, (): 17.25 g, (*): 19.29 g.

After adding the difference function, 12  , the pressure applied by different water levels ( −  ) can be calculated as follows: 

  −    

  !



  + 

 12  , For i=1-7

(5)

If the pressure at water level with index = 1 (in Table 1) is considered to be known, pressure at any other water level can be calculated as follows: 

  −     

**

−  +   −

**

−  +  

(6)

Figure 7 shows comparisons of calculated reservoir water volumes from pressure measurements using Eq 6 and estimated values from water masses and density presented in Table 1. Observed deviations can be attributed to uncertainty in measured radial positions during centrifugation and

ACS Paragon Plus Environment

13

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 37

effects of centrifugation and temperature on the pressure sensor. The largest deviations are observed at the lowest rotation speed of 500 rpm. The experimental technique presented in this work is used to monitor diluted bitumen volume evolution with time during centrifugal extraction from oil sand. Absolute filtrate volume error is not expected to have significant effect on estimated relative time evolution of filtrate volume during centrifugation.

Figure 7 – Water volumes in the reservoir: (symbols) calculated using pressure measurements and (lines) estimated from water masses presented in Table 1. (+): 8.16 g, (○):10.63 g, (∆): 13.27 g, (): 17.25 g, (*): 19.29 g. Calculations performed using measured radial position at reservoir bottom of 129.4 mm.

ACS Paragon Plus Environment

14

Page 15 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

2.2. CONVERSION OF PRESSURE MEASUREMENTS TO DILUTED BITUMEN VOLUMES DURING OIL SAND FILTRATION

After measuring the centrifugal pressure exerted by a known amount of water in the filtrate reservoir,   , diluted oil sand is added on a filter paper in the cell and then centrifuged. Measured pressure time variation !  can be converted to diluted bitumen liquid height   using the following equation: !  , −    

   



**

−   −

**

−    

 ,

(7)

where  is the density of diluted bitumen. The density can be measured at the end of the experiment after extracting water and diluted bitumen from the filtrate reservoir.

3. RESULTS AND DISCUSSION 3.1. INSTRUMENTED CENTRIFUGAL FILTRATION OF OIL SAND Water was used to fill the bottom of the filtrate reservoir and the volume connecting the reservoir to the pressure transducer to prevent sensor contamination with diluted bitumen and provide a minimum reaction pressure to counter the effect of centrifugal force deflection on the pressure sensor. The cell reservoir was initially filled with 10.05 g water. A syringe was used to pull vacuum in the reservoir and ensure that no air bubbles were trapped in the tubing and the fitting connecting the pressure transducer to the reservoir. The cell was then spun at 500 rpm for 5 min.

ACS Paragon Plus Environment

15

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 37

The reservoir was emptied using a long needle syringe and filled again with a water mass of 5.4 g. The cell was then closed and sealed with a sealing screw. It was placed in a swinging bucket of a JS-7.5 rotor in the Avanti J-30 I centrifuge and spun at 1000 rpm for 5 min. Recorded pressure measurements are presented in Figure 8. Initial absolute pressure measured before centrifugation was used as a reference and was subtracted from all subsequent pressure measurements.



Figure 8 – Pressure and temperature recorded for cell spinning at 1000 rpm and reservoir containing 5.4 g water.

A diluted oil sand sample was prepared by adding 49.9 g oil sand with 15.8 g toluene in a jar and mixing in a shaker for 1 h. Dean-Stark analysis of the oil sand sample gave the following composition by mass: 10.51% bitumen, 0.48% water, and 89.01% solids. Note that the oil sand sample also contained coal that could not be characterized with Dean Stark Analysis. The diluted

ACS Paragon Plus Environment

16

Page 17 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

bitumen layer above the oil sand cake was extracted after mixing the oil sand with toluene. A mass of 34.2 g diluted oil sand cake was transferred from the jar to the cell and compacted to form a cake of uniform thickness. A mass of 7.9 g diluted bitumen was added to the cake in the cell. Part of this liquid mass would form a layer above the oil sand cake and fill localized void space in the cake and between the cell wall and the o-ring and aluminum ring. The cell was then sealed and spun for 1 h at 1000 rpm. Recorded pressure and temperature data are presented in Figure 9.

Figure 9 – Pressure and temperature recorded during diluted oil sand filtration and desaturation at 1000 rpm. The filtrate reservoir initially contained 5.4 g water.

ACS Paragon Plus Environment

17

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 37

In order to take into account the effect of temperature changes in the sealed cell on air pressure during centrifugation, water and diluted oil sand pressure measurements are corrected using the following equation: " *   !  − 

#$%&  ' #$%&

(8)

where " and (" are initial air pressure and temperature before centrifugation. Original and temperature-corrected pressure data are presented in Figures 10. This figure shows that the pressure correction subtracts the effect of temperature increase (shown in Figure 10) on air pressure and demonstrates that the centrifugal pressure applied by diluted bitumen becomes nearly constant with time. Measured pressure becomes constant when equilibrium is reached between capillary and adsorptive forces and centrifugal forces in the porous cake. Adsorptive forces are responsible for binding a liquid layer on the surface of solid particles when the pores are desaturated [12]. Note that Temperature is measured inside the data logger. When there is rapid and steep change in air temperature caused by the centrifuge cooling system, for example, measured pressure would be instantly affected but temperature inside the data logger may require a time delay to adjust to the temperature change. This effect can make pressure correction for temperature variation challenging. Figure 11 shows that during the first few seconds of centrifugation, the initial rapid increase in pressure is caused mainly by water pressure in the reservoir (5.4 g water) during the acceleration phase from 0 to 1000 rpm. Theory has shown that, during filtration centrifugation at constant rpm, the flow rate is always decreasing with time [4]. Measured pressure reflects liquid volume

ACS Paragon Plus Environment

18

Page 19 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

in the reservoir. Thus, a criterion used to select time t = 0, at the end of acceleration and the initiation of filtration at constant rotation speed of 1000 rpm, is that the derivative of pressure with respect to time ,

 

, should become a decreasing function of time. Time t = 0 is indicated in

Figure 11 by the vertical dashed line.

Figure 10 – Pressure recorded during oil sand-solvent filtration and desaturation at 1000 rpm: (black): original data, (red): corrected data.

ACS Paragon Plus Environment

19

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 37

Figure 11 – Right axis: Pressure curve during the acceleration period and initiation of filtration. Left axis: rate of pressure change with time. Selected time t = 0 is indicated by the arrow in both figures. It corresponds to the initiation of filtration at constant rotation speed.

Corrected pressure measurements are converted into diluted bitumen volumes in Figure 12 using Eqs 7 and 8. Average measured water pressure,  (shown in Figure 9 and corrected with Eq 8), during centrifugation at 1000 rpm was used in Eq 7. Volume data derived from pressure measurements were fitted using a ratio of 3rd-order polynomials as shown in Figure 12. The flow rate plot presented in Figure 13, with the right axis, was obtained after calculating the time derivative of diluted bitumen volume evolution (Figure 12). Diluted bitumen volume and flow rate plots are qualitatively similar to the plots predicted numerically in our previous theoretical paper [4].

ACS Paragon Plus Environment

20

Page 21 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 12 – Diluted bitumen volume calculated from recorded pressure during diluted oil sand filtration and desaturation at 1000 rpm on semilog scale. (dots): volume converted from pressure data. (line): fitted function of 3rd-order polynomial ratio. Vertical dashed line corresponds to the transition time from filtration to desaturation.

Figure 13, with the left axis, shows the variations of the product (flow rate × time) vs time on logarithmic scale. The time at the maximum in these plots corresponds to the transition time from filtration to desaturation in the centrifugal extraction, as demonstrated in our previous paper [4]. Accordingly, it took 54 s of centrifugal filtration at 1000 rpm for the liquid layer above the cake to filter completely through the saturated oil sand layer.

ACS Paragon Plus Environment

21

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 37

Figure 13 – Right axis: Experimental flow rate of diluted bitumen during centrifugal filtration and desaturation. Vertical dashed line corresponds to the transition time from filtration to desaturation. Left axis: Product: Flow rate × time. Vertical dashed line corresponds to the transition time from filtration to desaturation.

3.2. EXTRACTION CHARACTERIZATION Samples were collected from the diluted bitumen layer above the cake before extraction and from the filtrate reservoir after extraction. The density of diluted bitumen samples was measured in a DMA 4500 densitometer. Bitumen content was measured by injecting a known mass of diluted bitumen on a filter paper. After 15 min of evaporation in a fume hood, the filter paper mass was measured to determine the mass fractions of bitumen and evaporated solvent in the diluted bitumen samples. Measurement results for density and bitumen content in diluted bitumen samples collected before and after extraction are presented in Table 2.

ACS Paragon Plus Environment

22

Page 23 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Table 2 – Experimental composition of original and diluted oil sand samples before and after extraction Original oil sand sample Bitumen mass fraction from Dean-Stark analysis (%)

10.5

Bitumen mass fraction from thermogravimetric analysis (%)

11.5

Sample composition before extraction Sand mass fraction in cake before extraction (%)

87.0

Bitumen and coal mass fraction in cake before extraction (%)

5.5

Solvent and light component fraction in cake before extraction (%)

7.5 3

Density of diluted bitumen layer above cake before extraction (g/cm ) Bitumen content in diluted bitumen before extraction (%)

0.8973 25.7

Sample composition after extraction Sand mass fraction in cake after extraction (%)

91.8

Bitumen and coal mass fraction in cake after extraction (%)

4.2

Solvent and light component fraction in cake after extraction (%) 3

3.9

Density of diluted bitumen in reservoir after extraction (g/cm )

0.8997

Bitumen mass fraction in diluted bitumen after extraction (%)

27.4

Cake porosity, 4! , was calculated using diluted oil sand and bitumen densities as follows: /

8/

4!  / 5$678/ 7%9:-.7 ;%9 5$67  0.341, 5$67

7%9:-.7