Experimental Investigation on Microscopic Residual Oil Distribution During CO 2 Huff-and-Puff Process in Tight Oil Reservoirs

The determination of microscopic residual oil distribution is beneficial for exploiting reservoirs to their maximum potential. In order to investigate microscopic residual oil during the carbon dioxide (CO2) huff-and-puff process in tight oil reservoirs, several CO2 huff-and-puff tests with tight sandstone cores were conducted at various conditions. Then, nuclear magnetic resonance (NMR) was used to determine the microscopic residual oil distribution of the cores. The experiments showed that the oil recovery factor increased from 27.22% to 52.56% when injection pressure increased from 5 MPa to 13 MPa. The oil recovery was unable to be substantially enhanced as the injection pressure further increased beyond the minimum miscible pressure. The lower limit of pore distribution where the oil was recoverable corresponded to relaxation times of 2.68 ms, 1.29 ms, and 0.74 ms at an injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively. Longer soaking time also increased the lower limit of the oil-recoverable pore distribution. However, more cycles had no obvious effect on expanding the interval of oil-recoverable pore distribution. Therefore, higher injection pressure and longer soaking time convert the residual oil in smaller and blind pores into recoverable oil. This investigation provides some technical ideas for oilfields in design development programs for optimizing the production parameters during the CO2 huff-and-puff process.


Introduction
The contribution of tight oil is becoming increasingly important for maximizing oil production around the world.[1][2][3][4].However, due to poor water injectivity, conventional water flooding is unsuitable after natural depletion of tight oil reservoirs.It has been proven that carbon dioxide (CO 2 ) injection is a reliable method to enhance oil recovery in tight oil reservoirs [5,6].The CO 2 injection technique can be generally categorized into CO 2 huff-and-puff and CO 2 flooding processes.In tight oil reservoirs, due to ultra-low permeability, it is difficult for the gas to drive oil from the injection well to the production well [7].However, tight oil reservoirs often have natural fractures or hydraulic fractures, which result in early break-through during the CO 2 flooding process, resulting in very low sweep efficiency [8,9].
The CO 2 huff-and-puff process is an effective single-well enhanced oil recovery (EOR) operation in tight oil reservoirs developed through laboratory experiments [10] and now used in field application [11].The main recovery mechanisms during the CO 2 huff-and-puff process include

Materials
In the experiment, the core samples were tight cores collected from Changqing Oilfield, China.It is noted that core samples with nearly the same gas permeability and pore distribution were selected from the same formation.The properties of cores are listed in Table 1.The stock tank oil (STO) sample was collected from the Changqing Oilfield, China.The density, molecular weight (MWoil), and viscosity of the cleaned STO was measured as 833.4 kg/m 3 , 229.7 g/mol and 4.80 mPa•s, respectively, at atmospheric pressure and 61 °C .The gas chromatography (GC) compositional analysis of the cleaned crude oil sample at 21 °C is shown in Figure 2. The brine was considered to be a calcium chloride water type of, which had the total dissolved solids (TDS) of 30,917.8mg/L.The brine viscosity was measured as 0.4 mPa•s at atmospheric pressure and 61 °C .The purity of the CO2 used in this study was 99.99%, and was supplied by Beijing Huayuan Gas Chemical Co. Ltd. (Beijing, China).

Materials
In the experiment, the core samples were tight cores collected from Changqing Oilfield, China.It is noted that core samples with nearly the same gas permeability and pore distribution were selected from the same formation.The properties of cores are listed in Table 1.The stock tank oil (STO) sample was collected from the Changqing Oilfield, China.The density, molecular weight (MW oil ), and viscosity of the cleaned STO was measured as 833.4 kg/m 3 , 229.7 g/mol and 4.80 mPa•s, respectively, at atmospheric pressure and 61 • C. The gas chromatography (GC) compositional analysis of the cleaned crude oil sample at 21 • C is shown in Figure 2.

Materials
In the experiment, the core samples were tight cores collected from Changqing Oilfield, China.It is noted that core samples with nearly the same gas permeability and pore distribution were selected from the same formation.The properties of cores are listed in Table 1.The stock tank oil (STO) sample was collected from the Changqing Oilfield, China.The density, molecular weight (MWoil), and viscosity of the cleaned STO was measured as 833.4 kg/m 3 , 229.7 g/mol and 4.80 mPa•s, respectively, at atmospheric pressure and 61 °C .The gas chromatography (GC) compositional analysis of the cleaned crude oil sample at 21 °C is shown in Figure 2. The brine was considered to be a calcium chloride water type of, which had the total dissolved solids (TDS) of 30,917.8mg/L.The brine viscosity was measured as 0.4 mPa•s at atmospheric pressure and 61 °C .The purity of the CO2 used in this study was 99.99%, and was supplied by Beijing Huayuan Gas Chemical Co. Ltd. (Beijing, China).The brine was considered to be a calcium chloride water type of, which had the total dissolved solids (TDS) of 30,917.8mg/L.The brine viscosity was measured as 0.4 mPa•s at atmospheric pressure and 61 • C. The purity of the CO 2 used in this study was 99.99%, and was supplied by Beijing Huayuan Gas Chemical Co. Ltd. (Beijing, China).

CO 2 Solubility and Oil Swelling Factor Measurements
Figure 3 shows the schematic diagram of the experimental apparatus used for CO 2 solubility and oil swelling factor measurements at various equilibrium pressures.The apparatus includes a high-pressure visual cell (HD-KS-02, Huada, Haian, China), a magnetic stirrer (85-2, Sile, Shanghai, China), high-pressure cylinders and a constant flow pump (260D, ISCO, Lincoln, NE, USA).A pressure transducer (HKY-06C, HKY, Beijing, China) was used to measure pressure inside the cell.The inner volume of the high-pressure cell was 80.00 cm 3 and the magnetic stirrer was used to accelerate the CO 2 dissolution into the oil.The crude oil sample was injected into the visual cell to a specific volume of V o,i = 30.00cm 3 .
Energies 2018, 11, x FOR PEER REVIEW 4 of 16 Figure 3 shows the schematic diagram of the experimental apparatus used for CO2 solubility and oil swelling factor measurements at various equilibrium pressures.The apparatus includes a highpressure visual cell (HD-KS-02, Huada, Haian, China), a magnetic stirrer (85-2, Sile, Shanghai, China), high-pressure cylinders and a constant flow pump (260D, ISCO, Lincoln, NE, USA).A pressure transducer (HKY-06C, HKY, Beijing, China) was used to measure pressure inside the cell.The inner volume of the high-pressure cell was 80.00 cm 3 and the magnetic stirrer was used to accelerate the CO2 dissolution into the oil.The crude oil sample was injected into the visual cell to a specific volume of Vo,i = 30.00cm 3 .Then, the cell was pressurized with CO2 to a prespecified pressure   .When the pressure inside the cell reached a stable value (  ) and did not change, the CO2-oil system was considered to have reached equilibrium.The height of the oil sample inside the visual cell was measured using a cathetometer.In this study, the CO2 solubility ( CO 2 ) is defined as the total moles of dissolved CO2 in 1 L of the original crude oil sample.Mass balance equations were used to calculate the CO2 solubility, which is shown follows: where   2 represents the moles of CO2 dissolved in the oil (mol);   2 , and   2 , are the moles of the CO2 of the gas phase at initial and final condition (mol), respectively;   and   are the initial pressure and final pressure of the system (MPa), respectively;   and   represent the initial and final volume of the CO2 of gas phase (cm 3 ), respectively;  is the universal gas constant (8.314J/(mol•k));  is the experimental temperature (K); and   and   are the gas compressibility factor at initial and final conditions, respectively.
where   2 is the solubility of CO2 in the oil (mol/L) and  , represents initial volume of the oil (cm 3 ).Then, the oil swelling factor (SF) was determined by the ratio of the final volume of the oil to its initial value at the beginning of the experiment: where SF is the oil swelling factor and  , represents the final volume of the oil (cm 3 ).Then, the cell was pressurized with CO 2 to a prespecified pressure P i .When the pressure inside the cell reached a stable value (P f ) and did not change, the CO 2 -oil system was considered to have reached equilibrium.The height of the oil sample inside the visual cell was measured using a cathetometer.In this study, the CO 2 solubility (χ CO 2 ) is defined as the total moles of dissolved CO 2 in 1 L of the original crude oil sample.Mass balance equations were used to calculate the CO 2 solubility, which is shown follows: where n CO 2 represents the moles of CO 2 dissolved in the oil (mol); n CO 2 ,i and n CO 2 , f are the moles of the CO 2 of the gas phase at initial and final condition (mol), respectively; P i and P f are the initial pressure and final pressure of the system (MPa), respectively; V i and V f represent the initial and final volume of the CO 2 of gas phase (cm 3 ), respectively; R is the universal gas constant (8.314J/(mol•k)); T is the experimental temperature (K); and Z i and Z f are the gas compressibility factor at initial and final conditions, respectively.
where χ CO 2 is the solubility of CO 2 in the oil (mol/L) and V o,i represents initial volume of the oil (cm 3 ).
Energies 2018, 11, 2843 5 of 16 Then, the oil swelling factor (SF) was determined by the ratio of the final volume of the oil to its initial value at the beginning of the experiment: where SF is the oil swelling factor and V o,i represents the final volume of the oil (cm 3 ).

CO 2 -oil IFT Measurements
A schematic diagram of the setup to measure the dynamic IFT between CO 2 and the crude oil is shown in shown in Figure 4.The core device of the set-up was the high-pressure IFT cell (IFT-10, Temco, Fremont, CA, USA) with maximum operating pressure and temperature of 69 MPa and 177 • C, respectively, and the designed volume of the IFT cell was 49.5 cm 3 .A pump (260D, ISCO, Lincoln, NE, USA) was used for injecting CO 2 into the optical cell and another syringe pump was used for injecting the crude oil into a stainless syringe needle.A microscopic camera was used to capture the image of the pendant drop and computer installed software (FTA, First Ten Angstroms Portsmouth, VA, USA) based on axisymmetric drop shape analysis (ADSA) was utilized to analyze the digital oil drop image.A densitometer (DMA512P, Anton Paar, Graz, Austria) was used to measure the densities of the oil during different tests.

CO2-oil IFT Measurements
A schematic diagram of the setup to measure the dynamic IFT between CO2 and the crude oil is shown in shown in Figure 4.The core device of the set-up was the high-pressure IFT cell (IFT-10, Temco, Fremont, CA, USA) with maximum operating pressure and temperature of 69 MPa and 177 °C , respectively, and the designed volume of the IFT cell was 49.5 cm 3 .A pump (260D, ISCO, Lincoln, NE, USA) was used for injecting CO2 into the optical cell and another syringe pump was used for injecting the crude oil into a stainless syringe needle.A microscopic camera was used to capture the image of the pendant drop and computer installed software (FTA, First Ten Angstroms Portsmouth, VA, USA) based on axisymmetric drop shape analysis (ADSA) was utilized to analyze the digital oil drop image.A densitometer (DMA512P, Anton Paar, Graz, Austria) was used to measure the densities of the oil during different tests.After cleaning the experimental setup and heating the setup to a reservoir temperature of 61 °C , CO2 was pumped into the optical cell to the designated pressure.A pendant drop of the oil was introduced and formed.A series of digital images of the pendant drop were captured at different times.The pendant drop remained two minutes at the tip of the syringe needle before the drop dropped off when equilibrium was achieved.The shapes of the pendant drop were analyzed by the ADSA technique to measure the dynamic IFT between the pendant oil drop and the CO2 phase.The IFT tests at different designed pressures were repeated three times and the measurement error between different tests should be less than 0.5 mJ/m 2 .

CO2 Huff-and-Puff Tests
The CO2 huff-and-puff tests were carried out with the apparatus shown in Figure 5.A constant flow pump (260D, ISCO, Lincoln, NE, USA) was applied to displace the crude oil, brine, and CO2 through the core plug inside a high-pressure stainless steel coreholder (TY-4, Huada, Haian, China) After cleaning the experimental setup and heating the setup to a reservoir temperature of 61 • C, CO 2 was pumped into the optical cell to the designated pressure.A pendant drop of the oil was introduced and formed.A series of digital images of the pendant drop were captured at different times.The pendant drop remained two minutes at the tip of the syringe needle before the drop dropped off when equilibrium was achieved.The shapes of the pendant drop were analyzed by the ADSA technique to measure the dynamic IFT between the pendant oil drop and the CO 2 phase.The IFT tests at different designed pressures were repeated three times and the measurement error between different tests should be less than 0.5 mJ/m 2 .

CO 2 Huff-and-Puff Tests
The CO 2 huff-and-puff tests were carried out with the apparatus shown in Figure 5.A constant flow pump (260D, ISCO, Lincoln, NE, USA) was applied to displace the crude oil, brine, and CO 2 through the core plug inside a high-pressure stainless steel coreholder (TY-4, Huada, Haian, China) with an inner and outer diameter of 25 mm and 40 mm, respectively.Three high-pressure cylinders were applied to store and deliver crude oil, brine and CO 2 , respectively.The confining pressure on the core plug was always maintained 2-3 MPa higher than the injection pressure with another ISCO syringe pump.An air bath that was heated by two electronic heat guns combined with a temperature controller was used to keep the system at the reservoir temperature of 61 • C. A back-pressure regulator (HDB-03, Huada, Haian, China) was used to target the desired production pressure during the coreflood test.The produced oil was collected in a burette and the volume of the produced gas was measured by a gas flow meter.The general procedure for the CO2 coreflood tests is briefly described as follows.
(1) Prior to each test, the core plugs were thoroughly cleaned using a Dean-Stark extractor (SXT-02, Shanghai Pingxuan Scientific Instrument Co., Ltd., Shanghai, China) for 20−30 days.After the core plugs were cleaned and dried at 100 °C .The gas permeability and porosity were measured with nitrogen (High-Pressure Gas Permeameter/Porosimeter, Temco, Tulsa, OK, USA).( 2) The core plug was placed in a high-pressure coreholder (TY-4, Huada, Haian, China) and vacuumed for 24 hours.Then, the formation brine was injected at a flow rate of 0.2 cm 3 /min to saturate the core plug.After that, the NMR apparatus (Mini-NMR, Niumag, Suzhou, China) was used to measure the transverse relaxation time T2 of the core sample under the initial watersaturated condition.The magnetic intensity, gradient value control precision, and frequency range of the NMR apparatus were 0.5 T, 0.025 T/m, 0.01 MHz, and 1-30 MHz, respectively.(3) The core was displaced with MnCl2 solution (15,000 mg/L) of 5 PV.Then, the saturated core was scanned again by the NMR apparatus to ensure the hydrogen signal of the brine was eliminated.(4) After that, 3.0 PV of the crude oil was pumped through the core plugs at a constant rate of 0.1 cm 3 /min until no water was produced to achieve the connate water saturation (Swc) and the initial oil saturation (Soi) at a reservoir temperature of 61 °C .The T2 spectrum was measured again after the core had been saturated with crude oil.( 5) In each test, the pressure of the CO2 in the high-pressure cylinder was increased to the prespecified injection pressure at the temperature of 61 °C .Then, the CO2 was injected into the oil saturated core at constant pressure to the desired pressure and the pump maintained the constant pressure condition for 30 min.After CO2 injection, CO2 was allowed to soak for tsoak = 6 The general procedure for the CO 2 coreflood tests is briefly described as follows.
(1) Prior to each test, the core plugs were thoroughly cleaned using a Dean-Stark extractor (SXT-02, Shanghai Pingxuan Scientific Instrument Co., Ltd., Shanghai, China) for 20-30 days.After the core plugs were cleaned and dried at 100 • C. The gas permeability and porosity were measured with nitrogen (High-Pressure Gas Permeameter/Porosimeter, Temco, Tulsa, OK, USA).(2) The core plug was placed in a high-pressure coreholder (TY-4, Huada, Haian, China) and vacuumed for 24 hours.Then, the formation brine was injected at a flow rate of 0.2 cm 3 /min to saturate the core plug.After that, the NMR apparatus (Mini-NMR, Niumag, Suzhou, China) was used to measure the transverse relaxation time T 2 of the core sample under the initial water-saturated condition.The magnetic intensity, gradient value control precision, and frequency range of the NMR apparatus were 0.5 T, 0.025 T/m, 0.01 MHz, and 1-30 MHz, respectively.(3) The core was displaced with MnCl 2 solution (15,000 mg/L) of 5 PV.Then, the saturated core was scanned again by the NMR apparatus to ensure the hydrogen signal of the brine was eliminated.(4) After that, 3.0 PV of the crude oil was pumped through the core plugs at a constant rate of 0.1 cm 3 /min until no water was produced to achieve the connate water saturation (S wc ) and the initial oil saturation (S oi ) at a reservoir temperature of 61 • C. The T 2 spectrum was measured again after the core had been saturated with crude oil.(5) In each test, the pressure of the CO 2 in the high-pressure cylinder was increased to the prespecified injection pressure at the temperature of 61 • C.Then, the CO 2 was injected into the oil saturated core at constant pressure to the desired pressure and the pump maintained the constant pressure condition for 30 min.After CO 2 injection, CO 2 was allowed to soak for t soak = 6 h.Then, the oil was produced from the same end of the coreholder at atmospheric pressure.The huff-and-puff cycles continued until no considerable oil production was obtained.The injection and production pressure were continuously monitored and recorded during the entire test.A video camera was utilized to measure the cumulative produced oil volume.(6) After the CO 2 huff-and-puff test, the T 2 transverse relaxation time of the core samples was measured with the NMR apparatus.As shown in Figure 6, CO 2 solubility increased with increasing equilibrium pressure.At the same time, the oil swelling factor increased due to more dissolved CO 2 in the crude oil.However, the oil swelling factor did not monotonously increase throughout the experiment.When the equilibrium pressure was higher than 9.51 MPa, the crude oil expanded by 28% and the oil swelling factor began to decrease.As the pressure continued to increase, the escaping rate of the light component molecules began to be greater than the rate of CO 2 dissolved in the liquid phase [30].When the equilibrium pressure reached 14.83 MPa, the oil swelling factor was only 0.61.

Results and Discussion
Energies 2018, 11, x FOR PEER REVIEW 7 of 16

CO2 Solubility and Oil Swelling Measurement
The CO2 solubility in the crude oil, which is the key parameter of the performance of the CO2 injection process, directly influences the oil swelling factor, oil viscosity, oil density, and CO2-oil IFT [29].The swelling/extraction test of the oil sample was conducted at a reservoir temperature of 61 °C .As shown in Figure 6, CO2 solubility increased with increasing equilibrium pressure.At the same time, the oil swelling factor increased due to more dissolved CO2 in the crude oil.However, the oil swelling factor did not monotonously increase throughout the experiment.When the equilibrium pressure was higher than 9.51 MPa, the crude oil expanded by 28% and the oil swelling factor began to decrease.As the pressure continued to increase, the escaping rate of the light component molecules began to be greater than the rate of CO2 dissolved in the liquid phase [30].When the equilibrium pressure reached 14.83 MPa, the oil swelling factor was only 0.61.The curve of the oil swelling factor at various equilibrium pressures was divided into two sections.When the oil swelling factor increases, the dominant effect was the oil swelling effect due to CO2 dissolved in the crude oil, though some light components in the liquid phase evaporated.As the oil swelling factor started to decrease, the extraction effect played a dominant role.The minimum extraction pressure (MEP) is defined as the pressure where the oil swelling factor begins to decrease.When the equilibrium pressure was between 9.51 MPa and 12.20 MPa, the oil swelling factor decreased rapidly because the light components are easily extracted.The larger the molecular weight, the smaller the rate of molecular motion at the same temperature [31].When the pressure was higher than 12.20 MPa, most light components were already extracted and molecules with higher weights is more difficult to be extracted by CO2.Thus, the curve of oil swelling factor with pressure decease The curve of the oil swelling factor at various equilibrium pressures was divided into two sections.When the oil swelling factor increases, the dominant effect was the oil swelling effect due to CO 2 dissolved in the crude oil, though some light components in the liquid phase evaporated.As the oil swelling factor started to decrease, the extraction effect played a dominant role.The minimum extraction pressure (MEP) is defined as the pressure where the oil swelling factor begins to decrease.When the equilibrium pressure was between 9.51 MPa and 12.20 MPa, the oil swelling factor decreased rapidly because the light components are easily extracted.The larger the molecular weight, the smaller the rate of molecular motion at the same temperature [31].When the pressure was higher than 12.20 MPa, most light components were already extracted and molecules with higher weights is more difficult to be extracted by CO 2 .Thus, the curve of oil swelling factor with pressure decease much more slowly.

IFT Measurement
In this study, the IFT between the CO 2 and crude oil was determined using the ADSA technique.Figure 7 shows the measured IFT of the CO 2 -crude oil system at various equilibrium pressure.The IFT of the CO 2 -oil system decreases with increasing pressure and the CO 2 -oil system was considered to be miscible when the IFT was 0 [32].The whole curve was divided into two sections to fit and extrapolate to intersect with the X axis, respectively (Figure 7).The two fitted straight lines intersect at 9.71 MPa, which was 0.2 MPa higher than the MEP of the crude oil (9.51 MPa).Therefore, 9.71 MPa was considered to be the critical point where the CO 2 -ROE mechanism changed from oil swelling effect to hydrocarbon extraction effect.The pressure in Range 1 was defined from 0.047 MPa to 9.71 MPa and the intercept of the linear regression of points in Range 1 was 11.83 MPa (Figure 7).Corresponding to the IFT tests, the phenomenon during the swelling/extraction tests was recorded by photographs, as shown in Figure 8.When the equilibrium pressure was 3.21 MPa, which was much lower than the MEP of the crude oil, the oil in the area below the phase interface was lighter in color due to CO2 dissolution.As the pressure increased to a MEP of 9.51 MPa, light-colored droplets appeared on the glass window.The time required to reach equilibrium between the gas and liquid phase increased with increasing equilibrium pressure.When the equilibrium pressure reached 11.73 MPa, a light-colored area that was about 5-6 mm appeared above the liquid level.This area was mainly composed of light and medium components as well as dissolved CO2.Multi-phase contact between CO2 and crude oil is necessary for CO2-oil system to attain miscibility.Therefore, 11.83 MPa was the multi-phase-contact minimum miscibility pressure of the CO2-oil system.When the pressure was higher than M-MMP, the oil swelling factor changed little.The second linear regression intersected with IFT = 0 at 21.89 MPa, which was interpreted as the first-contact minimum miscibility pressure because almost all medium and heavy components were also miscible with CO2 at this pressure.[33].The pressure in Range 1 was defined from 0.047 MPa to 9.71 MPa and the intercept of the linear regression of points in Range 1 was 11.83 MPa (Figure 7).Corresponding to the IFT tests, the phenomenon during the swelling/extraction tests was recorded by photographs, as shown in Figure 8.When the equilibrium pressure was 3.21 MPa, which was much lower than the MEP of the crude oil, the oil in the area below the phase interface was lighter in color due to CO 2 dissolution.As the pressure increased to a MEP of 9.51 MPa, light-colored droplets appeared on the glass window.The time required to reach equilibrium between the gas and liquid phase increased with increasing equilibrium pressure.When the equilibrium pressure reached 11.73 MPa, a light-colored area that was about 5-6 mm appeared above the liquid level.This area was mainly composed of light and medium components as well as dissolved CO 2 .Multi-phase contact between CO 2 and crude oil is necessary for CO 2 -oil system to attain miscibility.Therefore, 11.83 MPa was the multi-phase-contact minimum miscibility pressure of the CO 2 -oil system.When the pressure was higher than M-MMP, the oil swelling factor changed little.The second linear regression intersected with IFT = 0 at 21.89 MPa, which was interpreted as the first-contact minimum miscibility pressure because almost all medium and heavy components were also miscible with CO 2 at this pressure [33].
Energies 2018, 11, 2843 9 of 16 medium components as well as dissolved CO2.Multi-phase contact between CO2 and crude oil is necessary for CO2-oil system to attain miscibility.Therefore, 11.83 MPa was the multi-phase-contact minimum miscibility pressure of the CO2-oil system.When the pressure was higher than M-MMP, the oil swelling factor changed little.The second linear regression intersected with IFT = 0 at 21.89 MPa, which was interpreted as the first-contact minimum miscibility pressure because almost all medium and heavy components were also miscible with CO2 at this pressure.[33].

Effect of Injection Pressure on Residual Oil
Figure 9 shows the measured ORF versus cycle number at different injection pressures.The total cycle number of each CO 2 huff-and-puff test was lower at higher injection pressures.When the injection pressure was less than MEP of 9.51 MPa, seven cycles were conducted until no oil was obtained in each test.When the injection pressure was higher than the MMP of 11.83 MPa, the cyclic CO 2 injection test was terminated after five cycles.The ultimate ORF increased with the injection pressure and gradually slowed down (Figures 9 and 10).The ultimate ORF increased remarkably from 27.22% to 52.56% as the injection pressure increased from 5 MPa to 13 MPa.Obvious oil recovery was not obtained with further increase in the injection pressure, and the ORF only increased by 1.04% as the injection pressure increased from 13 MPa to 16 MPa.

Effect of Injection Pressure on Residual Oil
Figure 9 shows the measured ORF versus cycle number at different injection pressures.The total cycle number of each CO2 huff-and-puff test was lower at higher injection pressures.When the injection pressure was less than MEP of 9.51 MPa, seven cycles were conducted until no oil was obtained in each test.When the injection pressure was higher than the MMP of 11.83 MPa, the cyclic CO2 injection test was terminated after five cycles.The ultimate ORF increased with the injection pressure and gradually slowed down (Figures 9 and 10).The ultimate ORF increased remarkably from 27.22% to 52.56% as the injection pressure increased from 5 MPa to 13 MPa.Obvious oil recovery was not obtained with further increase in the injection pressure, and the ORF only increased by 1.04% as the injection pressure increased from 13 MPa to 16 MPa.The viscosity of the produced oil in the first cycle of the CO2 huff-and-puff tests performed at different injection pressures is shown in Figure 10.The viscosity of the produced oil decreased from 9.40 mPa•s to 2.44 mPa•s as the injection pressure increased.When the injection pressure was lower than the MEP, the interface of the CO2-oil was obvious.The oil swelling effect was determined as the dominant recovery mechanism and the extraction effect of the CO2 was weak.So, the viscosity of the produced oil and the composition of the residual oil in the formation changed little.The produced oil form each cycle was mainly recovered through the elastic energy of the reservoir.More cycles were necessary at lower injection pressures to reach the development capacity of the reservoir.
When the injection pressure was higher than the MEP, especially higher than the MMP, more light and medium components could be extracted by the injected CO2.The produced oil could be categorized into two portions: the lighter oil extracted by the injected CO2 and the oil recovered due to the elastic energy and oil swelling effect.The proportion of the lighter oil increased with injection When the injection pressure was higher than the MEP, especially higher than the MMP, more light and medium components could be extracted by the injected CO 2 .The produced oil could be categorized into two portions: the lighter oil extracted by the injected CO 2 and the oil recovered due to the elastic energy and oil swelling effect.The proportion of the lighter oil increased with injection pressure, so the viscosity of the produced oil was lower at higher injection pressure.As the huff-and-puff process progressed, the heavy components content of the residual oil increased.The solubility of the CO 2 in the residual oil decreased and the CO 2 played a weaker role in the later cycles.Therefore, the ORF in each cycle was higher and fewer cycles were needed at higher injection pressure during the CO 2 huff-and-puff processes.In order to investigate the effect of injection pressure on the microscopic residual oil distribution during CO2 huff-and-puff processes, the NMR technique was applied to measure core #1 after the first cycle of cyclic CO2 injection at injection pressures of 5 MPa, 11 MPa, and 16 MPa.As shown in Figure 11, the pore distribution of core #1 was bimodal with larger slanting degrees from the T2 spectrum of the core saturated with water.The T2 spectrum was artificially divided into four intervals: micro pores (0.1-1 ms), small pores (1-10 ms), medium pores (10-100 ms), and large pores (100-1000 ms).By comparing the differences between the T2 spectrum for core #1 saturated with water and that saturated with oil, the sum of the amplitude in micro, small, medium, and large pores decreased by 43.43%, 35.55%, 16.88%, and 6.14%, respectively.The irreducible water was mainly distributed in micro and small pores.
As shown in Figure 11, the amplitude of the T2 spectrum of medium and large pores (10-1000 ms) decreased homogeneously after the first cycle at different injection pressures.The ORF of the medium and large pores (10-1000 ms) was 10.85%, 21.37%, and 30.42% at the injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively, whereas the ORF of the micro and small pores (0.1-10 ms) was 4.16%, 9.56%, and 14.30%, respectively.It should be noted that the CO2 injection caused asphaltene precipitation in the pores, which influenced the T2 spectrum measured by the NMR technique [34].In the CO2 injection process, the CO2 molecules began to occupy the larger pores due to the lower oil-gas phase.The injected CO2 interacted with the crude oil and dissolved in the oil.As soaking progressed, the CO2 gradually diffused into smaller pores.Therefore, the escape of CO2 drove the crude oil into some larger pores.At the same time, the oil in smaller pores swelled and expanded into the larger pores of the core.In order to investigate the effect of injection pressure on the microscopic residual oil distribution during CO 2 huff-and-puff processes, the NMR technique was applied to measure core #1 after the first cycle of cyclic CO 2 injection at injection pressures of 5 MPa, 11 MPa, and 16 MPa.As shown in Figure 11, the pore distribution of core #1 was bimodal with larger slanting degrees from the T 2 spectrum of the core saturated with water.The T 2 spectrum was artificially divided into four intervals: micro pores (0.1-1 ms), small pores (1-10 ms), medium pores (10-100 ms), and large pores (100-1000 ms).By comparing the differences between the T 2 spectrum for core #1 saturated with water and that saturated with oil, the sum of the amplitude in micro, small, medium, and large pores decreased by 43.43%, 35.55%, 16.88%, and 6.14%, respectively.The irreducible water was mainly distributed in micro and small pores.
As shown in Figure 11, the amplitude of the T 2 spectrum of medium and large pores (10-1000 ms) decreased homogeneously after the first cycle at different injection pressures.The ORF of the medium and large pores (10-1000 ms) was 10.85%, 21.37%, and 30.42% at the injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively, whereas the ORF of the micro and small pores (0.1-10 ms) was 4.16%, 9.56%, and 14.30%, respectively.It should be noted that the CO 2 injection caused asphaltene precipitation in the pores, which influenced the T 2 spectrum measured by the NMR technique [34].In the CO 2 injection process, the CO 2 molecules began to occupy the larger pores due to the lower oil-gas phase.The injected CO 2 interacted with the crude oil and dissolved in the oil.As soaking progressed, the CO 2 gradually diffused into smaller pores.Therefore, the escape of CO 2 drove the crude oil into some larger pores.At the same time, the oil in smaller pores swelled and expanded into the larger pores of the core.Through the variation in the T2 spectrum of core #1 after the first cycle at different pressures and the T2 spectrum of core #1 saturated with oil, the interval of the pore distribution where the oil was recoverable expanded as the injection pressure increased (Figure 11).The lower limit of the pore distribution where the oil was recoverable corresponds to T2 values of 2.68 ms, 1.29 ms, and 0.74 ms at the injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively.This is because the CO2 overcomes the capillary pressure and enters smaller pores with higher injection pressure [35].Moreover, when the injection pressure is higher than the MEP, the injected CO2 extracts the oil in the smaller and blind pores, which converts the residual oil into recoverable oil.Therefore, higher injection pressure is beneficial for recovering the oil in smaller and blind pores.

Effect of Cycle Number on Residual Oil
As shown in Figure 12, the ORF in each cycle of different injection pressures decreased with the increase in cycle number.When the injection pressure was higher than the MMP, only five cycles were carried out until no oil was produced.Higher injection pressure is beneficial for improving the ultimate ORF and reducing the cumulative time of the total development process due to fewer huffand-puff cycles.As a result, a higher production rate was achieved.Through the variation in the T 2 spectrum of core #1 after the first cycle at different pressures and the T 2 spectrum of core #1 saturated with oil, the interval of the pore distribution where the oil was recoverable expanded as the injection pressure increased (Figure 11).The lower limit of the pore distribution where the oil was recoverable corresponds to T 2 values of 2.68 ms, 1.29 ms, and 0.74 ms at the injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively.This is because the CO 2 overcomes the capillary pressure and enters smaller pores with higher injection pressure [35].Moreover, when the injection pressure is higher than the MEP, the injected CO 2 extracts the oil in the smaller and blind pores, which converts the residual oil into recoverable oil.Therefore, higher injection pressure is beneficial for recovering the oil in smaller and blind pores.

Effect of Cycle Number on Residual Oil
As shown in Figure 12, the ORF in each cycle of different injection pressures decreased with the increase in cycle number.When the injection pressure was higher than the MMP, only five cycles were carried out until no oil was produced.Higher injection pressure is beneficial for improving the ultimate ORF and reducing the cumulative time of the total development process due to fewer huff-and-puff cycles.As a result, a higher production rate was achieved.Through the variation in the T2 spectrum of core #1 after the first cycle at different pressures and the T2 spectrum of core #1 saturated with oil, the interval of the pore distribution where the oil was recoverable expanded as the injection pressure increased (Figure 11).The lower limit of the pore distribution where the oil was recoverable corresponds to T2 values of 2.68 ms, 1.29 ms, and 0.74 ms at the injection pressure of 5 MPa, 11 MPa, and 16 MPa, respectively.This is because the CO2 overcomes the capillary pressure and enters smaller pores with higher injection pressure [35].Moreover, when the injection pressure is higher than the MEP, the injected CO2 extracts the oil in the smaller and blind pores, which converts the residual oil into recoverable oil.Therefore, higher injection pressure is beneficial for recovering the oil in smaller and blind pores.

Effect of Cycle Number on Residual Oil
As shown in Figure 12, the ORF in each cycle of different injection pressures decreased with the increase in cycle number.When the injection pressure was higher than the MMP, only five cycles were carried out until no oil was produced.Higher injection pressure is beneficial for improving the ultimate ORF and reducing the cumulative time of the total development process due to fewer huffand-puff cycles.As a result, a higher production rate was achieved.The NMR T 2 spectrum for core #1 after different numbers of cycles of CO 2 huff-and-puff tests at the injection pressure of 11 MPa is presented in Figure 13.The cumulative ORFs after the first, third, and sixth cycle of core #1 were 17.13%, 36.68%, and 48.48%, respectively.The initial oil saturation and the cumulative ORF contributed by different intervals of pore distribution are shown in Table 2.As the number of huff-and-puff process cycles increased, the ORF contributed by small, medium, and large pores (1-1000 ms) increased significantly, while the oil produced from micro pores (0.1-1 ms) had no obvious changes.The lower limit of the pore distribution where the oil was recoverable also did not increase with increasing cycle number.So, the oil in smaller pores cannot be recovered with increasing the number of huff-and-puff cycles.Residual oil in pores, which was distributed in the range of 1-1000 ms of the T 2 spectrum, was still the main force contributing to production.The NMR T2 spectrum for core #1 after different numbers of cycles of CO2 huff-and-puff tests at the injection pressure of 11 MPa is presented in Figure 13.The cumulative ORFs after the first, third, and sixth cycle of core #1 were 17.13%, 36.68%, and 48.48%, respectively.The initial oil saturation and the cumulative ORF contributed by different intervals of pore distribution are shown in Table 2.As the number of huff-and-puff process cycles increased, the ORF contributed by small, medium, and large pores (1-1000 ms) increased significantly, while the oil produced from micro pores (0.1-1 ms) had no obvious changes.The lower limit of the pore distribution where the oil was recoverable also did not increase with increasing cycle number.So, the oil in smaller pores cannot be recovered with increasing the number of huff-and-puff cycles.Residual oil in pores, which was distributed in the range of 1-1000 ms of the T2 spectrum, was still the main force contributing to production.

Effect of Soaking Time on Residual Oil
The range of soaking time in the field scale ranges from several days to tens of days, which has an significant effect on the oil recovery performance of CO2 huff-and-puff [36].Figure 14 shows the ORF and average recovery rate of the first CO2 huff-and-puff cycle versus the soaking time at the injection pressure of 11 MPa.The extension of soaking time resulted in a noticeable ORF improvement.The average recovery rate is defined as the ORF per hour during an entire huff-andpuff process, including the injection, soaking, and production period.The ORF increased from 13.20% to 23.91% when the soaking time increased from 3 h to 18 h and the average recovery rate decreased from 3.77%/h to 1.29%/h.As the soaking time increased from 18 h to 24 h, the ORF only increased by 1.03%.The contact between CO2 and oil is a dynamic process.The CO2 phase contacts the oil phase completely and dissolves into the oil phase fully with longer soaking time, due to viscosity reduction, oil swelling, solution gas drive, and extraction effect of the injected CO2, to improve the ORF.When

Effect of Soaking Time on Residual Oil
The range of soaking time in the field scale ranges from several days to tens of days, which has an significant effect on the oil recovery performance of CO 2 huff-and-puff [36].Figure 14 shows the ORF and average recovery rate of the first CO 2 huff-and-puff cycle versus the soaking time at the injection pressure of 11 MPa.The extension of soaking time resulted in a noticeable ORF improvement.The average recovery rate is defined as the ORF per hour during an entire huff-and-puff process, including the injection, soaking, and production period.The ORF increased from 13.20% to 23.91% when the soaking time increased from 3 h to 18 h and the average recovery rate decreased from 3.77%/h to 1.29%/h.As the soaking time increased from 18 h to 24 h, the ORF only increased by 1.03%.The contact between CO 2 and oil is a dynamic process.The CO 2 phase contacts the oil phase completely and dissolves into the oil phase fully with longer soaking time, due to viscosity reduction, oil swelling, solution gas drive, and extraction effect of the injected CO 2 , to improve the ORF.When the soaking time was longer than 18 h, the dissolved CO 2 concentration in the accessible oil tended to be saturated and the dissolution and diffusion rate slowed down.The CO 2 had no significant effect on improving the ORF.
Energies 2018, 11, x FOR PEER REVIEW 13 of 16 the soaking time was longer than 18 h, the dissolved CO2 concentration in the accessible oil tended to be saturated and the dissolution and diffusion rate slowed down.The CO2 had no significant effect on improving the ORF.The NMR T2 spectrum for core #2 with different soaking times for the first huff-and-puff cycle at the injection pressure of 11 MPa is shown in Figure 15.The amplitude of the micro pores (0.1-1 ms) had no obvious change.The amplitude of small pores (1-10 ms) decreased by 9.84%, 16.06%, and 22.49% and the amplitude of medium and large pores (10-1000 ms) decreased by 14.38%, 23.37%, and 29.78% with soaking time of 3 h, 12 h, and 24 h, respectively.The residual oil of the interval of pore distribution where the recoverable oil homogeneously decreased with longer soaking time.The lower limit of the pore distribution where the oil was recoverable expanded as the soaking time increased, which means crude oil in smaller pores turns from residual oil to recoverable oil.This is because the injected CO2 dissolved and diffused in smaller pores with longer soaking time.Although the ORF and the recoverable oil increased, the average recovery rate decreased with longer soaking time (Figure 14).Oilfields are usually assigned production tasks with a specified time.Therefore, under the premise of being able to meet the assigned production requirements, longer soaking times can fully exploit the potential of the reservoir and reduce the residual oil in smaller pores within a certain range.The NMR T 2 spectrum for core #2 with different soaking times for the first huff-and-puff cycle at the injection pressure of 11 MPa is shown in Figure 15.The amplitude of the micro pores (0.1-1 ms) had no obvious change.The amplitude of small pores (1-10 ms) decreased by 9.84%, 16.06%, and 22.49% and the amplitude of medium and large pores (10-1000 ms) decreased by 14.38%, 23.37%, and 29.78% with soaking time of 3 h, 12 h, and 24 h, respectively.The residual oil of the interval of pore distribution where the recoverable oil homogeneously decreased with longer soaking time.The lower limit of the pore distribution where the oil was recoverable expanded as the soaking time increased, which means crude oil in smaller pores turns from residual oil to recoverable oil.This is because the injected CO 2 dissolved and diffused in smaller pores with longer soaking time.Although the ORF and the recoverable oil increased, the average recovery rate decreased with longer soaking time (Figure 14).Oilfields are usually assigned production tasks with a specified time.Therefore, under the premise of being able to meet the assigned production requirements, longer soaking times can fully exploit the potential of the reservoir and reduce the residual oil in smaller pores within a certain range.

Conclusions
In this paper, a phase behavior study was carried out to clarify the CO2 recovery mechanisms during the CO2 injection process.Afterward, a series of CO2 huff-and-puff tests were conducted at various conditions.NMR was used to determine the microscopic residual oil distribution.From this

Conclusions
In this paper, a phase behavior study was carried out to clarify the CO 2 recovery mechanisms during the CO 2 injection process.Afterward, a series of CO 2 huff-and-puff tests were conducted at various conditions.NMR was used to determine the microscopic residual oil distribution.From this study, the following conclusions were drawn.
The MEP was determined to be 9.51 MPa through oil swelling tests in the visual cell.When the pressure increased higher than the MEP, the hydrocarbon extraction effect played a more dominant role in the CO 2 -EOR process and the gas-fluid interface became increasingly blurred.In the CO 2 huff-and-puff process, the ultimate ORF increased considerably as injection pressure increased, whereas the viscosity of the produced oil in the first cycle decreased from 9.40 mPa•s to 2.44 mPa•, due to the stronger hydrocarbon-extraction capacity of the injected CO 2 at higher pressures.As injection pressure increased, a higher ultimate ORF was obtained with fewer cycle numbers, so that a higher production rate was achieved to reach the development capacity of the reservoir.
CO 2 overcomes the capillary pressure and enters smaller pores to extract the oil in the smaller and blind pores at higher injection pressure, which converts the residual oil into recoverable oil.Longer soaking time also expands the lower limit of the pore distribution where the oil was recoverable.However, more cycles had no obvious effect on expanding the interval of pore distribution where oil was recoverable, which only homogeneously reduced the residual oil in the medium and large pores (10-1000 ms).Therefore, increasing injection pressure and extending soaking time are effective methods to recover the residual oil in smaller and blind pores within a certain range.The determination of microscopic residual oil distribution is not only beneficial for ensuring the EOR mechanisms in the microscopic pores of the CO 2 huff-and-puff, but also for optimizing production parameters to exploit the tight reservoirs to their maximum potential in the shortest possible time.

Figure 1 .
Figure 1.Logic structure diagram of this paper.

Figure 2 .
Figure 2. Gas chromatography (GC) compositional analysis of the crude oil sample.

Figure 1 .
Figure 1.Logic structure diagram of this paper.

Figure 2 .
Figure 2. Gas chromatography (GC) compositional analysis of the crude oil sample.

Figure 2 .
Figure 2. Gas chromatography (GC) compositional analysis of the crude oil sample.

Figure 3 .
Figure 3. Schematic diagram of the experimental apparatus used for carbon dioxide (CO2) solubility and oil swelling factor measurements.

Figure 3 .
Figure 3. Schematic diagram of the experimental apparatus used for carbon dioxide (CO 2 ) solubility and oil swelling factor measurements.

Figure 4 .
Figure 4. Schematic diagram of the experimental setup used for measuring the equilibrium interfacial tension (IFT) between CO2 and crude oil by applying the axisymmetric drop shape analysis (ADSA) technique for the pendant drop.

Figure 4 .
Figure 4. Schematic diagram of the experimental setup used for measuring the equilibrium interfacial tension (IFT) between CO 2 and crude oil by applying the axisymmetric drop shape analysis (ADSA) technique for the pendant drop.

Energies 2018 ,
11, x FOR PEER REVIEW 6 of 16regulator (HDB-03, Huada, Haian, China) was used to target the desired production pressure during the coreflood test.The produced oil was collected in a burette and the volume of the produced gas was measured by a gas flow meter.

Figure 5 .
Figure 5. Schematic diagram of the experimental setup used for CO2 huff-and-puff tests.

Figure 5 .
Figure 5. Schematic diagram of the experimental setup used for CO 2 huff-and-puff tests.

3. 1 .
Phase Behaviors of CO 2 -Oil System 3.1.1.CO 2 Solubility and Oil Swelling Measurement The CO 2 solubility in the crude oil, which is the key parameter of the performance of the CO 2 injection process, directly influences the oil swelling factor, oil viscosity, oil density, and CO 2 -oil IFT [29].The swelling/extraction test of the oil sample was conducted at a reservoir temperature of 61 • C.

Figure 6 .
Figure 6.CO2 solubility and oil swelling factor at various equilibrium pressures under the reservoir temperature of 61 °C .

Figure 6 .
Figure 6.CO 2 solubility and oil swelling factor at various equilibrium pressures under the reservoir temperature of 61 • C.

Energies 2018 ,
11,  x FOR PEER REVIEW 8 of 16 at 9.71 MPa, which was 0.2 MPa higher than the MEP of the crude oil (9.51 MPa).Therefore, 9.71 MPa was considered to be the critical point where the CO2-ROE mechanism changed from oil swelling effect to hydrocarbon extraction effect.

Figure 7 .
Figure 7.The IFT of the CO2-crude oil system at different equilibrium pressures under the temperature of 61 °C .

Figure 7 .
Figure 7.The IFT of the CO 2 -crude oil system at different equilibrium pressures under the temperature of 61 • C.

Figure 8 .
Figure 8. Images of the CO2-oil interface at different equilibrium pressure during swelling/extraction tests in the visual cell.

Figure 8 .
Figure 8. Images of the CO 2 -oil interface at different equilibrium pressure during swelling/extraction tests in the visual cell.

Figure 9 .
Figure 9. Cumulative oil recovery factor (ORF) versus cycle number at different injection pressures (core #1).The viscosity of the produced oil in the first cycle of the CO 2 huff-and-puff tests performed at different injection pressures is shown in Figure10.The viscosity of the produced oil decreased from 9.40 mPa•s to 2.44 mPa•s as the injection pressure increased.When the injection pressure was lower than the MEP, the interface of the CO 2 -oil was obvious.The oil swelling effect was determined as the dominant recovery mechanism and the extraction effect of the CO 2 was weak.So, the viscosity of the produced oil and the composition of the residual oil in the formation changed little.The produced oil

Energies 2018 , 16 Figure 10 .
Figure 10.Ultimate ORFs, ORFSs of the first cycle and the viscosity of the produced oil in the first cycle of CO2 huff-and-puff tests performed at different injection pressures (core #1).

Figure 10 .
Figure 10.Ultimate ORFs, ORFSs of the first cycle and the viscosity of the produced oil in the first cycle of CO 2 huff-and-puff tests performed at different injection pressures (core #1).

Figure 11 .
Figure 11.Nuclear magnetic resonance (NMR) T2 spectrum for core #1 after the first cycle at different injection pressures during CO2 huff-and-puff tests.

Figure 12 .
Figure 12.ORFs of each cycle number of the CO2 huff-and-puff tests at different injection pressures.

Figure 11 .
Figure 11.Nuclear magnetic resonance (NMR) T 2 spectrum for core #1 after the first cycle at different injection pressures during CO 2 huff-and-puff tests.

Figure 12 .
Figure 12.ORFs of each cycle number of the CO2 huff-and-puff tests at different injection pressures.Figure 12. ORFs of each cycle number of the CO 2 huff-and-puff tests at different injection pressures.

Figure 12 .
Figure 12.ORFs of each cycle number of the CO2 huff-and-puff tests at different injection pressures.Figure 12. ORFs of each cycle number of the CO 2 huff-and-puff tests at different injection pressures.

Figure 13 .
Figure 13.NMR T2 spectrum for core #1 after different cycle numbers of CO2 huff-and-puff tests at the injection pressure of 11 MPa.

Figure 13 .
Figure 13.NMR T 2 spectrum for core #1 after different cycle numbers of CO 2 huff-and-puff tests at the injection pressure of 11 MPa.

Figure 14 .
Figure 14.The ORFs and average recovery rate of first CO2 huff-and-puff cycle versus the soaking time at an operating pressure of 11 MPa (core #2).

Figure 14 .
Figure 14.The ORFs and average recovery rate of first CO 2 huff-and-puff cycle versus the soaking time at an operating pressure of 11 MPa (core #2).

Energies 2018 , 16 Figure 15 .
Figure 15.NMR T2 spectrum for core #2 with different soaking times for the first huff-and-puff cycle at the injection pressure of 11 MPa.

Figure 15 .
Figure 15.NMR T 2 spectrum for core #2 with different soaking times for the first huff-and-puff cycle at the injection pressure of 11 MPa.

Table 1 .
Basic properties of tight core plug samples.

Table 1 .
Basic properties of tight core plug samples.

Table 1 .
Basic properties of tight core plug samples.

Table 2 .
Initial oil saturation and cumulative ORFs after different huff-and-puff cycles.

Table 2 .
Initial oil saturation and cumulative ORFs after different huff-and-puff cycles.