Enhanced Solubility and Miscibility of CO₂-Oil Mixture in the Presence of Propane under Reservoir Conditions to Improve Recovery Efficiency

Abstract

The existence of propane (C₃H₈) in a CO₂-oil mixture has great potential for increasing CO₂ solubility and decreasing minimum miscibility pressure (MMP). In this study, the enhanced solubility, reduced viscosity, and lowered MMP of CO₂-saturated crude oil in the presence of various amounts of C₃H₈ have been systematically examined at the reservoir conditions. Experimentally, a piston-equipped pressure/volume/temperature (PVT) cell is first validated by accurately reproducing the bubble-point pressures of the pure component of C₃H₈ at temperatures of 30, 40, and 50 °C with both continuous and stepwise depressurization methods. The validated cell is well utilized to measure the saturation pressures of the CO₂-C₃H₈-oil systems by identifying the turning point on a P-V diagram at a given temperature. Accordingly, the gas solubilities of a CO₂, C₃H₈, and CO₂-C₃H₈ mixture in crude oil at pressures up to 1600 psi and a temperature range of 25–50 °C are measured. In addition, the viscosity of gas-saturated crude oil in a single liquid phase is measured using an in-line viscometer, where the pressure is maintained to be higher than its saturation pressure. Theoretically, a modified Peng–Robinson equation of state (PR EOS) is utilized as the primary thermodynamic model in this work. The crude oil is characterized as both a single and multiple pseudo-component(s). An exponential distribution function, together with a logarithm-type lumping method, is applied to characterize the crude oil. Two linear binary interaction parameters (BIP) correlations have been developed for CO₂-oil binaries and C₃H₈-oil binaries to accurately reproduce the measured saturation pressures. Moreover, the MMPs of the CO₂-oil mixture in the presence and absence of C₃H₈ have been determined with the assistance of the tie-line method. It has been found that the developed mathematical model can accurately calculate the saturation pressures of C₃H₈ and/or CO₂-oil systems with an absolute average relative deviation (AARD) of 2.39% for 12 feed experiments. Compared to CO₂, it is demonstrated that C₃H₈ is more soluble in the crude oil at the given pressure and temperature. The viscosity of gas-saturated crude oil can decrease from 9.50 cP to 1.89 cP and the averaged MMP from 1490 psi to 1160 psi at 50 °C with the addition of an average 16.02 mol% C₃H₈ in the CO₂-oil mixture.

1. Introduction

The injection of CO₂ into petroleum reservoirs can increase the oil-recovery factor by reducing the oil viscosity and interfacial tension, swelling the oil volume, extracting light hydrocarbons, and reaching miscibility with in situ crudes [1,2,3], which has been proven to be one of the most promising enhanced oil-recovery (EOR) techniques. The total CO₂-enhanced oil production was 273,000 barrels per day and approximately 142 CO₂ EOR projects were undertaken in the United States by 2020 [4]. In addition, it can substantially improve oil recovery, reduce greenhouse gas (GHG) emissions, and improve economic benefits because oil reservoirs with sealed cap rocks are ideal spaces for CO₂ sequestration. Gozalpour et al. [5] have claimed that up to 60% of the injected CO₂ could be stored in a depleted oil reservoir if the produced CO₂ was not re-injected in U.S. field projects. The application of CO₂ EOR technology not only improves oil recovery but also contributes to reducing greenhouse gas emissions and provides economic benefits by utilizing CO₂ sequestration in depleted reservoirs

Previous studies have proven that adding a certain amount of alkane solvents into the CO₂ stream is a practical approach to not only further improve the oil-recovery performance but also maintain an acceptable cost. Such an idea has been implemented in heavy crudes and bitumen EOR research. Talbi and Maini [6] have investigated the performance of CH₄-C₃H₈ and CO₂-C₃H₈ as solvents for the vapor extraction (VAPEX) process for the in situ recovery of heavy oil with a viscosity of 4500 cP. It has been found that the CO₂-C₃H₈ mixture provided approximately 45% of the original oil-in-place (OOIP) while the CH₄-C₃H₈ mixture is only 32% OOIP at 600 psig and 21 °C. Moreover, the viscosity of Athabasca bitumen can be significantly reduced from over 1000 cP to 28.8 cP by saturating 11.0 wt% of CO₂ and 13.5 wt% of C₃H₈ at a pressure of 4000 kPa and temperature of 20.2 °C [7]. Li et al. [8] have claimed that adding C₃H₈ and/or n-C₄H₁₀ into the CO₂ stream leads to a significant reduction of interfacial tension (IFT) between Lloydminster heavy oil and CO₂, which is from 19.92 mN/m to 10.18 mN/m by adding 24.0 mol% C₃H₈ at 3101 kPa and 21.0 °C and a decrease from 28.31 mN/m to 16.07 mN/m by adding 18.5 mol% n-C₄H₁₀ at 1101 kPa and 21.0 °C, respectively. Luo et al. [9] have claimed that a significantly increased solubility from 40.9 sm³/m³ to 52.4 sm³/m³ and an enhanced viscosity reduction from 412 cP to 195 cP of a CO₂/heavy oil system can be reached when C₃H₈ exists at about 4 MPa. Shen et al. [10] showed that the CO₂-C₃H₈ mixture solvent system achieved higher oil-recovery factors compared to pure CO₂. They claimed that, compared to a 19.4% recovery factor with pure CO₂, the 72% CO₂ and 28% C₃H₈ system achieved up to 23.44% recovery for heavy oil systems. Therefore, the investigation of C₃H₈’s effects on reservoir fluid phase behavior during CO₂-assisted EOR processes is crucial, both from a theoretical standpoint and for practical applications in the field.

The Peng–Robinson equation of state [11,12], i.e., PR EOS, is a common practice widely used for describing the phase behavior of hydrocarbon fluids. Though the Soave–Redlich–Kwong (SRK) Equation of State is also widely used in phased behavior, PR EOS shows better performance for liquid-phase prediction, especially for hydrocarbons or systems involving CO₂. SRK EOS is more suitable for predicting gas-phase properties [13,14]. Jaubert and Mutelet [15] have predicted the bubble point pressures of methane-n-decane-n-paraffins ranging from n-C₁₈ to n-C₃₀. They reported an average deviation of 1.4% in 54 bubble point pressures using PR EOS, which is three times better than the deviations obtained by the LCVM model (a linear combination of the Vidal and Michelsen rules) and 28 times better than by MHV2 (a group contribution equation of state based on the modified Huron–Vidal mixing rule) [16]. Li and Yang [17] proposed a new alpha (α) function for the PR EOS and compared it with three other α functions, i.e., the original Soave-type α function [18], and the α functions proposed by Gasem et al. [19] and Nji et al. [20], respectively. They concluded that their new modified α function for the PR EOS is capable of providing a more accurate prediction of vapor pressures with a percentage average absolute deviation of 1.90% for 59 non-hydrocarbon and hydrocarbon compounds. Later on, Li et al. [21,22,23,24] have utilized the aforementioned modified PR EOS to predict the saturation pressure and three phase boundaries of C₃H₈/n-C₄H₁₀-CO₂-heavy oil, with an absolute average relative deviation (AARD) of 5.07% and 4.58%, respectively. Additionally, they have found that the AARD decreases with the number of the pseudo-component, indicating treating crude oil as a single pseudo-component may cause erroneous results. Therefore, treating oil as multiple pseudo-components is desirable in phase behavior evaluation for reservoir fluids.

A number of phase behavior studies [25] have been conducted for the mixture of various gases and heavy oil/bitumen where the viscosity of crudes can be as high as 7700 cP. Nevertheless, few attempts have been made to quantify the solubility of CO₂, C₃H₈, and their mixture in the light oils. This is because the miscibility between CO₂ and light oil can be easily achieved when the reservoir pressure is higher than its minimum miscibility pressure (MMP). Studies [26,27] have shown that adding alkane solvents into the CO₂ stream could remarkably decrease the MMP of CO₂-rich oil as well. It is worthwhile noting that the gas solubility can be infinite at miscible conditions. As such, the MMP, instead of the solubility of gas, has commonly been the main focus for CO₂ application in light oil reservoirs. However, as for certain shallow light oil reservoirs, the formation pressures are too low to reach the MMP where the immiscible CO₂ EOR may occur, e.g., the Arbuckle reservoir in Kansas [28]. Under the immiscible condition, the solubility of the injected gas(es) in oil is a crucial parameter of the EOR performance. Moreover, in reservoirs with higher oil viscosity, lower oil gravity, low permeability, and severe heterogeneity, CO₂ miscibility can be achieved at lower pressures by adding C₃H₈. The increase in miscibility by adding C₃H₈ can effectively delay gas breakthrough and store more CO₂ in the reservoir. Compared to the immiscible case, creating near-miscible or completely miscible conditions by adding C₃H₈ can increase incremental oil recovery and minimize CO₂ breakthrough [29,30]. In addition, sparse viscosity data have been reported for gas–oil mixtures at high pressures and reservoir temperatures due to the restriction of the conventional viscometer.

In this paper, a systematic but very practical study has been conducted to both experimentally and theoretically determine the enhanced solubility, reduced viscosity, and lowered MMP of CO₂-saturated light oil in the presence of C₃H₈ at elevated temperatures. More specifically, the solubilities of 12 feeds of a CO₂-C₃H₈ mixture in the crude oil system are measured by the constant composition expansion method at a temperature range of 25–50 °C. The viscosities of the gas-saturated oil at high pressures are measured using an inline viscometer. An exponential distribution function together with a logarithm-type lumping method is applied to characterize the crude oil into 1–6 pseudo-component s). The exponents in the binary interaction parameter (BIP) correlations from PR EOS [11] for the C₃H₈-oil and CO₂-oil pair are, respectively, tuned to fit the experimentally measured saturation pressures. As a result, three pseudo-components of oil together with a modified PR EOS yield the most accurate computation of the saturation pressures. The validated PR EOS is then applied to predict the MMPs with the assistance of the tie-line method. It is found that the reduction of MMP and viscosity, as well as the improvement of gas solubility in CO₂ EOR, can be achieved significantly in the presence of C₃H₈.

2. Experimental

2.1. Materials

The crude-oil sample was collected from the Trembley area in Kansas, USA. It has a molecular weight of 234.667 g/mol measured by using the freezing-point depression method (Cryette WR, Natick, MA, USA). The Tertiary Oil Recovery Program (TORP), using gas chromatography simulated distillation method, measures the compositional analysis result of Trembley oil. It can be seen that there is no C₁–C₅, and the heaviest carbon number is C₄₂₊. The solvents of CO₂ and C₃H₈ used in this work have purities of 99.999% and 99.99% (Matheson, Las Colinas Irving, TX, USA), respectively.

2.1.1. Viscosity of Trembley Oil

A cone plate viscometer (DV2T, Brookfield Engineering Laboratories, Middleborough, MA, USA) with a measurement accuracy of 1.0% of the full-scale range is used to measure the viscosity of Trembley oil at the temperature range of 25–70 °C. The temperature is controlled by a heated circulating water bath (TC-650, Brookfield Corporation, Middleborough, MA, USA) with an accuracy of ±0.01 °C. At a given shear rate, the viscous drag of the fluid against the spindle that is immersed in the test fluid through a calibrated spring is measured by the spring deflection. The measured and calculated viscosities are plotted in Figure 1. It can be seen that the viscosity of the Trembley oil used in this work is 9.49 cP at 25 °C. The following equation fits the measured viscosity data well, with R² = 0.9996:

$${log}_{10}\mu =-1.207{log}_{10}T+2.665$$

(1)

where $\mu$ is the oil viscosity in cP, and $T$ is the temperature in °C.

2.1.2. Density of Trembley Oil

The density of Trembley oil is measured at a temperature range of 25–75 °C by using a densitometer (DMA4100M, Anton Paar, Graz, Austria) with an accuracy of ±0.0001 g/cm³. The measured and calculated densities are plotted in Figure 1. The Trembley oil has a density from 0.8552–0.8200 g/cm³ and an API gravity from 32.50–32.33 at a temperature range of 25–75 °C, indicating that a light oil sample is used in this work. The following equation fits the measured density data well, with R² = 1.0:

$$\rho =-0.0007T+0.8727$$

(2)

where $\rho$ is the oil density in g/cm³, and T is the temperature in °C.

2.2. Experimental Setup

All of the pressure–volume–temperature (PVT) measurements for the C₃H₈-CO₂-oil systems are performed using a PVT cylinder (100ML100MPA, Haian Corporation, Nantong, China), as shown in Figure 2. The PVT cell has an inner diameter of 3.643 cm and a total length of 17.51 cm, while it can sustain pressures up to 100.0 MPa over the temperature range of 10 °C to 150 °C. A floating piston in the PVT cell isolates the test fluid from the hydraulic oil, and a mixer is equipped at the bottom of the PVT cell to stir the mixture by making it rotate. A high-pressure syringe pump (500D, Teledyne ISCO Inc., Lincoln, NE, USA) with an accuracy of ±0.5% is employed to compress the mixture by moving the isolation piston downward/upward in the PVT cell. Two digital pressure gauges (MG1-3000-A-9V-R, SSI Technologies, Inc., Janesville, WI, USA) with an accuracy of ±1.0% are set up at the inlet and outlet of the PVT cell. A vacuum pump (DV-200N, JB Industries, Inc., Aurora, IL, USA) is utilized to put all of the apparatuses into a vacuum state. An in-line viscometer (VISCOpro 2000, PAC Corporation, Boston, MA, USA) equipped inside the air bath is used to measure the viscosity of the oil saturated with a CO₂, C₃H₈, and CO₂-C₃H₈ mixture at high pressures, respectively. It has an accuracy of 1.0% of full scale and has a measurement range of 0.2–20,000 cP. The temperature of the air bath is controlled by a fan heater (AF20-600-120-xx-10-4.7, Farnam Custom Products, Arden, NC, USA), which can maintain a temperature from 25–50 °C. A house-made Labview installed on the computer is used to record the flow rate, time, and pressure of the syringe pump.

2.3. Experimental Procedure

For the PVT system validation, prior to the phase behavior measurement for the gas–oil system, the PVT system is first validated by reproducing the vapor pressure of the pure substance of C₃H₈. In this work, two experimental procedures have been compared, namely stepwise and continuous depressurization methods. The stepwise method is considered to be more accurate because the equilibrium state should be reached at each pressure stage. However, it is a time-consuming process with the procedure as follows. The fluid is compressed into a single liquid phase with a pressure above its vapor pressure. The pressure in the PVT cylinder is reduced to the desired pressure by withdrawing the hydraulic oil and 5–6 h is sufficient to reach the equilibrium state. The corresponding pressure and volume are recorded. Subsequently, the pressure is further reduced to the next pressure stage by following the same procedure until the vapor pressure can be identified on the measured pressure–volume diagram. A more efficient approach is the continuous depressurization method, where the pressure is continuously reduced by expanding the cell volume at a constant flow rate. In this work, the rate of 3 cm³/h is adopted, since it has been verified in a number of sources in the literature [8,21,22,23,24,31]. Such a method is able to significantly reduce the experimental time.

For the saturation pressure measurement, the saturation pressures of 12 feeds of C₃H₈-CO₂-oil systems are measured. The compositions of 12 feeds and the measuring temperatures are listed in

Table 1. Compositions of C₃H₈-CO₂-Trembley oil systems with 12 feeds at various temperatures.

Feed #	Composition, mol%			Temperature, °C
	C3H8	CO2	Trembley Oil
1	34.62	0.0	65.38	25, 30, 40, 50
2	46.65	0.0	53.35	25, 30, 40, 50
3	60.08	0.0	39.92	25, 30, 40, 50
4	72.21	0.0	27.79	25, 30, 40, 50
5	0.0	29.29	70.71	25, 30, 40, 50
6	0.0	47.35	52.65	25, 30, 40, 50
7	0.0	60.39	39.61	25, 30, 40, 50
8	0.0	73.74	26.26	25, 30, 40, 50
9	13.03	46.04	40.57	25, 30, 40, 50
10	20.54	41.18	38.28	25, 30, 40, 50
11	14.34	51.05	34.61	25, 30, 40, 50
12	16.16	57.57	26.27	25, 30, 40, 50

. It can be seen that three types of fluids are prepared. Feeds #1–4 are the binary systems composed of C₃H₈ and Trembley oil, while Feeds #5–8 are also the binary systems made up of CO₂ and Trembley oil. Feeds #9–12 are ternary systems containing C₃H₈, CO₂, and Trembley oil. Such an experimental design allows us to determine the solubilities of CO₂, C₃H₈, and their mixture in Trembley oil. The measurements of saturation pressure, volume, and temperature have uncertainties of ±5.28 psi, ±0.3 cm³, and ±0.2 °C, respectively. The entire apparatus is first evacuated by a vacuum pump and the temperature of the air bath is set to the desired value at least 12 h prior to the experimental measurement. Then, C₃H₈ or CO₂ gas is charged into the PVT cell. The molar number of the injected gas can be determined based on the gas volume, pressure, and temperature. Subsequently, Trembley oil is pumped from the transfer vessel into the cell to be mixed with the gas. The amount of oil can be read from the syringe pump. The mixture is compressed into a single liquid phase by moving the piston downward. The single liquid phase in the cell can be confirmed by the pressure response when the pressure rapidly increases with a slight decrease in volume due to the extremely small compressibility of liquid. The equilibrium is considered to be achieved when the pressure is stabilized. Subsequently, the mixture is depressurized at the withdrawal rate of 3 cm³/h. Both pressure and cell volume are automatically recorded by the LabView program in real time. The bubble point pressure can be located by the intersection of two distinct linear lines from the pressure–volume diagram. Finally, the PVT cell is cleaned with toluene followed by acetone to be ready for the next feed.

For high-pressure in-line viscosity measurement, the in-line viscometer (VISCOpro 2000, PAC Corporation, MA, USA) is placed inside the air bath. It contains two magnetic coils inside the stainless-steel sensor. A low-mass stainless piston inside the measurement chamber is surrounded by the fluid sample and magnetically forced back and forth. The time required for the piston to move a fixed distance (about 0.2 inches) between two sensors is accurately related to the viscosity of the fluid in the chamber. An increase in viscosity is sensed as a slowed piston travel time. Since the viscosity of a fluid varies significantly with temperature, a platinum resistance temperature detector (RTD) mounted at the base of the measurement chamber is used to continuously measure the temperature. In this work, the viscosities of Feeds #1, 8, and 12 are measured at 25 °C and 1600 psi, which is higher than their saturation pressures. After the mixture is re-compressed into a liquid state and reaches equilibrium, the fluid sample is displaced from the PVT cell to the viscometer to measure the viscosity of the gas-saturated oil at high pressure. The viscosity measurement has an uncertainty of ±0.016 cP.

3. Mathematical Formulations

3.1. Oil Characterization

The properties of critical temperature, critical pressure, molecular weight, specific gravity, and acentric factor are required for each component in the vapor–liquid equilibrium calculations. Trembley oil is a light oil with 0.8485 mol% of the heaviest component C₄₂₊, as listed in

Table 2. Compositional analysis of the trembley oil.

Carbon No.	wt %	mol%	Carbon No.	wt%	mol%
C1	0.0000	0.0000	C23	2.4646	1.7589
C2	0.0000	0.0000	C24	2.4059	1.6459
C3	0.0000	0.0000	C25	2.1876	1.4370
C4	0.0000	0.0000	C26	2.0541	1.2977
C5	0.0000	0.0000	C27	2.0280	1.2340
C6	0.5450	1.4653	C28	1.8724	1.0989
C7	3.2287	7.4656	C29	1.8359	1.0405
C8	3.9144	7.9396	C30	1.9715	1.0802
C9	5.8007	10.4787	C31	1.8066	0.9581
C10	4.5822	7.4615	C32	1.4699	0.7553
C11	3.8471	5.7023	C33	1.4436	0.7194
C12	3.3481	4.5539	C34	1.2218	0.5910
C13	4.7586	5.9800	C35	1.3866	0.6517
C14	4.6292	5.4060	C36	0.9497	0.4340
C15	3.9684	4.3283	C37	1.3118	0.5833
C16	3.5751	3.6578	C38	1.1327	0.4905
C17	5.1290	4.9415	C39	0.6638	0.2801
C18	4.1833	3.8083	C40	0.8432	0.3469
C19	3.3873	2.9225	C41	0.7403	0.2972
C20	2.4659	2.0220	C42	0.8318	0.3260
C21	3.0100	2.3514	C42+	6.8063	0.8485
C22	2.1989	1.6402	Total	100	100

. It indicates that splitting the plus fraction is unnecessary for this oil sample. The properties of each single carbon number (SCN) can be calculated from empirical correlations as follows.

• A constant Watson factor K_w [32] is assumed based on the molecular weight and specific gravity of the oil sample:

  $$K_w\approx 4.5579M^{0.1578}{\gamma }^{-0.84573}$$

  (3)

  where M and $\gamma$ are the molecular weight and specific gravity of the oil sample, respectively;
• Specific gravity for each SCN [33] can be calculated according to the K_w:

  $${\gamma }_i=6.0108M_i^{0.17947}{K}_w^{-1.18241}$$

  (4)

  for ${\gamma }_{C_{42+}}$,

  $${\gamma }_{C_{42+}}={\displaystyle \frac{z_{C_{42+}}{M}_{C_{42+}}}{{\sum }_{i=1}^N\left(z_i{M}_i/{\gamma }_i\right)}}$$

  (5)

  where ${\gamma }_i$, $M_i$ and $z_i$ are the specific gravity, molecular weight, and mole fraction of the ith SCN;
• The boiling temperature $T_{bR}$ is calculated by using the Soreide [34] correlation, which is a function of ${\gamma }_i$ and $M_i$

  $$T_{bR}=1928.3-\left(1.695\times {10}^5\right)M_i^{-0.03522}{\gamma }_i^{3.266}\times exp\left[-\left(4.922\times {10}^{-3}\right)M_i-4.7685{\gamma }_i+\left(3.462\times {10}^{-3}\right)M_i{\gamma }_i\right]$$

  (6)

  where $T_{bR}$ is the boiling point temperature in °R;
• Then, the critical properties and acentric factor of SCN are estimated by the Kesler [35] correlations, which are the function of $T_{bR}$.

  $$T_{cR}=341.7+811{\gamma }_i+\left(0.4244+0.1174{\gamma }_i\right)T_{bR}+\left(0.4669-3.2623{\gamma }_i\right)\times {10}^5{T}_{bR}^{-1}$$

  (7)

  $$\begin{gathered} P_{cpsi}=exp\left\{8.3634-0.0566{\gamma }_i^{-1}-\left[\left(0.24244+2.898{\gamma }_i^{-1}\right)+0.118857{\gamma }_i^{-2}\times {10}^{-3}\right]T_{bR}+ \\ \left[\left(1.4685+3.648{\gamma }_i^{-1}+0.47227{\gamma }_i^{-2}\right)\times {10}^{-7}\right]T_{bR}^2-\left[\left(0.42019+1.6977{\gamma }_i^{-2}\right)\times {10}^{-10}\right]T_{bR}^3\right\} \end{gathered}$$

  (8)

  where $T_{cR}$ is the critical temperature in °R and $P_{cpsi}$ is the critical pressure in psi;
• Acentric factor $\omega$ [36] can be calculated from $T_{bR}$ as well:

  $$\omega ={\displaystyle \frac{-\ln \left(\frac{P_{cpsi}}{14.7}\right)+A_1+A_2{T}_{bR}^{-1}+A_3\mathrm{l}\mathrm{n}{T}_{bR}+A_4\mathrm{l}\mathrm{n}{T}_{bR}^6}{A_5+A_6{T}_{bR}^{-1}+A_7\mathrm{l}\mathrm{n}{T}_{bR}+A_8\mathrm{l}\mathrm{n}{T}_{bR}^6}}$$

  (9)
  If $T_{bR}=\left(T_{bR}/T_{cR}\right)<0.8$,

  $$\omega =-7.904+0.1352K_w-0.007456K_w^2+8.359T_{bR}+\left(1.408-0.01063K_w\right)T_{bR}^{-1}$$

  (10)
  If $T_{bR}=\left(T_{bR}/T_{cR}\right)>0.8$,
  where $A_1=-5.92714$, $A_2=6.09648$, $A_3=1.28862$, $A_4=-0.169347$, $A_5=15.2518$, $A_6=15.6875$, $A_7=-13.4721$, $A_8=0.43577$.
• Molecular weight of C₄₂₊ can be estimated from its known mass fraction and the number of moles:

  $${MW}_{C_{42+}}={\displaystyle \frac{w_i\times m}{n}}$$

  (11)

  where $w_i$, m, and n are the sample mass fraction, sample mass, and the number of moles of C₄₂₊, respectively.

3.2. Lumping

The crude oil may contain hundreds and thousands of components. The iteration of phase equilibrium calculations and computing time will dramatically increase if each component is treated individually. Therefore, it is indispensable to lump the SCNs and represent the oil as multiple pseudo-components (PCs). Appropriate lumping can decrease the simulation time by reducing the number of components and making the fluid properties of the grouped system similar to those from the original system at the same time [37]. Considering the influence of both the molecular weight and the mole fraction of each component, the lumping rule, which states that the summation of the mole fraction times the logarithm of the molecular weight $\left(\sum z_i\mathrm{l}\mathrm{n}{M}_i\right)$ and should be the same for each pseudo-component, has been widely used [38].

The mixing rule [39] employed to assign properties to the lumped pseudo-component is:

$${\eta }_j={\displaystyle \frac{{\sum }_{i=l}^u{z}_i{M}_i{\eta }_i}{{\sum }_{i=l}^u{z}_i{M}_i}}$$

(12)

where ${\eta }_j$ can be any property (e.g., $T_b$, $T_c$, and $P_c$) for the jth lumped pseudo-component-consisted SCNs from l to u.

3.3. PR EOS Model

PR EOS

Because of its wide application in the petroleum and chemical industries, the PR EOS model is chosen to quantify the phase behavior of C₃H₈-CO₂-oil systems in this work. It is expressed as:

$$P={\displaystyle \frac{RT}{V-b}}-{\displaystyle \frac{a}{V\left(V+b\right)+b\left(V-b\right)}}$$

(13)

with

$$a=a_c\alpha$$

(14)

$$\alpha ={\left[1+\left(0.37464+1.54226\omega -0.26992{\omega }^2\right)\left(1-\sqrt{T_r}\right)\right]}^2$$

(15)

$$a_c={\displaystyle \frac{0.457235R^2{T}_c^2}{P_c}}$$

(16)

$$b={\displaystyle \frac{0.0777969R{T}_c}{P_c}}$$

(17)

where $T_r$ is the reduced temperature, $\omega$ is the acentric factor, R is the universal gas constant, P is the pressure, V is the molar volume, and T is the temperature.

Despite the high prediction accuracy of the PR EOS, it shows significant deviations in the α function, especially for compounds with high and low $\omega$ values, because a change in saturation pressure can incur an exponential increase error in α [40]. Therefore, many modifications to the α function have been made. Among these modifications, the modified α function proposed by Li and Yang [17] shows great success in describing the phase behavior of both nonhydrocarbon and hydrocarbon compounds:

$$\alpha =exp\left\{\left(0.13280-0.05052\omega +0.25948{\omega }^2\right)\left(1-T_r\right)+0.81769\mathrm{l}\mathrm{n}{\left[1+\left(0.313355+1.86745\omega -0.52604{\omega }^2\right)\left(1-\sqrt{T_r}\right)\right]}^2\right\}$$

(18)

For a mixture system, the following van der Waals mixing rule is used:

$$a={\sum }_{i=1}^{nc}{\sum }_{j=1}^{nc}{x}_i{x}_j\left(1-{\delta }_{ij}\right)\sqrt{a_i{a}_j}$$

(19)

$$b={\sum }_{i=1}^{nc}{x}_j{b}_j$$

(20)

where nc is the number of components in the mixture; $x_i$ and $x_j$ are the mole fraction of the ith and jth component in the mixture; $a$ and $b$ are calculated from Equations (19) and (20) for the ith and jth components respectively; and ${\delta }_{ij}$ is the binary interaction parameter (BIP) between the ith and jth components. The program used in this work is written by using MATLAB R2023a, where the stability algorithm proposed by Michelsen [41] is applied.

The BIPs between the pseudo-components of the Trembley oil are set to zero due to their similar properties, and the BIP for the CO₂-C₃H₈ pair is set as 0.125 (CMG Winprop 2024). The BIPs between pseudo-components and CO₂/C₃H₈ can be calculated from:

$${\delta }_{ij}=1-{\left[{\displaystyle \frac{2\sqrt{V_{ci}^{1/3}{V}_{cj}^{1/3}}}{V_{ci}^{1/3}+V_{cj}^{1/3}}}\right]}^{\beta }$$

(21)

where ${\delta }_{ij}$ is the BIP between the ith and jth component, $V_c$ is the critical molar volume in m³/kmol, and $\beta$ is an adjustable parameter.

In this work, $\beta$ has been tuned to find the optimal BIP to match the experimental data. Based on the optimized BIPs, we propose the following two new BIP correlations for C₃H₈-pseudo-components and CO₂-pseudo-components pairs, respectively.

$${\delta }_{{\mathrm{C}}_3{\mathrm{H}}_8-{\mathrm{P}\mathrm{C}}_{\mathrm{i}}}={\displaystyle \frac{0.0583T}{T_{C_i}}}-0.0109{\omega }_i+0.3422{\gamma }_i-0.2523$$

(22)

$${\delta }_{{\mathrm{C}\mathrm{O}}_2-{\mathrm{P}\mathrm{C}}_{\mathrm{i}}}={\displaystyle \frac{0.0096T}{T_{C_i}}}-0.0198{\omega }_i+0.0021{\gamma }_i-0.0145$$

(23)

where ${\delta }_{{\mathrm{C}}_3{\mathrm{H}}_8-{\mathrm{P}\mathrm{C}}_{\mathrm{i}}}$ is the BIP between C₃H₈, and PC_i, PC_i is the ith pseudo-component. $T_{C_i}$, ${\omega }_i$ and ${\gamma }_i$ are the critical temperature in °C, acentric factor, and specific gravity of ith pseudo-component. The standard errors of the coefficients in Equation (22) are 0.0272, 0.0137, 0.0413, and 0.0314, respectively. ${\delta }_{{\mathrm{C}\mathrm{O}}_2-{\mathrm{P}\mathrm{C}}_{\mathrm{i}}}$ is the BIP between CO₂ and PC_i. The standard errors of the coefficients in Equation (23) are 0.0119, 0.0181, 0.0061, and 0.0137. Note that the proposed correlations (i.e., Equations (22) and (23)) are valid when three or more pseudo-components’ lumping schemes are applied.

According to the measured and calculated saturation pressures, the absolute average relative deviation (AARD) is used in this study to evaluate the prediction accuracy,

$$\mathrm{A}\mathrm{A}\mathrm{R}\mathrm{D}={\displaystyle \frac{1}{n}}{\sum }_{i=1}^n\left|{\displaystyle \frac{P_i^{cal}-P_i^{exp}}{P_i^{exp}}}\right|$$

(24)

where n is the number of data points, $P^{cal}$ is the calculated saturation pressure, and $P^{exp}$ is the measured saturation pressure.

3.4. MMP

The minimum miscibility pressure (MMP) is a significant parameter during gas-assisted EOR processes. The maximum oil recovery can be achieved when the displacement pressure is higher than MMP. As aforementioned, although it may be difficult to achieve CO₂-oil miscibility at low reservoir pressures, adding a certain amount of C₃H₈ in the CO₂ stream can lower the MMP between gas and oil so that the miscibility may be reached at lower reservoir pressures. In this work, the MMP is calculated by using the tie line method [42,43,44], which is adopted by CMG:

• Start with initial equations at the initial pressure and given temperature:

  $$\left(1-{\sigma }_1\right)x_i^1+{\sigma }_1{y}_i^1=z_i^{oil};i=1,2,\cdots ,N$$

  (25)

  $$\left(1-{\sigma }_{N-1}\right)x_i^{N-1}+{\sigma }_{N-1}{y}_i^{N-1}=z_i^{inj};i=1,2,\cdots ,N$$

  (26)

  where $z_i^{oil}$ is the mole fraction of component i in the original reservoir oil, $z_i^{inj}$ is the mole fraction of component i in injection gas, N is the number of components, $x_i$ and $y_i$ are the mole fractions of component i in the equilibrium liquid phase and gas phase, respectively, and $\sigma$ is a parameter that can be anywhere in the interval from $-\infty$ to $+\infty$. In this work, the first guesses of x_i and y_i are the mole fraction of C₃H₈ or CO₂ or their mixture and Trembley oil;
• To obtain a set of intersecting tie lines, the following equations are applied:

  $$\left(1-{\sigma }_j\right)x_i^j+{\sigma }_j{y}_i^j=\left(1-{\sigma }_{j+1}\right)x_i^{j+1}+{\sigma }_{j+1}{y}_i^{j+1};i=1,2,\cdots ,N,j=1,2,\cdots N-1$$

  (27)

  $$x_i^j{\phi }_i^{L,j}=y_i^j{\phi }_i^{V,j};i=1,2,\cdots ,N,j=1,2,\cdots N-1$$

  (28)

  $${\sum }_{i=1}^N\left(y_i^j-x_i^j\right)=0;j=1,2,\cdots N-1$$

  (29)

  where ${\phi }_i$ is the fugacity coefficient of component i in the gas (V) or liquid (L) phase;
• Increase the pressure slightly. Equations (28) and (29) are solved again using the solution for the previous pressure;
• Repeat Step #c until one of the tie lines shrinks to a point at which condition is fulfilled for the nth tie line if

  $$L_n=\sqrt{{\sum }_{i=1}^N{\left(y_i^n-x_i^n\right)}^2}\approx 0$$

  (30)

  where L_n is the length of one of the tie lines between the injected fluid and oil. The pressure where this happens is the MMP. When the length is zero, the tie line connects two identical critical compositions. Finally, the MMP can be calculated by using the PR EOS with the determined gas fraction (x_i) and liquid fraction (y_i).

4. Results and Discussions

4.1. PVT Equipment Validation

Table 3. Vapor pressure of C₃H₈ with two depressurization methods.

Temperature (°C)	Reference Value (psi)	Continuous Method		Stepwise Method
		Measured (psi)	Error (%)	Measured (psi)	Error (%)
30	158.4	163.2	3.03	161.6	2.02
40	190.3	195.4	2.67	192.7	1.26
50	226.8	232.6	2.56	230.1	1.45

summarizes the measured vapor pressures of pure substance C₃H₈ by using both continuous and stepwise depressurization methods, respectively. The well-accepted referential value from CMG is used as the true vapor pressure of C₃H₈, which are 158.4, 190.3, and 226.8 psi at 30, 40, and 50 °C, respectively. As can be seen, the vapor pressures measured by using both continuous and stepwise methods are close to the referential value. It indicates that the PVT equipment is valid for performing the experiments in this work. The errors of the continuous depressurization method are 3.03%, 2.67%, and 2.56%, respectively. Meanwhile, the stepwise depressurization method yields errors of 2.02%, 1.26%, and 1.45%, respectively. In terms of accuracy, as we expected, the time-consuming stepwise method can provide a more accurate measurement, since there is sufficient time for the mixture to reach equilibrium at each pressure stage. However, the errors of the continuous method are slightly higher than that of the stepwise method. Compared with 48 h for the stepwise method, it only takes 3–4 h to measure a vapor pressure point using a continuous method. Therefore, the continuous depressurization method is applied in this work.

4.2. Characterization of Trembley Oil

For the physical and critical properties, the constant Watson factor $K_w$ is calculated to be 11.824 by using Equation (3). The Trembley oil is characterized into one to six pseudo-component(s). The physical and critical properties of each pseudo-component in different lumping schemes are tabulated in

Table 4. Physical and critical properties of each pseudo-component in different lumping schemes.

No. of PCs *	MW, g/mol	PC, atm	TC, K	γ	Tb, K	ω
1	234.667	17.008	753.715	0.863	573.479	0.114
2	150.212	23.711	641.937	0.794	459.751	0.212
	490.845	14.631	869.390	0.954	687.960	0.636
3	129.247	25.914	610.064	0.774	427.981	0.176
	213.037	18.350	727.432	0.846	546.376	0.312
	580.628	13.101	910.548	0.982	728.647	0.739
4	116.781	27.578	588.588	0.760	407.150	0.153
	163.012	21.979	664.193	0.807	481.339	0.236
	231.871	17.262	748.773	0.860	568.598	0.337
	604.239	12.911	919.991	0.990	737.828	0.766
5	116.782	27.579	588.589	0.761	407.517	0.153
	163.013	21.980	664.193	0.808	481.773	0.237
	219.448	17.853	735.917	0.852	555.609	0.322
	296.820	14.796	808.874	0.899	632.341	0.408
	745.441	12.270	966.642	1.030	782.285	0.933
6	107.672	28.999	571.533	0.750	390.998	0.136
	134.178	25.045	619.963	0.780	437.200	0.186
	171.067	21.184	676.108	0.814	493.328	0.250
	219.447	17.852	735.916	0.852	555.108	0.322
	290.818	14.952	804.197	0.896	626.856	0.403
	718.796	12.343	958.868	1.022	774.496	0.901

* Note: PC is the abbreviation of pseudo-component.

. It can be seen that the physical properties of each pseudo-component change with the different lumping schemes.

For the optimum lumping scheme, even though the oil can be lumped as either a single or as multiple pseudo-component(s), as listed in Table 4, an optimum lumping scheme is preferred for determining Trembley oil. The BIPs are tuned to best match the bubble point pressures of Feeds #1–12 by applying each of the six lumping schemes. The optimized exponent β in the BIP formula (see Equation (21)) can be found in

Table 5. Optimum BIP exponent (β) in each of six lumping schemes.

Binary and Ternary Systems	Feed No.	No. of PCs
		Temperature (°C)	1	2	3	4	5	6
C3H8-oil	1	25	1.99	1.21	1.32	1.47	1.22	1.31
	2	30	2.21	1.30	1.48	1.52	1.25	1.34
	3	40	2.44	1.54	1.55	1.57	1.29	1.39
	4	50	2.68	1.41	1.54	1.63	1.33	1.44
CO2-oil	5	25	0.19	0.27	0.04	0.05	0.03	0.05
	6	30	0.09	0.00	0.01	0.00	0.00	0.00
	7	40	0.92	0.06	0.10	0.11	0.08	0.12
	8	50	0.38	0.00	0.01	0.02	0.00	0.03
C3H8-CO2-oil (β for C3H8-PC)	9	25	1.60	1.36	1.36	1.36	1.22	1.36
	10	30	1.69	1.30	1.40	1.42	1.28	1.39
	11	40	2.68	1.36	1.36	1.36	1.29	1.34
	12	50	1.72	1.36	1.37	1.36	1.33	1.20
C3H8-CO2-oil (β for CO2-PC)	9	25	0.22	0.01	0.01	0.01	0.03	0.01
	10	30	0.31	0.01	0.02	0.04	0.03	0.06
	11	40	0.50	0.01	0.01	0.01	0.08	0.01
	12	50	0.34	0.01	0.02	0.01	0.00	0.01

. Subsequently, the AARDs in predicting saturation pressures of all feeds are quantified by using the optimized BIPs. Thus, Figure 3 plots the AARDs as a function of the number of pseudo-components. It is found that the AARDs decrease with the number of pseudo-components for all three systems. Although the minimum AARD is achieved when the six pseudo-components lumping scheme is employed, three pseudo-components are finally selected to perform further calculations in this work because the AARD is only improved slightly, but the computing time increases greatly when more than three pseudo-components are applied. With the lumping scheme of three pseudo-components, the overall AARDs of the C₃H₈-oil, CO₂-oil, and C₃H₈-CO₂-oil systems are 2.88%, 1.34%, and 3.28%, respectively.

4.3. Validation of the Developed BIP Correlations

For BIP optimization, for each lumping scheme, the exponent β in Equation (21) is tuned for C₃H₈-oi, CO₂-oil, and C₃H₈-CO₂-oil mixture, respectively. As for the ternary systems of Feeds #9–12, the BIP between C₃H₈ and the pseudo-component and the BIP between CO₂ and the pseudo-component are tuned separately. As a result, the optimal exponents in each lumping scheme are listed in Table 5. It shows that the BIP is affected by both the temperature and the lumping scheme.

For the validation of the BIP correlations, with the assistance of the optimum exponents in Table 5, two new BIP correlations, i.e., Equations (22) and (23), are, respectively, developed for the C₃H₈-PC pair and CO₂-PC pair when the oil is characterized as three or more PCs. Figure 3 exhibits the overall AARD for C₃H₈-oil, CO₂-oil, and C₃H₈-CO₂-oil systems as a function of the number of pseudo-components. As shown, the largest value of AARD is when applying four pseudo-component lumping schemes for C₃H₈-oil systems, which is about 4.4%. The rest of the AARDs are all less than 4.0%, implying that these two new correlations are capable of calculating BIP in the phase behavior evaluation. It is noteworthy that these two correlations are applicable for three or more pseudo-component lumping schemes, whereas they are not suitable for single or two pseudo-component(s) lumping schemes, which have an overall AARD greater than 30%. In general, Equations (22) and (23) are able to estimate the BIP between CO₂-oil and C₃H₈-oil for more than three pseudo-components with acceptable AARDs.

4.4. Saturation Pressure and Solubility

As aforementioned, the intersection of two trend linear lines on a P-V diagram represents the bubble point. Figure 4 plots the measured pressure versus volume for Feed #4 (72.21 mol% C₃H₈, 27.79 mol% oil) at 50 °C, indicating the bubble point is 199.9 psi. It is noted that the abscissa of the P-V diagram is the withdrawal volume of hydraulic oil. Figure 4 also indicates that the solubility of C₃H₈ gas in Trembley oil is 60.08 mol% at 199.9 psi and 25 °C. Figure 5, Figure 6 and Figure 7 exhibit the measured P-V diagrams for 12 feeds at 25, 30, 40, and 50 °C, respectively, where both the saturation pressures and solubilities of gas and a gas mixture can be quantified.

For the C₃H₈-oil binary system, Figure 8 exhibits the measured and calculated saturation pressures for Feeds #1–4 over the temperature range of 25–50 °C. It can be seen that the saturation pressure increases with the temperature. When the C₃H₈ solubility is 60.08 mol% (Feed #3), for instance, the saturation pressure is raised from 106 psi to 177 psi if the temperature is increased from 25 °C to 50 °C. It indicates that temperature is an important impact factor for saturation pressure. Also, the solubility of C₃H₈ in Trembley oil can be significantly improved by raising pressure. For example, at a given temperature of 50 °C, the C₃H₈ solubility changes from 34.62 mol% (Feed #1) to 72.21 mol% (Feed #4) when the pressure increases from 117 psi to 200 psi. In addition, the saturation pressure of the C₃H₈-oil system increases with the amount of C₃H₈ for a given temperature. For instance, when the temperature is 25 °C, the saturation pressure for Feed #1 (34.62 mol%) is around 65 psi, while Feed #3 (60.08%) is 118 psi. Overall, the calculated saturation pressures, by using the developed PR EOS model, can reach a good agreement with the measured one, yielding an overall AARD of 2.30% for the 16 data points in Figure 8.

For the CO₂-oil binary system, when compared with the C₃H₈-oil mixture, much higher saturation pressures can be found for the CO₂-oil mixture as can be seen in Figure 9. Although Feed #3 and Feed #7 contain similar gas mole fractions in the mixtures (60.08 mol% C₃H₈ in Feed #3 and 60.39 mol% CO₂ in Feed #7), the saturation pressures are measured to be 177 psi and 670 psi at 50 °C for Feed #3 and Feed #7, respectively. It implies that high pressure is required during CO₂-assistant EOR processes to reach the maximum swelling effect. At a given temperature of 50 °C, the CO₂ solubility is 29.29 mol% at 290 psi (see Feed #5), whereas 73.74 mol% of solubility can be reached when the pressure is up to 920 psi (see Feed #8). Similar to the C₃H₈-oil system, the saturation pressure increases with the temperature. Taking Feed #8 as an example, an approximate 57% increase in saturation pressure (from 584 to 920 psi) is obtained when the temperature changes from 25 °C to 50 °C. Similar to the C₃H₈-oil system, the saturation pressure increases with the amount of CO₂ in oil for a given temperature. Additionally, the developed BIP correlation (i.e., Equation (23)) incorporated in the PR EOS exhibits excellent performance in the prediction of the saturation pressure, with an overall AARD of 1.62% for the 16 data points in Figure 9.

For the C₃H₈-CO₂-oil ternary system, when compared with CO₂-oil binary systems, the saturation pressures of the ternary system are greatly reduced by introducing C₃H₈ in the system, as presented in Figure 10. Feed #12 consists of 16.16 mol% C₃H₈ and 57.57 mol% CO₂, resulting in a total of 73.73 mol% CO₂-C₃H₈ gas mixture. The saturation pressure of Feed #12 is measured to be 505 psi, which is much lower than the 790 psi of Feed #8, which includes 73.74 mol% pure CO₂ gas at the same temperature of 40 °C. Such a finding indicates that adding C₃H₈ into the CO₂ stream can increase the gas solubility in oil and decrease the saturation pressure of the reservoir fluids. It also can be seen that temperature is an important impact factor for the ternary system as well. The saturation pressure increases from 350 psi to 500 psi when the temperature is elevated from 25 °C to 50 °C for Feed #9. Moreover, the pressure has a great impact on gas solubility for Feeds #9–12 at a given temperature. For instance, the solubility of the gas mixture increases from 59.07 mol% (Feed #9) to 73.73 mol% (Feed #12) with the pressure changing from 435 psi to 648 psi at 40 °C. It is noteworthy that the measured saturation pressure for Feed #12 at 50 °C is abnormal, which may be caused by the non-equilibrium state of the system. However, the overall AARD for the four feeds in Figure 10 is 3.26%, yielding an acceptable prediction result. Finally, the overall AARD for all 12 feeds is found to be 2.39%, which illustrates the capability of two newly developed BIP correlations.

In addition, Figure 11 plots the measured data in terms of the solubility of CO₂, C₃H₈, and their mixture as a function of saturation pressure at 25 °C. The dashed lines representing the regressions of solubility show an excellent agreement with the measured data points, with R² of 0.9855, 0.9352, and 0.9986 for C₃H₈, the C₃H₈-CO₂ mixture, and CO₂, respectively. It can be seen that the gas solubility increases with pressure at a given temperature for all three systems. Furthermore, C₃H₈ has the highest solubility, while CO₂ has the lowest one at the same pressure and temperature. For example, at 400 psi, the solubility of CO₂ is 65 mol%, and C₃H₈ has a solubility far greater than 100 mol% according to the regression equation, indicating the MMP between C₃H₈ and Trembley oil might have already been reached at this given pressure of 400 psi. The solubility of their mixture, however, exhibits a value of 75 mol%, higher than that of CO₂ (65 mol%) at the same pressure. Such results reconfirm that C₃H₈ has a higher solubility than CO₂ in Trembley oil. Therefore, adding a certain amount of C₃H₈ in CO₂ can improve the CO₂ EOR performance at an acceptable cost.

4.5. Viscosity of the Gas-Saturated Oil at High Pressures

After the saturation pressure measurements, the viscosities of Feeds #1, 8, and 12 have been measured above their saturation pressures using an in-line viscometer at 25 °C. The viscosity results are listed in

Table 6. Viscosity of mixtures at 25 °C and high pressure.

System	μ (cP)
C3H8-oil (Feed #1)	2.33
CO2-oil (Feed #8)	3.03
C3H8-CO2-oil (Feed #12)	1.89
Dead oil at 1600 psi	9.50
Dead oil at room pressure	9.49

. In this study, a pressure of 1600 psi is selected to measure the viscosities, since all three mixtures are in a single liquid phase at such a high pressure. The viscosity of dead oil at 1600 psi and room pressure are 9.50 cP and 9.49 cP, respectively. The viscosities of gas-saturation oil mixtures are measured to be 2.33 cP (Feed #1), 3.03 cP (Feed #8), and 1.89 cP (Feed #12) at 1600 psi, respectively. Compared with the Trembley oil without solution gas, it is found that the viscosities of gas-saturation mixtures are reduced by adding C₃H₈ or CO₂ or their mixture.

4.6. Miscibility between Gas and Oil

The MMPs are also computed using the tie-line method in CMG with three pseudo-components lumping scheme. Figure 12 exhibits the average simulated MMPs under temperature ranges from 25 °C to 50 °C. The y-axis is the average simulated MMP of C₃H₈-oil for Feeds #1–4 (average C₃H₈ of 53.36 mol%), CO₂-oil for Feeds #5–8 (average CO₂ of 52.69 mol%), and C₃H₈-CO₂-oil for Feeds #9–12 (average C₃H₈ of 16.02 mol% and CO₂ of 48.96 mol%). Generally, C₃H₈-oil systems indicate the smallest MMPs, while CO₂-oil systems show the largest MMPs. The average MMPs for C₃H₈-oil systems are predicted from 150 psi to 250 psi at a temperature range of 25–50 °C, whereas the MMPs for CO₂-oil systems can be as high as 950–1500 psi. The presence of C₃H₈ (average C₃H₈ of 16.02 mol%) in the CO₂-oil system can lower the MMP from 1490 psi to 1160 psi at 50 °C. It means that it is easier to achieve the miscible condition with the assistance of C₃H₈ during CO₂ EOR processes. Hence, it is feasible to implement the miscible gas EOR, even in shallow reservoirs with low reservoir pressures.

When C₃H₈ is added to a CO₂-oil mixture, it lowers the MMP by altering the structure and interactions of the crude-oil components. C₃H₈ is highly soluble in oil and mixes well with hydrocarbons, especially lighter fractions, which reduces phase segregation and improves the uniformity of the oil phase. This enhanced mixing allows CO₂ to dissolve more easily into the oil, aided by the reduction in interfacial tension between CO₂ and the oil. Additionally, C₃H₈ causes the oil to swell, increasing its volume and providing more space for CO₂ to diffuse, further facilitating miscibility at lower pressures. C₃H₈ also modifies the critical properties and phase behavior of the CO₂-oil system, making the oil phase less dense and altering the phase boundaries, which shifts the conditions needed for miscibility. These changes lead to a lower MMP because the fluid system can achieve a single miscible phase more easily, making propane a valuable additive for EOR processes using CO₂ injection.

5. Conclusions

By systematically investigating the effects of C₃H₈ on CO₂-saturated crude oil, this study provides crucial insights into optimizing CO₂ EOR processes, demonstrating the practical benefits of using C₃H₈ to enhance recovery efficiency under various reservoir conditions. The important highlights of this work are summarized by the following key points.

(1)

Under reservoir conditions, the saturation pressures of the C₃H₈-oil mixture, CO₂-oil mixture, and C₃H₈-CO₂-oil mixture have been experimentally and theoretically determined. It has been found that the saturation pressure increases with the temperature, as well as with the amount of C₃H₈, CO₂, or CO₂-C₃H₈ gas mixture in the oil;

(2)

The saturation pressures are successfully predicted by using the well-developed PR EOS model, together with two newly developed BIP correlations with an overall AARD of 2.39%, where three pseudo-component lumping schemes are applied;

(3)

The correlations of Trembley oil’s viscosity and density have been successfully generated for temperatures ranging from 20 °C to 75 °C. The measured viscosity and saturation pressure data confirm that the presence of C₃H₈ in the CO₂ stream can enhance oil-recovery performance by achieving miscibility at lower pressures and delaying gas-breakthrough time, ultimately leading to more CO₂ being stored in the reservoir;

(4)

The enhancement in miscibility through C₃H₈ addition can effectively delay gas breakthrough and increase incremental oil recovery compared to immiscible or near-miscible CO₂ flooding, thus minimizing CO₂ breakthrough and improving the overall EOR efficiency.