Investigation of Polymer-Assisted CO2 Flooding to Enhance Oil Recovery in Low-Permeability Reservoirs

CO2 flooding is a favorable technical means for the efficient development of low-permeability reservoirs, and it can also contribute to the realization of net-zero CO2 emissions. However, due to the unfavorable viscosity ratio and gravity overriding effect, CO2 channeling will inevitably occur, seriously affecting its storage and displacement effects. This paper conducts a systematic study on the application of polymer-assisted CO2 flooding in low-permeability reservoirs. Firstly, the polymer agent suitable for low-permeability reservoirs is optimized through the viscosity-increasing, rheological, and temperature- and salt-resistant properties of the solution. Then, the injectivity performance, resistance-increasing ability, and profile-improving effect of the polymer solution were evaluated through core experiments, and the optimum concentration was optimized. Finally, the enhanced oil recovery (EOR) effects of polymer-assisted and water-assisted CO2 flooding were compared. The results show that the temperature-resistant polymer surfactant (TRPS) has a certain viscosity-increasing performance, good temperature resistance performance, and can react with CO2 to increase the solution viscosity significantly. Meanwhile, TRPS has good injection performance and resistance-increasing effect. The resistance increasing factor (η and η′) of TRPS-assisted CO2 flooding increases with increased permeability, the concentration of TRPS solution, and injection rounds. Considering η′ and the profile improvement effect comprehensively, the application concentration of TRPS should be 1000 mg/L. The EOR effect of TRPS-assisted CO2 flooding is 8.21% higher than that of water-assisted CO2 flooding. The main effective period is in the first and second rounds, and the best injection round is three. The research content of this paper can provide data support for the field application of polymer-assisted CO2 flooding in low-permeability reservoirs.


Introduction
With increased international energy consumption and oil exploration and development intensification, low-permeability reservoirs have contributed to Chinese oil and gas resource production. Developing low-permeability oilfields is generally dominated by water injection, and gas flooding, especially CO 2 flooding, is an efficient development technology for low-permeability reservoirs with excellent prospects [1,2]. CO 2 flooding can play the role of high-efficiency oil displacement and geological storage simultaneously, in line with the policy of carbon peak and carbon neutrality, and has good application advantages. However, due to unfavorable mobility ratio and gravity overriding, CO 2 will channel in the reservoir, significantly reducing the development effect [3][4][5]. Therefore, an economical and effective method for mitigating gas channeling is the key to further improving CO 2 flooding in low-permeability reservoirs. Polymer flooding is the primary EOR technology in China [6], and it has been successfully applied in Daqing Oilfield, Xinjiang Oilfield, Dagang Oilfield, etc. [7,8]. Polymers can reduce the fluidity of injected fluids, reduce the permeability of advantageous channels through adsorption and retention, and ultimately expand the swept volume [9,10]. The resistance coefficient and residual resistance coefficient are usually used to evaluate the injectivity performance of polymer solutions, and having a suitable resistance coefficient and residual resistance coefficient is the prerequisite for the successful application of polymers [11]. Indoor research results show that polyacrylamide (HPAM) polymers can increase oil recovery by 10-12% of OOIP after water flooding [12]. Meanwhile, the molecular modification of the polymer can play an interface effect, stripping, dispersing, and carrying oil droplets [13]. In addition, polymer gels [14], polymer microspheres [15][16][17], polymeric surfactants [18], and other agents [19] can play a role in profile control and flooding, further improving oil recovery. A comparison of the EOR effects of different enhanced CO 2 injection methods is shown in Appendix A Table A1. It can be found that as the system strength increases, the EOR gradually increases. However, due to the high viscosity characteristics of polymers, it must keep a good reservoir matching relationship while exerting the viscosity-increasing effect [20,21], so it is usually used in medium-high permeability reservoirs.
It is of excellent research significance to explore using polymers to assist CO 2 flooding in low-permeability reservoirs to improve gas channeling and expand the swept volume. Dann et al. [22] compared the injection performance of polymer solutions in different lowpermeability cores through core flooding experiments and pointed out that reservoir core characteristics other than permeability can dominate polymer performance. In addition, Ghosh et al. [23] conducted polymer injectivity experiments on two cores with a permeability of 15 mD but with significant differences in pore throat distribution. The results showed that the polymer could flow smoothly in the cores with a bimodal distribution of pore throats. But in another core with unimodal distribution, the flow is difficult. The above two studies show that the polymer solution has the potential to flow smoothly in the low-permeability reservoir, but it is necessary to meet the matching of the polymer and the reservoir. Marliere et al. [24] used cores with an effective permeability of 1-3 mD to carry out oil displacement experiments with polymers (molecular weight of 3 million Da and viscosity of 2 cP), which can increase oil recovery by 20% to 40% based on water flooding. Bennetzen et al. [25] proved by experiments that partially hydrolyzed polyacrylamide (HPAM, with a molecular weight of 8 million Da and a concentration of less than 5000 mg/L) solution injected into a 0.3 mD carbonate reservoir core at a rate of 0.5 mL/h will not produce core plugging. Mohammed Taha et al. [26] studied the application of lowsalinity polymer flooding in high-temperature, high-salt, and low-permeability reservoirs. The polymers can maintain a viscosity of 2-3 cP (2500-4000 mg/L), with good injection and enhanced oil recovery effects. Leon et al. [27] reported a successful field pilot of the polymer in the low-permeability Palogrande-Cehu field. Considering the low permeability of the reservoir (6-150 mD), three polymers with different molecular weights (average concentration 1021 mg/L, viscosity 3.43 cP) were used, and the viscosity change was monitored. It is estimated that EOR efficiency can reach up to 8%, and the water cut of some wells is reduced by as much as 14%. The above studies show that the successful application of polymers in low-permeability reservoirs requires low viscosity and low injection velocity. For CO 2 flooding in low-permeability reservoirs, a CO 2 response mechanism based on the above studies [28][29][30] can be used to prevent gas channeling. Inject a low-viscosity polymer surfactant into the reservoir, which can react with CO 2 to form worm-like micelles, significantly increasing the viscosity of the system [31,32]. However, due to the high viscosity of this type of micelles, it is mainly used for the plugging of fractures in low-permeability reservoirs [33][34][35]. Through the above literature review, two problems exist in the current research on polymer-assisted CO 2 flooding in low-permeability reservoirs: (1) how to optimize the polymer agents with reservoir adaptability; and (2) how to apply the CO 2 response characteristics in low-permeability reservoirs matrix.
Aiming at the above problems, the present paper studies a temperature-resistant CO 2responsive polymer surfactant (TRPS) to assist CO 2 flooding to improve the oil recovery of low-permeability reservoirs. The advantages of it as an assisted agent were evaluated by comparing its static properties, such as viscosifying properties, rheological properties, and temperature and CO 2 resistance with xanthan gum. Then, the injection performance, resistance-increasing performance, and profile improvement effect of TRPS were evaluated through core injectability experiments, and the optimal injection concentration was determined. Finally, the enhanced oil recovery effects of water-assisted CO 2 flooding and TRPS-assisted CO 2 flooding were compared. The research in this paper can provide the experimental basis and data support for agent selection, performance evaluation, and injection parameter optimization in the application process of polymer-assisted CO 2 EOR of low-permeability reservoirs.

Material
Polymers: The injectivity performance of polymer solution in the low-permeability reservoirs can only be satisfied when the polymer molecular weight and solution viscosity are low enough. Low-molecular-weight polymers and polymeric surfactants can be used as two typical low-molecular and low-viscosity polymer systems. Here, xanthan gum (molecular weight is about 5 × 10 6 Da) and a kind of CO 2 -responsive temperatureresistant polymeric surfactant (molecular weight is about 10 6 Da) are selected as representatives of two polymer systems to conduct the following research. TRPS with certain interfacial activity and viscosity-increasing properties was obtained by grafting functional functional groups onto the main chain of polyacrylamide. Meanwhile, it has the CO 2 corresponding characteristics and can significantly increase the viscosity of the solution in a CO 2 environment.
Liquids: The inorganic salt purchased from Shanghai Aladdin Reagent Co. (Shanghai, China) was proportionally added to the deionized water (DI water) to prepare the simulated formation water, and the total salinity was 7000 mg/L. DI water was prepared by UPT-I-10T Ultra-pure Water Purifier from Chengdu Youpu Super Pure Technology Co. (Chengdu, China). The simulated oil is white oil purchased from, Shanghai Aladdin Co. (Shanghai, China), with a viscosity of 15.5 cP at 80°C (target reservoir temperature).
Cores: Cylindrical Berea cores with a size of 3.8 × 30 cm were used for polymerassisted CO 2 flooding experiments, and the gas permeabilities are about 5 mD, 10 mD, and 20 mD, respectively. The effective permeability of each core was specifically tested before each experiment.

Viscosity-Increasing Properties
The viscosity-increasing property of a polymer refers to the ability of a polymer to dissolve in water to increase the viscosity of the aqueous phase and is the most critical parameter for evaluating a polymer agent. Use DI water to prepare XG and TRPS mother liquor with a concentration of 5000 mg/L, and stir for 2 h with an electronic stirrer at 400 rpm. Then, the mother liquid could be diluted using simulated formation water to prepare the target polymer solution (100 mg/L, 300 mg/L, 500 mg/L, 700 mg/L, and 900 mg/L). The viscosities of the polymer solutions will be tested by a Brookfield viscometer after stirring for 1 h with an electronic stirrer at 200 rpm. The shear rate of the Brookfield viscometer is 7.4 1/s, and the test temperature is 80 • C.
Next, 100 mL of the XG and TRPS solutions with the target concentration were poured into the Warring agitator, respectively, sheared at 3500 rpm for 1 min, and then we tested the viscosity of the solution again and calculated the viscosity retention rate.

Polymer Rheological Properties
The viscosity of polymer solution decreases with the increase in shear rate, which has a typical shear thinning property. Evaluating the viscosity of polymers at different shear rates is of great significance for testing their EOR efficiency. The rheological curves of XG and TRPS solutions at concentrations of 100 mg/L, 300 mg/L, 500 mg/L, and 1000 mg/L were tested using a Haake rheometer (HAAKE RS6000, Thermo Electron Karlsruhe GmbH, Karlsruhe, Germany), and the shear rate was 0.1 1/s to 1000 1/s, and the test temperature was 80°C.

Polymer Reservoir Adaptability
After the polymer is injected into the reservoir, it needs to stay in the reservoir environment for a long time, and its performance under this condition determines its effect. The temperature resistance and CO 2 resistance performance requirements of the polymer solution are evaluated by the viscosity retention rate to meet the application conditions of the oil reservoir. Put XG and TRPS solution with a concentration of 200 mg/L, 500 mg/L, 800 mg/L, 1000 mg/L, and 1500 mg/L in a high-temperature aging tank separately and aged in an oven at 80 • C for 30 days. Samples were taken at regular intervals (2, 4, 6, 8, 10, 15, 20, and 30 d) to test the viscosities using a Brookfield viscometer. In addition, injecting CO 2 into piston containers with polymer solutions of 300 mg/L, 500 mg/L, and 800 mg/L until the pressure reaches 10 MPa. Then, seal the piston and place it at 80°C for ten days for a certain period to test the viscosity of the solution.

The Injectivity Ability of TRPS
The injectivity performance of polymer solution is mainly quantitatively evaluated by Resistance Factor (R F ) and Residual Resistance Factor (R FF ). The specific experimental procedure is as follows: (1) After the core was vacuumed for 3 h, it was saturated with simulated water by self-priming for more than 4 h, and the porosity was calculated; (2) According to Figure 1a, the simulated formation water was injected with the ISCO pump at a constant rate, and the pressure difference between the two ends of the core was recorded as ∆P 1 after the pressure became stable, and the core water permeability was calculated using Darcy law; (3) The ISCO pump was used to inject TRPS solution at a constant rate until the pressure is stable, and the pressure difference of the core was recorded as ∆P 2 ; and (4) The ISCO pump was used to inject simulated formation water at a constant rate until the pressure was stable, and the pressure difference at both ends of the core was recorded as ∆P 3 . The injection rate is 0.3 mL/min, and the TRPS solution needs to be injected continuously for at least 2 PV. Finally, the FR = ∆P 2 /∆P 1 and the FRR = ∆P 3 /∆P 1 can be calculated. The specific experimental scheme is shown in Table 1.
In addition, a zeta potential analyzer (Zetasizer Nano ZS, Malvern Panalytical, malvern city, England) was used to test the hydrodynamic size of TRPS solutions with different concentrations, and the Poiseuille formula was used to calculate the mean pore throat size of the cores. The injectivity performance results of TRPS can be analyzed by comparing the above two sizes.

Resistance-Increasing Performance of TRPS-Assisted CO2 Flooding
After clarifying the injectivity performance of TRPS, it is necessary to evaluate the resistance-increasing effect of TRPS in the presence of CO2. Evaluation is also carried out by RF and RFF. The specific experimental procedure is the same as that in Section 2.3.1, except that step (3) is to inject CO2 and TRPS solution into the core simultaneously using the ISCO pump. The specific experimental scheme is shown in Table 2. Meanwhile, it is necessary to evaluate the resistance-increasing performance of the TRPS-assisted CO2 flooding process. Berea cores with a size of 3.8 cm × 20 cm and a permeability of 5 mD, 10 mD, and 20 mD were selected to evaluate the resistance-increasing effect of TRPS. The specific experimental process is as follows: (1) The core was vacuumed for 3 h and then saturated with water for more than 4 h by self-priming and calculating the porosity; (2) Connected the experiment flow chart in Figure 1a and carried out water flooding and gas flooding at a constant rate using the ISCO pump until the pressure is stable. Recorded the stable pressure P0 as the basic gas flooding pressure; (3) Injected TRPS

Resistance-Increasing Performance of TRPS-Assisted CO 2 Flooding
After clarifying the injectivity performance of TRPS, it is necessary to evaluate the resistance-increasing effect of TRPS in the presence of CO 2 . Evaluation is also carried out by RF and RFF. The specific experimental procedure is the same as that in Section 2.3.1, except that step (3) is to inject CO 2 and TRPS solution into the core simultaneously using the ISCO pump. The specific experimental scheme is shown in Table 2. Meanwhile, it is necessary to evaluate the resistance-increasing performance of the TRPS-assisted CO 2 flooding process. Berea cores with a size of 3.8 cm × 20 cm and a permeability of 5 mD, 10 mD, and 20 mD were selected to evaluate the resistance-increasing effect of TRPS. The specific experimental process is as follows: (1) The core was vacuumed for 3 h and then saturated with water for more than 4 h by self-priming and calculating the porosity; (2) Connected the experiment flow chart in Figure 1a and carried out water flooding and gas flooding at a constant rate using the ISCO pump until the pressure is stable. Recorded the stable pressure P 0 as the basic gas flooding pressure; (3) Injected TRPS solutions with concentrations of 200 mg/L, 500 mg/L, and 1000 mg/L at a constant rate, and recorded the stable pressure P 1 , respectively; (4) Carried out gas flooding again at a constant injection rate to record the pressure P 1 before gas channeling; and (5) Repeated Step (3) Step (4) multiple times to obtain the pressures P i and P i . TRPS resistance increase coefficient η (i) = P i /P 0 and resistance increase coefficient of gas flooding η (i) = P i /P 0 .

Profile Control Performance of TRPS
The 3.8 cm × 20 cm Berea cores with permeability of 5 mD and 20 mD were connected in parallel to simulate reservoir heterogeneity (permeability difference ratio is 4) to carry out the TRPS-assisted CO 2 flooding experiment without oil. The specific experimental procedure is as follows: (1) After the core is evacuated for 3 h, it is saturated with simulated water by self-priming for more than 4 h, and the porosity is calculated; (2) The simulated formation water was injected at a constant rate by an ISCO pump, and the pressure at both ends of the core is recorded after the pressure is stable. The stable pressure difference is recorded as ∆P 1 , and the core water permeability is calculated by Darcy law; (3) According to Figure 1b, the experimental process was connected and simulated water and CO 2 were co-injected using the ISCO pump. Recorded the liquid production conditions and injection pressure of the cores; (4) TRPS and CO 2 were co-injected using the ISCO pump at a constant rate until the pressure is stable; and recorded the liquid production conditions and injection pressures of the two cores. The combined injection rate of liquid and gas is 0.6 mL/min (1:1 volume ratio), and if there is no CO 2 injection, the TRPS injection rate is 0.6 mL/min. The specific experimental scheme is shown in Table 3.

EOR Efficiency of TRPS-Assisted CO 2 Flooding
The oil displacement experiments were carried out using Berea cores (gas permeability is 10 mD and the water permeability is 5.19 mD). Comparing the oil recovery and pressure change law of water-assisted CO 2 flooding and TRPS-assisted CO 2 flooding, the advantages of TRPS-assisted CO 2 flooding in improving oil recovery were clarified, and the best effect period was determined. The specific experimental procedure is as follows: (1) After the core is evacuated for 3 h, it is saturated with water by self-priming for more than 4 h, and the porosity is calculated; (2) The simulated formation water is injected at a constant rate by an ISCO pump, and the pressure at both ends of the core is recorded after the pressure is stable. The pressure difference is ∆P 1 , and the core water permeability is calculated by Darcy law; (3) According to the connection experiment process in Figure 1a, the ISCO pump is used to inject crude oil into the core at a constant rate until there is no water production, and the oil saturation is calculated; (4) Simulated formation water was injected with ISCO pump at a constant rate until the water cut reaches 90% and then switched to CO 2 flooding until gas channeling without producing oil; (5) Injected TRPS solution (1000 mg/L) with ISCO pump at a constant rate of 0.2 PV and then switched to CO 2 flooding until no oil is made again, this is a round of TRPS-assisted CO 2 flooding process; (6) Repeated step (5) for a total of three times, that is, three rounds of TRPS-assisted CO 2 flooding are completed. During the experiment, the fluid production and injection pressure of the core were recorded. The injection rate of liquid and gas was 0.3 mL/min. During the experiment, a back pressure of 10 MPa was applied at the outlet end of the core by a back pressure valve. Figure 2 shows the relationship curves of the viscosity of XG and TRPS polymer solutions with the change of solution concentration at 80°C. The results show that both polymers have certain viscosity-increasing properties in terms of demand in the field of oil and gas field development engineering, and the viscosity-increasing effect of XG is more significant. When the solution concentration exceeds 300 mg/L, the viscosity-increasing effect of the XG solution is significantly improved, much higher than that of TRPS at the same concentration. When the solution concentration was less than 1000 mg/L, the viscosity-increasing effect of TRPS gradually increased with the increase in the solution concentration, and the change was stable. When the concentration is 1000 mg/L, the viscosities of XG and TRPS solutions are 15.21 cP and 4.57 cP, respectively. Within this concentration range, XG and TRPS have the potential of low-permeability reservoir injectability [36], meeting the requirements of follow-up research. In addition, the viscosity of TRPS solutions with concentrations of 700 mg/L and 1000 mg/L is similar to that of XG solutions with concentrations of 100 mg/L and 300 mg/L, respectively. The performance of the two solutions at the same viscosity will be compared through the above two concentrations.

Static Performance Comparison between XG and TRPS
3.1.1. Viscosity-Increasing Performance Figure 2 shows the relationship curves of the viscosity of XG and TRPS polymer solutions with the change of solution concentration at 80 ℃. The results show that both polymers have certain viscosity-increasing properties in terms of demand in the field of oil and gas field development engineering, and the viscosity-increasing effect of XG is more significant. When the solution concentration exceeds 300 mg/L, the viscosity-increasing effect of the XG solution is significantly improved, much higher than that of TRPS at the same concentration. When the solution concentration was less than 1000 mg/L, the viscosity-increasing effect of TRPS gradually increased with the increase in the solution concentration, and the change was stable. When the concentration is 1000 mg/L, the viscosities of XG and TRPS solutions are 15.21 cP and 4.57 cP, respectively. Within this concentration range, XG and TRPS have the potential of low-permeability reservoir injectability [36], meeting the requirements of follow-up research. In addition, the viscosity of TRPS solutions with concentrations of 700 mg/L and 1000 mg/L is similar to that of XG solutions with concentrations of 100 mg/L and 300 mg/L, respectively. The performance of the two solutions at the same viscosity will be compared through the above two concentrations.   Figure 3c. The viscosity retention rates of the two polymer solutions after shearing are both above 80%. Whether the viscosity is the same or the concentration is the same, the viscosity retention rate of TRPS is greater than that of XG because the viscosity-increasing mechanism of XG and TRPS mainly includes: (1) the polymer molecules in water are entangled to form a structure; (2) The hydrophilic groups in the polymer chains are solvated in water, and the apparent molecular volume of the polymer increases. XG has a more considerable molecular weight, and the larger molecular aggregates formed are more easily damaged by shearing.   Figure 3c. The viscosity retention rates of the two polymer solutions after shearing are both above 80%. Whether the viscosity is the same or the concentration is the same, the viscosity retention rate of TRPS is greater than that of XG because the viscosity-increasing mechanism of XG and TRPS mainly includes: (1) the polymer molecules in water are entangled to form a structure; (2) The hydrophilic groups in the polymer chains are solvated in water, and the apparent molecular volume of the polymer increases. XG has a more considerable molecular weight, and the larger molecular aggregates formed are more easily damaged by shearing.  showing typical shear thinning characteristics. The shear rates corresponding to the rheological curves of XG and TRPS solutions reaching the second Newton zone are about 10 1/s and 5 1/s, respectively, and the molecular conformation transition has reached the limit value. Because in the practice of oil field production, the shear rate of the polymer system in the underground is about 1-10 1/s, combined with the rheological curve, it is easier for TRPS to reach the second Newton zone in this shear rate range. The adaptability to the reservoir shearing of TRPS is more robust. Meanwhile, comparing the rheological curves of TRPS (1000 mg/L) and XG (300 mg/L) solutions with the same viscosity, it can be found that the viscosity of XG solution is higher at low shear rate, and the viscosity of the two is similar in the second Newton zone. This is mainly because the molecular weight of XG is large, and the viscosity-increasing property of the molecular coil is stronger at a lower shear rate, and as the shear rate increases, the difference between the two molecular weights for the viscosity-increasing performance decreases.  Figure 4 shows the rheological curves of XG and TRPS with four concentrations. Figure 4 shows that the rheological curves of XG and TRPS solutions are power-law curves, showing typical shear thinning characteristics. The shear rates corresponding to the rheological curves of XG and TRPS solutions reaching the second Newton zone are about 10 1/s and 5 1/s, respectively, and the molecular conformation transition has reached the limit value. Because in the practice of oil field production, the shear rate of the polymer system in the underground is about 1-10 1/s, combined with the rheological curve, it is easier for TRPS to reach the second Newton zone in this shear rate range. The adaptability to the reservoir shearing of TRPS is more robust. Meanwhile, comparing the rheological curves of TRPS (1000 mg/L) and XG (300 mg/L) solutions with the same viscosity, it can be found that the viscosity of XG solution is higher at low shear rate, and the viscosity of the two is similar in the second Newton zone. This is mainly because the molecular weight of XG is large, and the viscosity-increasing property of the molecular coil is stronger at a lower shear rate, and as the shear rate increases, the difference between the two molecular weights for the viscosity-increasing performance decreases.  3.1.3. Temperature-Resistance and CO2-Resistance Performance Figure 5 compares the viscosity retention curves of XG and TRPS solutions after aging in high-temperature and CO2 environments. Figure 5a shows that high temperature has a more significant effect on the viscosity of XG solution but has little impact on the viscosity of TRPS. The viscosity of the XG solution decreased rapidly with the increase in aging time, then reduced steadily after aging for ten days, and the viscosity retention rate after aging for 30 days was less than 20%. The viscosity of the TRPS solution decreased slowly with the increase in aging time, and the final viscosity retention rate was about 80%. This is because high-temperature conditions will gradually untangle the initially entangled polymer coils, and long-term aging will cause further breakage of the stretched molecular bonds, resulting in a significant decrease in the viscosity of the solution. However, TRPS has a temperature-resistant monomer and a short molecular coil, showing good temperature resistance.   Figure 5a shows that high temperature has a more significant effect on the viscosity of XG solution but has little impact on the viscosity of TRPS. The viscosity of the XG solution decreased rapidly with the increase in aging time, then reduced steadily after aging for ten days, and the viscosity retention rate after aging for 30 days was less than 20%. The viscosity of the TRPS solution decreased slowly with the increase in aging time, and the final viscosity retention rate was about 80%. This is because high-temperature conditions will gradually untangle the initially entangled polymer coils, and long-term aging will cause further breakage of the stretched molecular bonds, resulting in a significant decrease in the viscosity of the solution. However, TRPS has a temperatureresistant monomer and a short molecular coil, showing good temperature resistance.  3.1.3. Temperature-Resistance and CO2-Resistance Performance Figure 5 compares the viscosity retention curves of XG and TRPS solutions after aging in high-temperature and CO2 environments. Figure 5a shows that high temperature has a more significant effect on the viscosity of XG solution but has little impact on the viscosity of TRPS. The viscosity of the XG solution decreased rapidly with the increase in aging time, then reduced steadily after aging for ten days, and the viscosity retention rate after aging for 30 days was less than 20%. The viscosity of the TRPS solution decreased slowly with the increase in aging time, and the final viscosity retention rate was about 80%. This is because high-temperature conditions will gradually untangle the initially entangled polymer coils, and long-term aging will cause further breakage of the stretched molecular bonds, resulting in a significant decrease in the viscosity of the solution. However, TRPS has a temperature-resistant monomer and a short molecular coil, showing good temperature resistance.   Figure 5b shows that the presence of CO 2 significantly reduces the viscosity of the XG solution, and the viscosity retention rate drops to about 20% after aging for ten days. However, after aging in a CO 2 environment, the viscosity retention rate of TRPS solution increased first (up to more than 150%), then decreased slowly, and finally remained at about 120%. CO 2 will increase H + in the polymer solution and destroy the biomolecular chains in the XG solution, thereby losing the viscosity-increasing effect [37]. The TRPS molecular chain contains tertiary amine groups with CO 2 response characteristics. The primary tertiary amine group can undergo an acid-base neutralization reaction with dissolved CO 2 in aqueous solution to form a bicarbonate structure. The tertiary amine group is protonated to form a quaternary salt cation structure, and the molecular chain is hydrophilic enhanced. Therefore, the presence of CO 2 leads to larger TRPS molecular aggregates and an increase in the viscosity of the system [38,39]. The specific reaction mechanism is shown in Figure 6.  Figure 5b shows that the presence of CO2 significantly reduces the viscosity of the XG solution, and the viscosity retention rate drops to about 20% after aging for ten days. However, after aging in a CO2 environment, the viscosity retention rate of TRPS solution increased first (up to more than 150%), then decreased slowly, and finally remained at about 120%. CO2 will increase H + in the polymer solution and destroy the biomolecular chains in the XG solution, thereby losing the viscosity-increasing effect [37]. The TRPS molecular chain contains tertiary amine groups with CO2 response characteristics. The primary tertiary amine group can undergo an acid-base neutralization reaction with dissolved CO2 in aqueous solution to form a bicarbonate structure. The tertiary amine group is protonated to form a quaternary salt cation structure, and the molecular chain is hydrophilic enhanced. Therefore, the presence of CO2 leads to larger TRPS molecular aggregates and an increase in the viscosity of the system [38,39]. The specific reaction mechanism is shown in Figure 6. In addition, it can be seen from Figure 5 that regardless of the concentration of the TRPS solution, the viscosity retention rate after high-temperature aging and CO2 environment aging is higher than that of XG. This also shows that under the same viscosity, the temperature resistance and CO2 resistance of the TRPS solution are better than XG. Figure 7 shows the TRPS injection pressure curves under three concentrations and three core permeabilities. The injection pressure increased significantly during the TRPS flooding, and it decreased slightly and then remained stable in the subsequent water displacement stage. It also showed a certain resistance-increasing ability under the condition of no CO2. RF and RFF increase with the increase in TRPS concentration and the decrease in core permeability. When the concentration of TRPS is fixed at 1000 mg/L, RF and RFF in 5 mD core are 1.36 and 1.28, respectively, indicating good injectivity. However, the RF and RFF of 1000 mg/L TRPS solution in 20 mD cores are both less than 1.1, which means that although its resistance-increasing effect is poor, its injectivity is good. RF and RFF of 500 mg/L, 1000 mg/L, and 1500 mg/L TRPS solutions in 10 mD cores are, respectively, distributed between 1.08-1.27 and 1.01-1.15, indicating that the concentration of TRPS solution will not significantly influence its injectivity. When the core permeability is greater than 5 mD, and the TRPS solution concentration is less than 1500 mg/L, RFF is close to 1, indicating that the TRPS solution has less adsorption and retention in the formation, less damage to the formation, and has good injection performance. In addition, it can be seen from Figure 5 that regardless of the concentration of the TRPS solution, the viscosity retention rate after high-temperature aging and CO 2 environment aging is higher than that of XG. This also shows that under the same viscosity, the temperature resistance and CO 2 resistance of the TRPS solution are better than XG. Figure 7 shows the TRPS injection pressure curves under three concentrations and three core permeabilities. The injection pressure increased significantly during the TRPS flooding, and it decreased slightly and then remained stable in the subsequent water displacement stage. It also showed a certain resistance-increasing ability under the condition of no CO 2 . R F and R FF increase with the increase in TRPS concentration and the decrease in core permeability. When the concentration of TRPS is fixed at 1000 mg/L, R F and R FF in 5 mD core are 1.36 and 1.28, respectively, indicating good injectivity. However, the R F and R FF of 1000 mg/L TRPS solution in 20 mD cores are both less than 1.1, which means that although its resistance-increasing effect is poor, its injectivity is good. R F and R FF of 500 mg/L, 1000 mg/L, and 1500 mg/L TRPS solutions in 10 mD cores are, respectively, distributed between 1.08-1.27 and 1.01-1.15, indicating that the concentration of TRPS solution will not significantly influence its injectivity. When the core permeability is greater than 5 mD, and the TRPS solution concentration is less than 1500 mg/L, R FF is close to 1, indicating that the TRPS solution has less adsorption and retention in the formation, less damage to the formation, and has good injection performance. The core permeability is 10 mD, and the TRPS concentration is 500 mg/L, 1000 mg/L, and 1500 mg/L. Figure 8 compares the hydrodynamic size of the TRPS solution and the mean pore throat size of the core. It can be found that as the concentration of TRPS solution increases, its hydrodynamic size gradually increases from 150 nm to 350 nm, and no obvious intermolecular association occurs. As the core permeability increases, the mean pore throat size increases from 470 nm to 650 nm. The mean pore throat size of TRPS is smaller than that of the core, ensuring its good injectivity performance. Meanwhile, TRPS molecules can effectively plug pore throats through the 1/2 and 1/3 bridging theory. This can explain why TRPS in Figure 7 can effectively increase the injection pressure while ensuring a low resistance coefficient.   Figure 8 compares the hydrodynamic size of the TRPS solution and the mean pore throat size of the core. It can be found that as the concentration of TRPS solution increases, its hydrodynamic size gradually increases from 150 nm to 350 nm, and no obvious intermolecular association occurs. As the core permeability increases, the mean pore throat size increases from 470 nm to 650 nm. The mean pore throat size of TRPS is smaller than that of the core, ensuring its good injectivity performance. Meanwhile, TRPS molecules can effectively plug pore throats through the 1/2 and 1/3 bridging theory. This can explain why TRPS in Figure 7 can effectively increase the injection pressure while ensuring a low resistance coefficient. The core permeability is 10 mD, and the TRPS concentration is 500 mg/L, 1000 mg/L, and 1500 mg/L. Figure 8 compares the hydrodynamic size of the TRPS solution and the mean pore throat size of the core. It can be found that as the concentration of TRPS solution increases, its hydrodynamic size gradually increases from 150 nm to 350 nm, and no obvious intermolecular association occurs. As the core permeability increases, the mean pore throat size increases from 470 nm to 650 nm. The mean pore throat size of TRPS is smaller than that of the core, ensuring its good injectivity performance. Meanwhile, TRPS molecules can effectively plug pore throats through the 1/2 and 1/3 bridging theory. This can explain why TRPS in Figure 7 can effectively increase the injection pressure while ensuring a low resistance coefficient.   Figure 9 shows the injection pressure curves of the TRPS injection, HPAM injection, and co-injection of TRPS and CO 2 . The injection pressure rises rapidly after HPAM injection, which is much higher than that of TRPS injection, showing that even if the viscosity of the polymer solution is the same, the molecular weight of the polymer also affects its injection performance and resistance-increasing performance. The R F of HPAM injection is 7.67, which is not significant in value, but the injection pressure rises above 2.0 MPa, and there will be a problem with too high injection pressure during the oilfield application. Meanwhile, the R FF of HPAM injection is 4.45, indicating that HPAM has a large amount of retention and adsorption in the low-permeability layer, which is not conducive to the subsequent development of the formation. However, the R F and R FF of TRPS injected are 1.13 and 1.09, respectively, and the resistance-increasing effect is poor. The injection pressure during the co-injection of TRPS and CO 2 increases significantly because the more complex molecular aggregates formed after the mixing of TRPS and CO 2 significantly increase its flow resistance (Figure 6). At this time, R F and R FF rose to 2.67 and 1.23, respectively, and the effect of increasing resistance was improved. Figure 9 shows that it is difficult for conventional polymers (relatively high molecular) to achieve the coordination of injectivity and resistance-increasing in low-permeability reservoirs, while the response of TRPS and CO 2 allows it to be injected into the reservoir smoothly and achieve resistance increase inside the reservoir effect.  Figure 9 shows the injection pressure curves of the TRPS injection, HPAM injection, and co-injection of TRPS and CO2. The injection pressure rises rapidly after HPAM injection, which is much higher than that of TRPS injection, showing that even if the viscosity of the polymer solution is the same, the molecular weight of the polymer also affects its injection performance and resistance-increasing performance. The RF of HPAM injection is 7.67, which is not significant in value, but the injection pressure rises above 2.0 MPa, and there will be a problem with too high injection pressure during the oilfield application. Meanwhile, the RFF of HPAM injection is 4.45, indicating that HPAM has a large amount of retention and adsorption in the low-permeability layer, which is not conducive to the subsequent development of the formation. However, the RF and RFF of TRPS injected are 1.13 and 1.09, respectively, and the resistance-increasing effect is poor. The injection pressure during the co-injection of TRPS and CO2 increases significantly because the more complex molecular aggregates formed after the mixing of TRPS and CO2 significantly increase its flow resistance (Figure 6). At this time, RF and RFF rose to 2.67 and 1.23, respectively, and the effect of increasing resistance was improved. Figure 9 shows that it is difficult for conventional polymers (relatively high molecular) to achieve the coordination of injectivity and resistance-increasing in low-permeability reservoirs, while the response of TRPS and CO2 allows it to be injected into the reservoir smoothly and achieve resistance increase inside the reservoir effect. The injection pressures of multiple rounds of TRPS-assisted CO2 flooding with three kinds of permeability and three kinds of concentrations are collected and the resistance increase coefficients η and η′ of each round are shown in Table 4. Table 4 shows that η is greater than η′ in the same round because the resistance of the injected liquid must be greater than the resistance of the injected CO2. Under the same permeability conditions, η and η′ increased significantly after increasing the concentration of TRPS. When the concentration is 1000 mg/L, η can reach more than three times that when the concentration is 200 mg/L. η′ can reflect the resistance-increasing effect after the reaction of TRPS and CO2, The injection pressures of multiple rounds of TRPS-assisted CO 2 flooding with three kinds of permeability and three kinds of concentrations are collected and the resistance increase coefficients η and η of each round are shown in Table 4. Table 4 shows that η is greater than η in the same round because the resistance of the injected liquid must be greater than the resistance of the injected CO 2 . Under the same permeability conditions, η and η increased significantly after increasing the concentration of TRPS. When the concentration is 1000 mg/L, η can reach more than three times that when the concentration is 200 mg/L. η can reflect the resistance-increasing effect after the reaction of TRPS and CO 2 , so the relationship between η of the above rounds and the change of core permeability is plotted into a resistance-increasing coefficient chart, as shown in Figure 10. so the relationship between η′ of the above rounds and the change of core permeability is plotted into a resistance-increasing coefficient chart, as shown in Figure 10.   When the concentration of TRPS is lower than 1000 mg/L, the increase in core permeability and alternate rounds will increase η . This is because although the increase in permeability will reduce the injection pressure of TRPS-assisted CO 2 flooding, the gas channeling pressure of CO 2 flooding will also decrease, eventually increasing η . The molecular aggregates produced by the reaction of TRPS and CO 2 will be adsorbed and retained in the core, so η increases with the increase in injection rounds. When the concentration of TRPS is 1000 mg/L, the relationship curves of η change abnormally. This also shows once again the matching between TRPS and the reservoir (the low-permeability reservoir will not have unusually high pressure caused by the poor injectivity of molecular aggregates produced after the reaction of TRPS and CO 2 ) and the cumulative effect of multiple rounds of resistance-increasing performance. When the permeability is 5 mD, the molecular aggregates formed by the reaction of TRPS and CO 2 are poorly compatible with the core, resulting in an abnormal increase in injection pressure and a significant increase in η . Then, the pore throats of the 10 mD and 20 mD cores became more prominent, and the flow capacity increased, which relieved the contradiction between the poor matching between TRPS and the core, and the change law of η was normal. In addition, the irregularity of η of the TRPS solution with a concentration of 1000 mg/L increases with the increase in injection rounds, mainly in the core with a permeability of 5 mD. This is due to the increase in retention of TRPS in the core due to the decrease in permeability. Figure 11 shows the injection pressure and fractional flow rate curves of the TRPSassisted CO 2 flooding in a two-core parallel model. Figure 11 shows that the injection pressure of 1000 mg/L TRPS-assisted CO 2 flooding is higher than that of TRPS flooding separately, which also leads to a more obvious reduction in the fractional flow rate of the increased permeability layer. The injection pressure of TRPS-assisted CO 2 flooding with a concentration of 500 mg/L is lower than that of TRPS separately (1000 mg/L), showing that TRPS needs to reach a certain concentration before fully reacting with CO 2 . Combined with the resistance-increasing effect of TRPS, it can be determined that the optimal concentration of TRPS should be 1000 mg/L for reservoirs with a permeability above 5 mD and 500 mg/L for reservoirs below 5 mD. resulting in an abnormal increase in injection pressure and a significant increase in η′. Then, the pore throats of the 10 mD and 20 mD cores became more prominent, and the flow capacity increased, which relieved the contradiction between the poor matching between TRPS and the core, and the change law of η′ was normal. In addition, the irregularity of η′ of the TRPS solution with a concentration of 1000 mg/L increases with the increase in injection rounds, mainly in the core with a permeability of 5 mD. This is due to the increase in retention of TRPS in the core due to the decrease in permeability. Figure 11 shows the injection pressure and fractional flow rate curves of the TRPSassisted CO2 flooding in a two-core parallel model. Figure 11 shows that the injection pressure of 1000 mg/L TRPS-assisted CO2 flooding is higher than that of TRPS flooding separately, which also leads to a more obvious reduction in the fractional flow rate of the increased permeability layer. The injection pressure of TRPS-assisted CO2 flooding with a concentration of 500 mg/L is lower than that of TRPS separately (1000 mg/L), showing that TRPS needs to reach a certain concentration before fully reacting with CO2. Combined with the resistance-increasing effect of TRPS, it can be determined that the optimal concentration of TRPS should be 1000 mg/L for reservoirs with a permeability above 5 mD and 500 mg/L for reservoirs below 5 mD.  Figure 12 shows the produced fluids and oil recovery curves of each stage for waterassisted CO2 flooding and TRPS-assisted CO2 flooding. CO2 can dissolve in crude oil and cause it to expand, reducing its viscosity [40,41], and has a higher recovery rate than air and N2. The oil recovery after CO2 flooding to gas channeling is about 46%, and liquidassisted CO2 flooding can effectively control the gas channeling and further improve the oil recovery. Whether it is water or TRPS solution, the enhanced oil recovery of each round of the liquid injection stage is higher than that of the CO2 injection stage of the same round. TRPS-assisted CO2 flooding enhances oil recovery mainly in rounds 1 and 2, and the EOR effects are 20.71% and 10.00%, respectively. This also led to poorer EOR effects in subsequent rounds, in which the EOR in the third round was lower than that of water-assisted CO2 flooding. The main effect period of water-assisted CO2 flooding is in the first three rounds, but the overall EOR effect is 8.21% lower than that of TRPS-assisted CO2 flooding. The ultimate recovery of TRPS-assisted CO2 flooding and water-assisted CO2 flooding can reach 78.93% and 71.07%, respectively.  Figure 12 shows the produced fluids and oil recovery curves of each stage for waterassisted CO 2 flooding and TRPS-assisted CO 2 flooding. CO 2 can dissolve in crude oil and cause it to expand, reducing its viscosity [40,41], and has a higher recovery rate than air and N 2 . The oil recovery after CO 2 flooding to gas channeling is about 46%, and liquid-assisted CO 2 flooding can effectively control the gas channeling and further improve the oil recovery. Whether it is water or TRPS solution, the enhanced oil recovery of each round of the liquid injection stage is higher than that of the CO 2 injection stage of the same round. TRPS-assisted CO 2 flooding enhances oil recovery mainly in rounds 1 and 2, and the EOR effects are 20.71% and 10.00%, respectively. This also led to poorer EOR effects in subsequent rounds, in which the EOR in the third round was lower than that of water-assisted CO 2 flooding. The main effect period of water-assisted CO 2 flooding is in the first three rounds, but the overall EOR effect is 8.21% lower than that of TRPS-assisted polymer-assisted CO2 displacement, the TRPS used in this paper has a significantly better EOR effect. In the literature review, only two scenarios obtained higher EOR than this paper (the bolded items in Appendix A Table A1). The first scenario is for heavy oil reservoirs, where CO2 or WAG recovery is very low due to the extremely unfavorable mobility ratio, so the effect of polymer-assisted CO2 displacement is remarkable. The second scenario is gel particle-assisted CO2 flooding, and the research target is parallel cores. Because the profile improvement effect of gel particles is stronger than that of polymers, its EOR effect in heterogeneous reservoirs is remarkable. Therefore, the TRPS-assisted CO2 flooding in this paper has a considerable EOR effect of 32.93% compared with CO2 flooding and 8.21% compared with water-assisted CO2 flooding.  Figure 13 is the pressure difference comparison curve of water-assisted CO2 flooding and TRPS-assisted CO2 flooding. The increase in injection pressure during the liquid injection is more significant than that of the gas injection, which can also explain the better EOR effect of each round of liquid injection in Figure 12. For water-assisted CO2 flooding, the gas-injection pressure in each round is close to the initial gas channeling pressure, indicating that gas-water alternation can increase the flow resistance of two-phase flow to a certain extent, but the gas channeling will quickly occur and form a dominant channel again. However, the CO2 injection pressure of each round of TRPS-assisted CO2 flooding is significantly higher than that of water-assisted CO2 flooding because the molecular aggregates produced by TRPS can effectively increase the flow resistance and adsorption  Table A1 compares the EOR effects of different CO 2 enhancement injections. It can be found that compared with the ultrasonic physical method and another polymer-assisted CO 2 displacement, the TRPS used in this paper has a significantly better EOR effect. In the literature review, only two scenarios obtained higher EOR than this paper (the bolded items in Appendix A Table A1). The first scenario is for heavy oil reservoirs, where CO 2 or WAG recovery is very low due to the extremely unfavorable mobility ratio, so the effect of polymer-assisted CO 2 displacement is remarkable. The second scenario is gel particle-assisted CO 2 flooding, and the research target is parallel cores. Because the profile improvement effect of gel particles is stronger than that of polymers, its EOR effect in heterogeneous reservoirs is remarkable. Therefore, the TRPS-assisted CO 2 flooding in this paper has a considerable EOR effect of 32.93% compared with CO 2 flooding and 8.21% compared with water-assisted CO 2 flooding. Figure 13 is the pressure difference comparison curve of water-assisted CO 2 flooding and TRPS-assisted CO 2 flooding. The increase in injection pressure during the liquid injection is more significant than that of the gas injection, which can also explain the better EOR effect of each round of liquid injection in Figure 12. For water-assisted CO 2 flooding, the gas-injection pressure in each round is close to the initial gas channeling pressure, indicating that gas-water alternation can increase the flow resistance of two-phase flow to a certain extent, but the gas channeling will quickly occur and form a dominant channel again. However, the CO 2 injection pressure of each round of TRPS-assisted CO 2 flooding is significantly higher than that of water-assisted CO 2 flooding because the molecular aggregates produced by TRPS can effectively increase the flow resistance and adsorption time in the channeling channel. Figure 13 shows that, whether water or TRPS solution, the injection pressure of each round of the water injection stage and the end of the CO 2 injection stage is significantly higher than that of the previous round, which has a significant cumulative effect. The higher injection pressure is the main reason for the quicker effect of TRPS-assisted CO 2 flooding and better EOR effect. The injection pressure of liquid-assisted CO 2 flooding will continue to increase with the injection round, but the EOR will drop significantly after the third round, so the optimal development round should be three. In addition, the increased injection pressure difference of liquid-assisted CO 2 displacement can not only expand the swept volume [42], but also effectively promote the dissolution of CO 2 into crude oil and reduce its viscosity. The high-pressure CO 2 environment will reduce the viscosity of the TRPS solution, but the viscosity of crude oil will be reduced to a greater extent, so TRPS-assisted CO 2 flooding has good EOR potential.  Figure 13 shows that, whether water or TRPS solution, the injection pressure of each round of the water injection stage and the end of the CO2 injection stage is significantly higher than that of the previous round, which has a significant cumulative effect. The higher injection pressure is the main reason for the quicker effect of TRPS-assisted CO2 flooding and better EOR effect. The injection pressure of liquid-assisted CO2 flooding will continue to increase with the injection round, but the EOR will drop significantly after the third round, so the optimal development round should be three. In addition, the increased injection pressure difference of liquid-assisted CO2 displacement can not only expand the swept volume [42], but also effectively promote the dissolution of CO2 into crude oil and reduce its viscosity. The high-pressure CO2 environment will reduce the viscosity of the TRPS solution, but the viscosity of crude oil will be reduced to a greater extent, so TRPS-assisted CO2 flooding has good EOR potential.

Conclusions
This paper investigated the improvement of CO2 channeling in low-permeability reservoirs by polymer solutions. Through the experimental evaluation of the static properties, flow properties, resistance-increasing effect, and EOR effect of the polymer solution, the polymer surfactant (TRPS) with CO2 response was selected as the assisted agent, which has a good application effect. The specific conclusions are as follows: (1) TRPS has a certain viscosity-increasing property, and the viscosity of the solution is 4.57 cP when the concentration is 1000 mg/L. It has good temperature resistance, and its viscosity retention rate is above 80% after aging at 80 °C for 30 days. In addition, TRPS can react with CO2 to increase the size of molecular aggregates significantly, and the viscosity retention rate is about 120% after aging in a CO2 environment for 10 days.

Conclusions
This paper investigated the improvement of CO 2 channeling in low-permeability reservoirs by polymer solutions. Through the experimental evaluation of the static properties, flow properties, resistance-increasing effect, and EOR effect of the polymer solution, the polymer surfactant (TRPS) with CO 2 response was selected as the assisted agent, which has a good application effect. The specific conclusions are as follows: (1) TRPS has a certain viscosity-increasing property, and the viscosity of the solution is 4.57 cP when the concentration is 1000 mg/L. It has good temperature resistance, and its viscosity retention rate is above 80% after aging at 80 • C for 30 days. In addition, TRPS can react with CO 2 to increase the size of molecular aggregates significantly, and the viscosity retention rate is about 120% after aging in a CO 2 environment for 10 days.
(2) The R F and R FF of the TRPS solution with a concentration of 1000 mg/L in 5 mD cores are 1.36 and 1.28, respectively, showing good injectivity performance. Increasing the concentration of TRPS to 1500 mg/L had little effect on its injectivity performance. (3) The injection pressure of TRPS and CO 2 co-injection is between the injection of HPAM with the same viscosity and the injection of TRPS solution alone, which has good flow performance and resistance-increasing effect. The η and η of TRPS-assisted CO 2 flooding increase with increased permeability, concentration of TRPS solution, and injection rounds. When the permeability is 5 mD, the base pressure of gas channeling is high, which will reduce the matching system between TRPS and the reservoir, thus affecting the change law of η . (4) The effect of TRPS solution on profile improvement: 1000 mg/L TPRS + CO 2 > 1000 mg/L TRPS > 500 mg/L TRPS + CO 2 , considering the η and profile improvement effect, the application concentration of TRPS should be 1000 mg/L. (5) The EOR effect of TPRS-assisted CO 2 flooding is 8.21% higher than that of waterassisted CO 2 flooding. The EOR effect of TRPS-assisted CO 2 displacement is mainly reflected in the first to second rounds, while the EOR effect of water-assisted CO 2 displacement is primarily reflected in the first to third rounds. The injection pressure of liquid-assisted CO 2 flooding has a cumulative impact of multiple rounds, so the optimal injection round is 3.