SPE 129728

Swelling/Extraction Test of a Small Sample Size for Phase Behavior Study

J.S. Tsau, L.H. Bui, and G.P. Willhite, SPE, University of Kansas

Copyright 2010, Society of Petroleum Engineers

This paper was prepared for presentation at the 2010 SPE Improved Oil Recovery Symposium held in Tulsa, Oklahoma, USA, 24–28 April 2010.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its
officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to
reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract

Swelling/extraction tests are single-contact phase-behavior experiments to measure the solubility of CO2 dissolved in crude
oil and the amount of hydrocarbon that CO2 can extract or vaporize from crude oil. The tests are commonly conducted in a
visual PVT cell with a large sample size (40-100cc). In this paper, an easy operated apparatus capable of determining phase
behavior with a significantly smaller sample size (3 to 14 cc) is described. The apparatus consists of a high-pressure view
cell, high-pressure and precision syringe pump filled with CO2, a water bath, and accessories to measure the temperature and
pressure. The device is capable of determining vapor-liquid, liquid-liquid and vapor-liquid-liquid equilibrium commonly
observed in a high pressure CO2 enhanced oil recovery process. The solubility of CO2 in oil, the expansion volume of oil due
to the dissolution of CO2 as well as the phase transition during the test were quantified with excellent reproducibility. The
molar volume of oil saturated with CO2 correlated linearly with the mole fraction of dissolved CO2 suggesting ideal mixing in
the liquid phase. The phase behavior between CO2 and crude oil samples with different composition, temperature and
pressure is discussed.

Introduction

A swelling/extraction test is a common phase behavior experiment to determine reservoir fluid volume and composition
changes due to CO2 dissolution at reservoir temperature. The test is usually conducted in a constant volume, high pressure
view cell initially filled with a predetermined amount of stock-tank oil. CO2 is injected progressively with pressure increase
in a number of discrete steps. The change of crude oil volume due to the swelling is measured and the amount of CO2
dissolved in the oil is calculated by assuming that vaporization of crude oil components into the equilibrium vapor phase is
neglible until the MMP is approached. At a higher pressure, light hydrocarbon components are vaporized or extracted and the
hydrocarbon rich oil phase shrinks with increasing pressure. Phase behavior at high pressures can be viewed. The saturation
pressure, solubility of CO2 and swelling factor are commonly used for tuning the equation of state (EOS) for modeling phase
behavior.

Various sizes of view cells, 140 cc (Hand et al., 1990), 170 cc (Harmon et al., 1988), 190 cc (Orr et al., 1981) and
450 cc from commercial laboratory have been used to study the phase behavior. The sample size for the swelling test
described by Holm and Josendal (1982) is 30% of the cell volume. Thus, the test performed in these cells would require 40cc
to more than 100 cc of liquid sample. An inherent difficulty of large volume cells is the length of time required to reach
equilibrium after the pressure is changed and the difficulty of mixing large volumes of oil and gas under pressure.

This paper describes a small volume high pressure view cell that was developed to investigate the mass-transfer
process occurring in swelling/extraction tests when CO2 dissolves in the oil phase. The total cell volume is 26 cc with the
sample size adjustable for the test. This paper summarizes the design of apparatus, experimental procedure and the tests
conducted on CO2 with three crude oils produced from reservoirs in Kansas.

Experimental Apparatus and Procedure

Apparatus. The apparatus consists of a high-pressure view cell, high-pressure and precision syringe pump filled with CO2, a
water bath, and accessories to measure the temperature and pressure. Figure 1 shows a diagram of the experimental
2 SPE 129728

apparatus setup including an ISCO 100DM syringe pump, a high-pressure view cell, a water bath, a cathetometer,
temperature and pressure measurement devices connecting to a computer data acquisition system (Ren et al., 2008). The
temperature of the water bath is controlled by a Fisher Isotemp Immersion Circulator. The high pressure view cell is
equipped with high pressure gauge glass window to allow visual observations of fluids under experimental conditions. The
view cell is made of stainless steel. The gauge glass window is rated at a maximum temperature of 280oC and pressure of
4000 psi. Pressure in the view cell is measured by a 5000 psi Heise DXD Series 3711 precision digital pressure transducer.
Mixing is accomplished by moving a PTFE coated stir bar inside the view cell using a rare-earth magnet in a slot on the
outside of the cell. The magnet is raised and lowered manually by a pulley system but could be automated. An Eberbach
5160 cathetometer is used to measure the height of the liquid in the view cell.

Procedure. In a typical swelling experiment, the ISCO pump is filled with CO2. Pressure of the pump is set at the maximum
anticipated pressure with a constant-pressure mode. The pump automatically adjusts the volume of CO2 to maintain the
desired pressure. Temperatures of the gas lines are maintained above the critical temperature of CO2 to avoid CO2
condensation inside the lines. Temperature of the water bath is set at the desired temperature. A predetermined volume of
crude oil is carefully injected into the view cell to avoid liquid droplets on the wall of the view cell. The view cell is
connected to the gas lines and then immersed into the water bath. The height of the liquid sample inside the view cell is
measured using the cathetometer. The volume of the liquid sample is calculated using a pre-calibrated curve which correlates
the volume with the measured height.

When the system is thermally equilibrated, the gas lines and the view cell are flushed with CO2 at low pressure to
remove any residual gas or air. The cell pressure is increased in discrete steps by CO2 injection from the top of the view cell.
CO2 injection is stopped when a desired pressure is achieved. During pressurization process, the stir bar inside the view cell
is used to mix the liquid and vapor phases, accelerating the mass transfer between the gas and liquid phase. When the system
is in equilibrium, the height of the liquid sample in the view cell is measured with a cathetometer. Equilibrium between
pressure changes takes from 30-60 minutes after vigourous mixing. The pump condition (temperature, pressure & final
volume of CO2), temperature of gas lines and the view cell condition (temperature & pressure) are recorded. All the data are
transferred into a specially designed spreadsheet to calculate the solubility, density of the liquid solution, molar volume and
volume expansion. As noted earlier, the material balance calculations are based on the assumption that the amount of
hydrocarbon extracted into the vapor phase is negligible. The composition of the liquid phase is based on the mass balance
by determining the amount of CO2 dissolved in the liquid. This assumption appears to be valid over a wide range of
pressures. This method yields high resolution of solubility data (often better than ±0.0001), pressure with accuracy of ±3 psi,
temperature of ±0.01 °C, density up to ±0.4% and volume expansion to ±0.05% (Ren et al., 2008). At the end of the
experiment after depressurization, the view cell is cleaned with methylene chloride, acetone solution and blown dry with
compressed air.

Apparatus Verification. The apparatus and method were verified by comparing the phase equilibrium data of carbon
dioxide in n-decane at 71.1 °C with published data. Figure 2 illustrates the p-x phase equilibrium of CO2 and n-decane binary
system. The experimental data of this setup are in excellent agreement with literature data (Nagarajan et al., 1986, Jennings
et al., 1996). The principal assumption in analyzing these data is that the amount of liquid component in the equilibrium
vapor phase is negligible. This assumption was verified by Ren et al. (2008) who demonstrated that the percentage of decane
in CO2 vapor phase was less than 0.13%.

Results and Discussion

Table 1 gives the laboratory measured properties of oil used in this study. Figure 3 shows the carbon number distributions
for each of stock-tank oil determined by the gas chromatograph. Oil A is the lightest oil among all three oils as its
composition has the least mole fraction of detectable C30+. As the mole fraction of heavy component C30+ increases from
0.031 in oil A to 0.229 in oil C, the API gravity of oil decreases from 39.6 to 24.3.

The data obtained from the view cell includes solubility of CO2 in oil, density and molar volume of CO2-saturated
liquid solution, and the relative volume change as a result of swelling/extraction at reservoir temperature and specified
pressures. It is useful to examine possible correlations between composition and pressure for oil saturated with carbon
dioxide. Orr and Silva (1983) correlated the molar density of decane with mole fraction of carbon dioxide for the decane data
at 71 ºC over a wide range of pressures assuming ideal mixing. Equation 1 gives the molar volume of a carbon dioxide-oil
solution assuming that ideal mixing occurs when carbon dioxide dissolves in oil and the molar volumes of the carbon dioxide
and oil are independent of pressure at a given temperature.
SPE 129728 3

V = xCO2 VCO2 + (1 − xCO2 )Vo ...............................(1)

or
V = Vo + (VCO2 − Vo ) xCO ...................................(2)
2

where xCO2 is the mole fraction of CO2, Vo and VCO2 are the molar volume of oil and CO2, respectively. V is the molar
volume of oil saturated with carbon dioxide. Equation 2 suggests that when ideal mixing exists, graph of V versus xCO2 will
be a straight line with a slope of (VCO2 − Vo ) and an intercept of Vo . Figure 4 is a plot of the molar volume of saturated
decane versus mole fraction of carbon dioxide. The molar volume of decane at 71 °C is 200.65 cc/mol compared to 201
cc/mole determined experimentally. The molar volume of carbon dioxide in solution is 42.32 cc/mol.

Figure 5 is a plot of the molar volume of different crude oils saturated with carbon dioxide at its reservoir
temperature. The molar volume of crude oil A at 105 ºF is 266.12 cc/mol compared to the measured molar volume of 263.41
cc/mol. The molar volume of carbon dioxide in solution is 30.09 cc/mol. The molar volume of crude oil B at 110 ºF is
306.99 cc/mol compared to the measured molar volume of 303.44 cc/mol. The molar volume of carbon dioxide in solution is
26.23 cc/mol. The molar volume of crude C at 78 ºF is 402.52 cc/mol compared to the measured molar volume of 399.20
cc/mol. The molar volume of carbon dioxide is solution is 34.39 cc/mol. These correlations suggest that carbon dioxide
dissolves in oil at constant temperature to form an ideal solution. The molar volumes of dissolved carbon dioxide and crude
oil are not a strong function of pressure over the range of pressure investigated. The correlation of molar volume data for
these crude oils suggests that the swelling behavior over a wide range of pressures could be obtained by measuring swelling
at two pressures.

Figure 6 shows swelling/extraction curve for the stock-tank-oil B/CO2 system studied at 110 ºF. The swelling factor,
defined as the ratio of the oil volume at a given pressure to its initial volume at atmospheric condition, increases with the
pressure as the dissolution of CO2 in the oil increases. The swelling factor of this oil increases to 1.21 with the maximum
volume expansion of 21% when 0.728 mole fraction of CO2 was dissolved in it at 1150 psig. At a higher pressure, the light
components of oil are extracted into CO2 rich phase and the oil starts to shrink and the composition of the liquid phase can no
longer be computed by material balance. Analysis of either the equilibrium liquid or vapor phases would be necessary to
determine the swelling. However, it appears that rate of extraction is faster than the rate of swelling as CO2 dissolved in the
oil. At the end of this experiment, 39.2 volume % of the initial oil has been extracted by carbon dioxide. Two phases exist
over the pressure range tested. At low pressure, the CO2 vapor phase is immiscible with hydrocarbon liquid phase while at
moderate high pressure, critical CO2 is equilibrated with hydrocarbon liquid phase.

Effect of Temperature. The effect of temperature on solubility of CO2 in oil and the corresponding change of swelling
factor are presented in Figure 7 and 8, respectively. At a given pressure, the solubility of CO2 decreases with the temperature
as indicated in Figure 7. As a result, the swelling factor of oil decreases with the temperature as less amount of CO2 is
dissolved in the oil.

Extraction of light hydrocarbons starts at a lower pressure at a lower temperature. For example, in Figure 8, the
extraction starts (inferered by decrease in the swelling factor) at 1150 psig at 105 °F, whereas it starts at 1350 psig at
temperature of 125 °F. The ability of CO2 to extract hydrocarbon from the crude oil depends on its density. At higher
temperature, a higher pressure results in a density equivalent to its density at a lower temperature. Holm and Josendal report
that the extraction of liquid hydrocarbons into CO2-rich vapor phase occurs when the density of CO2 is at least 0.25 to 0.35
gm/cc. The extraction of oil B starts at density of CO2, 0.26 gm/cc at 105 ºF and 1150 psig. At 125 ºF, the pressure of CO2
needs to increase to have an equivalent density to start the extraction and it is in the neighborhood of 1300 psia.

Similar behavior is also observed for oil A. Figure 9 shows the swelling/extraction curves of stock-tank oil A with
carbon dioxide. The extraction starts at 1150 psig at 105 °F whereas it starts at 1100 psig at 98 °F. The maximum swelling
factor is 1.41 at lower temperature of 98 °F as compared to 1.28 at 105 ºF.

Three-Phase Region. Oil A/CO2 exhibits liquid-liquid-vapor separation in a specific range of pressures consistent with
observations of phase behavior of CO2 and Maljamar separator oil (Orr et al., 1981). Figure 10 shows the volume fraction of
phases observed in the view cell at pressures between 1100 and 1225 psig. At pressure below 1150 psig, liquid-vapor phase
coexists with liquid vol% increasing with pressure as the oil swells. As the pressure approaches 1160 psig, a CO2-rich
middle phase appears. The vol% of this middle phase increases with the pressure until 1225 psig wherein it merges with the
vapor phase and the system becomes two phases (L+V). This three phase region exists in a narrow pressure range of 50 psi
which is difficult to identify without careful experimental measurements.
4 SPE 129728

Oil C was tested at 78 ºF, a temperature below critical temperature of carbon dioxide (87.98˚F). The
swelling/extraction behavior of Oil C is shown in Figure 11. A maximum of 15% of volume expansion was observed before
the system pressure reached the saturation pressure of 930 psig where the CO2 vapor phase became a CO2 rich-liquid phase.
Extraction starts at 872 psig at where three phases are observed, 1) Oil-rich liquid phase, 2) CO2-rich liquid phase and 3) CO2
vapor phase. The three phases exist in a very narrow pressure range between 872 and 930 psig. When the pressure is above
930 psig, the CO2 vapor phase merges with the CO2-rich liquid phase to form two liquid phases, one hydrocarbon rich and
one CO2-rich.

Effect of Initial Oil Volume. The effect of initial oil volume on the results from swelling/extraction tests is demonstrated in
Figure 12. The comparisons are made among curves at different initial oil volume of 3, 9 and 14 cc which represent 12, 35
and 54 % of total cell volume. All curves display both swelling and extraction behavior wherein the rate of extraction is
fastest when the initial sample volume is the smallest. The change of relative volume of oil due to the extraction varies with
the initial oil volume. The smaller initial oil volume with larger gas-phase volume, the greater amount the hydrocarbon can
be extracted from the oil. This similar effect has been reported by Hand et al. (1990) and was first noted by Holm and
Josendal (1982) in their work where a 30% of initial volume of sample was used for the swelling test.

Estimation of MMP. The relationship between the phase behavior observed in swelling/extraction tests and slim-tube
minimum miscibility pressure (MMP) has been investigated by several researchers. Harmon and Grigg (1988) reported a
relationship between the pressure required to initiate significant extraction in swelling/extraction tests and the MMP from
slim-tube experiment. They proposed that a rapid rise in CO2 upper-phase density of measurement due to the extraction of
hydrocarbon components from the crude oil corresponds to the process by which multiple-contact miscibility is developed.
However, Hand and Plnczewski (1990) concluded no such direct relationship between the two tests because the vapor phase
density, dominated by high solvent CO2 concentration, is not a sensitive indicator of the onset of major extraction, or of
MMP. In this work, we observed MMP can be graphically derived from the extraction test. By examining the extraction test
results with MMP measured from the slim-tube experiment, a relationship exists between these two tests if the initial oil
volume tested in view cell is small (12%) and the relative volume of oil due to extraction falls below 0.8 over the pressure
range investigated. Figure 13 present swelling/extraction test curves of oil B/CO2 system at 110 ºF and 125 ºF. The rate of
slope changes in two distinct stages in each of the two extraction curves. Drawing lines through the major extraction and
secondary stages, the pressure at the intersection of these two lines is close to MMP determined with the slim-tube
experiment. As shown in this figure, the pressures at the intersection point are 1340 psig and 1640 psig at 110 ºF and 125 ºF,
respectively. The MMP determined from slim-tube for oil B/CO2 were 1350 psig and 1650 psig (Bui et al., 2010). Figure 14
shows another swelling/extraction curve for oil A/CO2 at 105 ºF. The MMP determined graphically from this test is 1240
psig while MMP measured from slim-tube experiment was 1250 psig.

For oil C/CO2 system, the MMP can not be determined graphically from the plot (see Figure 11) as it lacks of
secondary stage of extraction as measured from this setup. In addition, the relative volume measured is 1.0 indicating very
low concentration of extractable hydrocarbons. MMP was not determined for this oil, however, the slim-tube experiment
showed a low recovery efficiency of 50% at pressure of 1000 psi (Tsau et al., 2008).

Conclusions

1. The swelling of crude oil due to the dissolution of CO2 was determined accurately in a new apparatus using small sample
sizes.
2. Swelling of crude oil by dissolved CO2 correlated with the molar volume of crude oil and apparent molar volume of CO2
over the range of pressure where neglibile oil was extracted into the CO2 phase.
3. The pressure at which extraction starts by CO2 depends on the initial volume of oil and temperature tested.
4. A three phase region (VLL) equilibrium was identified with the apparatus and occurred over a narrow range of pressure.
5. Using 12 % of cell volume sample size, the MMP estimated by the swelling/extraction test graphically is close to what
determined from the slim-tube experiment.

Acknowledgements

The authors wish to acknowledge the funding support by Research Partnership to Secure Energy for America (RPSEA) small
producer program, RPSEA Contract DE-AC26-07NT42677/Subcontract 07123-03, and Scott Ramskill of TORP for help in
design and manufacturing of the apparatus.
SPE 129728 5

References

Bui, L.H., Tsau, J. S., and Willhite, G. P.: “Laboratory Investigations of CO2 Near-miscible Application in Arbuckle Reservoir,” paper
SPE 129710 to be presented at the Improved Oil Recovery Symposium, Tulsa, OK. April 24-28, 2010.
Hand, J. L. and Plnczewshl, W. V.: “Interpretation of Swelling/Extraction Tests,” SPERE (November 1990) 595-600.
Harmon, R. A. and Grigg, R. B.: “Vapor-Density Measurement for Estimating Minimum Miscibility Pressure,” SPERE (November 1988)
1215-1220.
Holm, L.W. and Josendal, V.A.: “Effect of Oil Composition on Miscible-Type Displacement by Carbon Dioxide,” SPEJ (February 1982)
87-98.
Jennings, D. W. and Schucker, R. C. J.: “Comparison of High-Pressure Vapor-Liquid Equilibria of Mixtures of CO2 or Propane with
Nonane and C9 Alkylbenzens,” J. Chem. Eng. Data, 41, 831, 1996
Nagarajan, N. and Robinson, R. L. J.: “Equilibrium Phase Compositions, Phase Densities, and Interfacial Tensions for CO2 + Hydrocarbon
Systems, 2. CO2 + n-Decane,” J. Chem. Eng. Data, 31, 168. 1986
Orr, F. M., Yu, A. D. and Lien, C. L.: “Phase Behavior of CO2 and Crude Oil in Law-Temperature Reservoirs,” SPEJ (August 1981) 480-
492.
Orr, F. M. and Silva, M. K.: “Equilibrium Phase Compositions of CO2/Hydrocarbon Mixtures – Part 1: Measurement by a Continuous
Multiple-Contact Experiment,” SPEJ (April 1983) 272-280.
Ren, Wei and Scurto, A. M.: “High –Pressure Phase Equilibria with Compressed Gases,” Review of Scientific Instruments, 78, 125104,
2007.
Tsau, J.S., Bustamante, V, Green, D., Barnett, B., and Dale, J. L..: “Evaluation of Manson Lease for Improved Oil Recovery Process,”
paper SPE 113985 presented at the Improved Oil Recovery Symposium, Tulsa, OK. April 19-23, 2008.

Table 1 Properties of crude oil used in this study

Oil A Oil B Oil C

Molecular weight 198.0 228.7 340.0
API gravity 39.6 33.3 24.3
Reservoir temperature (ºF) 105 110 78
Viscosity at reservoir
2.85 4.95 117
temperature (cp)

o o
CO2 + n-Decane 71.10 C (160 F)
2000
This work #1
This work # 2
1600
Nagarajan & Robinson Jr.
Pressure (psi)

Jennings & Schucker

1200

800

400

0
0.0 0.2 0.4 0.6 0.8 1.0
Mole fraction of CO2 in the liquid phase
Figure 1 Experimental setup of swelling/extraction test Figure 2 Phase equilibrium data of n-decane with carbon
dioxide
6 SPE 129728

1.4 1.0

1.2
0.8
1.0

CO 2 solubility
Swelling Factor
0.6
0.8

0.6
0.4

0.4
0.2
0.2 Swelling Factor
CO2 solubility
0.0 0.0
0 500 1000 1500 2000 2500
Pressure, psi
Figure 3 Carbon number distributions of crude oils Figure 6 Swelling factor and solubility of carbon dioxide in
crude oil B at 110 ºF

1400

1200

Pressure, psia 1000

800

600
105F
400 110F
115F
200 120F
125F
0
0.0 0.2 0.4 0.6 0.8 1.0
Mole fraction of CO2 in the liquid phase
Figure 4 Molar volume of carbon dioxide saturated n-decane
at 71.1 ºC Figure 7 Effect of temperature on solubility of carbon dioxide
in crude oil B

1.4

1.2

1.0
Swelling Factor

0.8

0.6
105F
0.4 110F
115F
0.2 120F
125F
0.0
0 500 1000 1500 2000 2500
Pressure, psi
Figure 5 Molar volume of carbon dioxide saturated crude oils Figure 8 Effect of temperature on swelling factor of crude oil
at different temperature B
SPE 129728 7

1.6

1.4

1.2

Swelling factor
1.0

0.8

0.6

0.4 14 cc
9 cc
0.2
3 cc
0.0
0 500 1000 1500 2000 2500
Figure 9 Effect of temperature on swelling/extraction of crude Pressure, psi
oil A
Figure 12 Effect of initial sample volume on
swelling/extraction of crude oil B

Figure 13 Estimation of MMP from extraction test of crude oil

Figure 10 Three-phase region LLV of crude oil A at 98 ºF B

Figure 14 Estimation of MMP from extraction test of crude oil

Figure 11 Three- phase region LLV of crude oil C at 78˚F A
(25.56˚C)