Effect of CO2 Partial Pressure on the Corrosion Behavior of J55 Carbon Steel in 30% Crude Oil/Brine Mixture

The influence of CO2 partial pressure on the corrosion properties, including corrosion rate, morphology, chemical composition, and corrosion depth, of J55 carbon steel in 30% crude oil/brine at 65 °C was investigated. A corrosion mechanism was then proposed based on the understanding of the formation of localized corrosion. Results showed that localized corrosion occurred in 30% crude oil/brine with CO2. The corrosion rate sharply increased as the CO2 partial pressure (Pco2) was increased from 0 to 1.5 MPa, decreased from Pco2 = 1.5 MPa to Pco2 = 5.0 MPa, increased again at Pco2 = 5.0 MPa, and then reached a constant value after Pco2 = 9.0 MPa. The system pH initially decreased, rapidly increased, and then stabilized as CO2 partial pressure was increased. In the initial period, the surface of J55 carbon steel in the CO2/30% crude oil/brine mixtures showed intense corrosion. In conclusion, CO2 partial pressure affects the protection performance of FeCO3 by changing the formation of corrosion scale and further affecting the corrosion rate.


Introduction
In recent years, the carbon dioxide flooding enhanced oil recovery (CO 2 -EOR) technology has been widely applied worldwide [1][2][3] and has made a positive contribution to the geological reserves of carbon. However, CO 2 -EOR is expected to significantly increase the corrosion failure risk of tubes [4,5]. The acceptable rate of wellbore corrosion in China is less than 0.076 mm·year −1 [6], and the qualitative categorization of carbon steel corrosion rates for oil production systems in the US includes low (<0.025 mm·year −1 ), moderate (0.025-0.12 mm·year −1 ), high (0.13-0.25 mm·year −1 ), and severe (>0.25 mm·year −1 ) [7]. When water cut is greater than 50%, the corrosion rates of carbon steel (API 5CT L80) and P110 steel are 3.4-34.2 and 0.03-5.0 mm·year −1 , respectively [8,9], which are far beyond the acceptable range. Thus, many studies have focused on CO 2 corrosion, especially on the effect of environment on corrosion.
Mass loss during CO 2 corrosion is generally related to environmental conditions, such as temperature, pressure, salt concentration, solution pH, and CO 2 partial pressure. CO 2 partial pressure and protective scale considerably impact corrosion rate. Many studies demonstrated that the CO 2 corrosion rate of carbon steel increases with increasing CO 2 pressure [10][11][12]. The concentration of H 2 CO 3 increases as CO 2 partial pressure increases, which accelerates the cathodic reactions and increases the corrosion rate [13][14][15][16]. CO 2 partial pressure affects the protective properties and components of the corrosion product layer by changing the system pH. Other studies [15][16][17][18][19][20][21] indicated In CO 2 -EOR, gas channelling often occurs [22,23], during which the CO 2 partial pressure rises from the bottom hole to no more than 15 MPa in the C 8 reservoir. Corrosion test was carried out in the PARR-4578 autoclave (Parr Instrument Company, Champaign, IL, USA) by using the weight-loss method, and the schematic is shown in Figure 1. A 1 L aliquot of the mixture of 30% crude oil/brine was added to the autoclave, and the dissolved oxygen was purged in the solution with a small amount of nitrogen gas for 4 h under a pressure of 0.5 MPa [13] and a temperature of 65 • C. The autoclave was pressured with pure N 2 gas to the experimental values (total pressure value-CO 2 partial pressure) and with CO 2 gas to a total pressure value of 15 MPa for 2 days at the running speed of 0.5 m·s −1 (200 r·min −1 ). In CO2-EOR, gas channelling often occurs [22,23], during which the CO2 partial pressure rises from the bottom hole to no more than 15 MPa in the C8 reservoir. Corrosion test was carried out in the PARR-4578 autoclave (Parr Instrument Company, Champaign, IL, USA) by using the weight-loss method, and the schematic is shown in Figure 1. A 1 L aliquot of the mixture of 30% crude oil/brine was added to the autoclave, and the dissolved oxygen was purged in the solution with a small amount of nitrogen gas for 4 h under a pressure of 0.5 MPa [13] and a temperature of 65 °C. The autoclave was pressured with pure N2 gas to the experimental values (total pressure value-CO2 partial pressure) and with CO2 gas to a total pressure value of 15 MPa for 2 days at the running speed of 0.5 m·s −1 (200 r·min −1 ). After corrosion induction, the three corroded samples were divided into two groups for scanning electron microscope (SEM), energy dispersive spectrometer (EDS), and X-ray diffraction (XRD) analyses of the corrosion scales formed on the steel surface. After these tests, the three corroded samples were subjected to mass loss tests to determine the average corrosion rate.
The corrosion rate of the steel was determined by the mass loss technique in accordance with the ASTM (American Society for Testing Materials) G1-03-Standard practice for preparing, cleaning, and evaluating corrosion [24]. Immediately after corrosion induction, the samples were rinsed with distilled water and the crude oil on the surface was removed with acetone. Corrosion products were removed with an ultrasonic cleaner. Then, the samples were immersed in an acid cleaning solution (500 mL of HCl and 3.5 g of hexamethylenamine diluted with water to 1000 mL) for 10 min, and the corrosion products on the surface were removed. After being immersed, the samples were thoroughly washed with distilled water until the acid cleaning solution on the surface was completely removed. Then, the samples were placed in ethanol for cleaning and dehydration twice. The samples were dried in cold air, packed with filter paper, and then placed in the dryer for 4-7 h. Finally, the samples were weighed to within an accuracy of 0.1 mg. The corrosion rate was calculated as follows: (1) where rcorr is the average corrosion rate, mm·year −1 ; m is the weight of the test sheet before the experiment, g; mt is the weight of the test sheet after the experiment, g; S is the whole surface contacted with solution, cm 2 ; ρ is the density of tested steel, g·cm −3 , which is 7.86 g·cm −3 in the case of After corrosion induction, the three corroded samples were divided into two groups for scanning electron microscope (SEM), energy dispersive spectrometer (EDS), and X-ray diffraction (XRD) analyses of the corrosion scales formed on the steel surface. After these tests, the three corroded samples were subjected to mass loss tests to determine the average corrosion rate.
The corrosion rate of the steel was determined by the mass loss technique in accordance with the ASTM (American Society for Testing Materials) G1-03-Standard practice for preparing, cleaning, and evaluating corrosion [24]. Immediately after corrosion induction, the samples were rinsed with distilled water and the crude oil on the surface was removed with acetone. Corrosion products were removed with an ultrasonic cleaner. Then, the samples were immersed in an acid cleaning solution (500 mL of HCl and 3.5 g of hexamethylenamine diluted with water to 1000 mL) for 10 min, and the corrosion products on the surface were removed. After being immersed, the samples were thoroughly washed with distilled water until the acid cleaning solution on the surface was completely removed. Then, the samples were placed in ethanol for cleaning and dehydration twice. The samples were dried in cold air, packed with filter paper, and then placed in the dryer for 4-7 h. Finally, the samples were weighed to within an accuracy of 0.1 mg. The corrosion rate was calculated as follows: where r corr is the average corrosion rate, mm·year −1 ; m is the weight of the test sheet before the experiment, g; m t is the weight of the test sheet after the experiment, g; S is the whole surface contacted with solution, cm 2 ; ρ is the density of tested steel, g·cm −3 , which is 7.86 g·cm −3 in the case of carbon steel; and t is the immersion duration, h. The mean corrosion rate error was calculated using three parallel specimens in each test. The surface microstructure of the corrosion product scales on the surface of corroded samples was analyzed via SEM (FEI Quanta 600F microscope, FEI Corporation, Hillsboro, TX, USA). The elemental compositions of the corrosion product scales were estimated by EDS (OXFORD INCA energy 350, Oxford Instrument, Oxford, UK). The composition of the corroded samples was performed with XRD (Bruker D8 XRD, Bruker Corporation, Karlsruhe, Germany).
The maximum corrosion depth of the corroded samples was analyzed with an optical digital microscope (OLYMPUS DSX500, Olympus Corporation, Tokyo, Japan) after removal of the corrosion product layers by using the acid cleaning solution. Under bright-field mode, the corroded sample surface was subjected to grand horizon three dimensions (3D) image capture using adjacent visual synthetic diagram mode. The magnification was 100 times, with a 3 × 3 nine-image synthetic diagram and an overlap ratio of 10%. Four points on the front and back surfaces of the samples were collected, as shown in Figure 2. The area of the 3 × 3 nine-image synthetic diagram was 7612 µm × 7612 µm, the total area of image acquisition was 4.63 cm 2 , and 43.27% of the exposed surface area was occupied, which was much larger than that in other studies [15][16][17][18][19][20][21]. The maximum corrosion depth could be acquired by comparing the corrosion depth measured in different areas. Therefore, the method can also accurately reflect the maximum corrosion depth of the corroded samples. carbon steel; and t is the immersion duration, h. The mean corrosion rate error was calculated using three parallel specimens in each test. The surface microstructure of the corrosion product scales on the surface of corroded samples was analyzed via SEM (FEI Quanta 600F microscope, FEI Corporation, Hillsboro, TX, USA). The elemental compositions of the corrosion product scales were estimated by EDS (OXFORD INCA energy 350, Oxford Instrument, Oxford, UK). The composition of the corroded samples was performed with XRD (Bruker D8 XRD, Bruker Corporation, Karlsruhe, Germany).
The maximum corrosion depth of the corroded samples was analyzed with an optical digital microscope (OLYMPUS DSX500, Olympus Corporation, Tokyo, Japan) after removal of the corrosion product layers by using the acid cleaning solution. Under bright-field mode, the corroded sample surface was subjected to grand horizon three dimensions (3D) image capture using adjacent visual synthetic diagram mode. The magnification was 100 times, with a 3 × 3 nine-image synthetic diagram and an overlap ratio of 10%. Four points on the front and back surfaces of the samples were collected, as shown in Figure 2. The area of the 3 × 3 nine-image synthetic diagram was 7612 μm × 7612 μm, the total area of image acquisition was 4.63 cm 2 , and 43.27% of the exposed surface area was occupied, which was much larger than that in other studies [15][16][17][18][19][20][21]. The maximum corrosion depth could be acquired by comparing the corrosion depth measured in different areas. Therefore, the method can also accurately reflect the maximum corrosion depth of the corroded samples.  Figure 3 shows the macroscopic morphologies of the J55 carbon steel before corrosion test and after the removal of corrosion scales under different CO2 partial pressures. Localized corrosion occurred on the surface of the J55 carbon steel. As shown in Figure 4, the average corrosion rate of the J55 carbon steel after immersing in a CO2/crude oil/brine environment initially increased and then decreased with increasing CO2 partial pressure before finally stabilizing. When the CO2 partial pressure was increased from 0 to 1.5 MPa, the corrosion rate of J55 increased sharply. The concentration of H2CO3 increased as the partial pressure of CO2 was increased, which decreased the system pH and therefore increased the corrosion rate [13][14][15][16]. When the CO2 partial pressure was increased from 1.5 MPa to 5.0 MPa, the corrosion rate of J55 decreased. With the continuous increase in CO2 partial pressure, a protective layer gradually formed on the surface of the J55 carbon steel. When the CO2 partial pressure was increased from 5.0 to 9.0 MPa, the corrosion rate of J55 increased. The protective layer formed on the surface of J55 may be dissolved gradually, thereby increasing the corrosion rate [14,15]. When the CO2 partial pressure was increased from 9.0 MPa to 15.0 MPa, the corrosion rate of J55 was almost constant. A protective layer formed faster on the steel surface as the CO2 partial pressure was increased [17][18][19]. When CO2 dissolved in water equilibrium, CO2  Figure 3 shows the macroscopic morphologies of the J55 carbon steel before corrosion test and after the removal of corrosion scales under different CO 2 partial pressures. Localized corrosion occurred on the surface of the J55 carbon steel. As shown in Figure 4, the average corrosion rate of the J55 carbon steel after immersing in a CO 2 /crude oil/brine environment initially increased and then decreased with increasing CO 2 partial pressure before finally stabilizing. When the CO 2 partial pressure was increased from 0 to 1.5 MPa, the corrosion rate of J55 increased sharply. The concentration of H 2 CO 3 increased as the partial pressure of CO 2 was increased, which decreased the system pH and therefore increased the corrosion rate [13][14][15][16]. When the CO 2 partial pressure was increased from 1.5 MPa to 5.0 MPa, the corrosion rate of J55 decreased. With the continuous increase in CO 2 partial pressure, a protective layer gradually formed on the surface of the J55 carbon steel. When the CO 2 partial pressure was increased from 5.0 to 9.0 MPa, the corrosion rate of J55 increased. The protective layer formed on the surface of J55 may be dissolved gradually, thereby increasing the corrosion rate [14,15]. When the CO 2 partial pressure was increased from 9.0 MPa to 15.0 MPa, the corrosion rate of J55 was almost constant. A protective layer formed faster on the steel surface as the CO 2 partial pressure was increased [17][18][19]. When CO 2 dissolved in water equilibrium, CO 2 solubility almost no longer increased with increasing CO 2 partial pressure. Thus, the system pH was almost invariable [25], and the protective layer was not dissolved.

Weight Loss Tests
Materials 2018, 11, x FOR PEER REVIEW 5 of 14 solubility almost no longer increased with increasing CO2 partial pressure. Thus, the system pH was almost invariable [25], and the protective layer was not dissolved.

Microstructure and Composition of the Corrosion Scale
Figures 5-9 show the SEM images of the corrosion scales formed on the J55 steel surface as a function of CO2 partial pressure in 30% crude oil/brine mixtures at the same magnification (×100 or ×2000). EDS was performed on the corrosion product scales of the tested samples. Table 3 shows the EDS spectra of the corrosion scale in the inner surface of the blue line region in Figures 4-8, respectively. Figure 5 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 0 MPa and 65 °C. The polishing marks were still visible on the surface of the J55 steel, and no visible signs of corrosion were observed on the sample. The corrosion product mainly consisted of Fe3C (the content ratio of Fe and C atoms is about 1:3) and minor constituents of alloying elements from the carbon steel matrix. Figure 6 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 1.5 MPa and 65 °C. The surface was severely attacked and showed disperse FeCO3 and CaCO3 scales and minor constituents of alloying elements from the carbon steel matrix. Figure 7 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 5.0 MPa and 65 °C. A large part of the surface was attacked and fully covered by FeCO3 and CaCO3. Figures 8 and 9 show the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 9.0 and 15 MPa. The surface was almost covered by the protective FeCO3 layer and few CaCO3. At Pco 2 = 15.0 MPa, the corrosion product layer was thicker and denser than that at Pco 2 = 9.0 MPa. solubility almost no longer increased with increasing CO2 partial pressure. Thus, the system pH was almost invariable [25], and the protective layer was not dissolved.

Microstructure and Composition of the Corrosion Scale
Figures 5-9 show the SEM images of the corrosion scales formed on the J55 steel surface as a function of CO2 partial pressure in 30% crude oil/brine mixtures at the same magnification (×100 or ×2000). EDS was performed on the corrosion product scales of the tested samples. Table 3 shows the EDS spectra of the corrosion scale in the inner surface of the blue line region in Figures 4-8, respectively. Figure 5 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 0 MPa and 65 °C. The polishing marks were still visible on the surface of the J55 steel, and no visible signs of corrosion were observed on the sample. The corrosion product mainly consisted of Fe3C (the content ratio of Fe and C atoms is about 1:3) and minor constituents of alloying elements from the carbon steel matrix. Figure 6 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 1.5 MPa and 65 °C. The surface was severely attacked and showed disperse FeCO3 and CaCO3 scales and minor constituents of alloying elements from the carbon steel matrix. Figure 7 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 5.0 MPa and 65 °C. A large part of the surface was attacked and fully covered by FeCO3 and CaCO3. Figures 8 and 9 show the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 9.0 and 15 MPa. The surface was almost covered by the protective FeCO3 layer and few CaCO3. At Pco 2 = 15.0 MPa, the corrosion product layer was thicker and denser than that at Pco 2 = 9.0 MPa.

Microstructure and Composition of the Corrosion Scale
Figures 5-9 show the SEM images of the corrosion scales formed on the J55 steel surface as a function of CO 2 partial pressure in 30% crude oil/brine mixtures at the same magnification (×100 or ×2000). EDS was performed on the corrosion product scales of the tested samples. Table 3 shows the EDS spectra of the corrosion scale in the inner surface of the blue line region in Figures 4-8, respectively. Figure 5 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 0 MPa and 65 • C. The polishing marks were still visible on the surface of the J55 steel, and no visible signs of corrosion were observed on the sample. The corrosion product mainly consisted of Fe 3 C (the content ratio of Fe and C atoms is about 1:3) and minor constituents of alloying elements from the carbon steel matrix. Figure 6 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 1.5 MPa and 65 • C. The surface was severely attacked and showed disperse FeCO 3 and CaCO 3 scales and minor constituents of alloying elements from the carbon steel matrix. Figure 7 shows the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 5.0 MPa and 65 • C. A large part of the surface was attacked and fully covered by FeCO 3 and CaCO 3 . Figures 8 and 9 show the SEM images of the corrosion scales formed on the J55 steel surface at Pco 2 = 9.0 and 15 MPa. The surface was almost covered by the protective FeCO 3 layer and few CaCO 3 . At Pco 2 = 15.0 MPa, the corrosion product layer was thicker and denser than that at Pco 2 = 9.0 MPa.      The main elements of the corrosion products in the 30% crude oil/brine environment without CO2 were Fe and C, and the content ratio of iron and carbon atoms was about 1:3, indicating that the corrosion products consisted mainly of FeC3. The main elements of the corrosion products in the CO2/30% crude oil/brine environment were O, Fe, and Ca, indicating that the corrosion products consisted mainly of FeC3 and mixed carbonate (FexCa1−xCO3) [26,27]. At Pco 2 = 1.5 and 5.0 MPa, the   The main elements of the corrosion products in the 30% crude oil/brine environment without CO2 were Fe and C, and the content ratio of iron and carbon atoms was about 1:3, indicating that the corrosion products consisted mainly of FeC3. The main elements of the corrosion products in the CO2/30% crude oil/brine environment were O, Fe, and Ca, indicating that the corrosion products consisted mainly of FeC3 and mixed carbonate (FexCa1−xCO3) [26,27]. At Pco 2 = 1.5 and 5.0 MPa, the  The main elements of the corrosion products in the 30% crude oil/brine environment without CO 2 were Fe and C, and the content ratio of iron and carbon atoms was about 1:3, indicating that the corrosion products consisted mainly of FeC 3 . The main elements of the corrosion products in the CO 2 /30% crude oil/brine environment were O, Fe, and Ca, indicating that the corrosion products consisted mainly of FeC 3  the minor constituents of alloying elements from the carbon steel were detected, indicating that the surface was not fully covered by the corrosion product layer. At Pco 2 = 9.0 and 15.0 MPa, the minor constituents of alloying elements from the carbon steel were not detected, suggesting that the surface was fully covered by the corrosion product layer. Figure 10 shows the XRD spectra of the surface layer on the corroded samples immersed in CO 2 /30% crude oil/brine mixtures. Related research [15][16][17][18][19][20][21] reported that the main CO 2 corrosion product of carbon steel is FeCO 3 . The compositions of the corrosion product layer in the CO 2 /30% crude oil/brine mixtures were similar and mainly consisted of the complex salt of CaCO 3 and FeCO 3 . This result may be attributed to the presence of metal cation isomorphous substitution in CO 2 corrosion [27]. When the [Fe 2+ ] × [CO 3 2− ] in the medium exceeds FeCO 3 solubility product K sp (FeCO 3 ), that is, when the FeCO 3 supersaturation in the medium is the FeCO 3 would be deposited on the metal surface. As shown in the following form [27]: The Ca 2+ in the solution to the replacement in FeCO 3 crystal Fe 2+ and the formation of the Fe (Ca) CO 3 complex can be expressed as Materials 2018, 11, x FOR PEER REVIEW 8 of 14 minor constituents of alloying elements from the carbon steel were detected, indicating that the surface was not fully covered by the corrosion product layer. At Pco 2 = 9.0 and 15.0 MPa, the minor constituents of alloying elements from the carbon steel were not detected, suggesting that the surface was fully covered by the corrosion product layer. Figure 10 shows the XRD spectra of the surface layer on the corroded samples immersed in CO2/30% crude oil/brine mixtures. Related research [15][16][17][18][19][20][21] reported that the main CO2 corrosion product of carbon steel is FeCO3. The compositions of the corrosion product layer in the CO2/30% crude oil/brine mixtures were similar and mainly consisted of the complex salt of CaCO3 and FeCO3. This result may be attributed to the presence of metal cation isomorphous substitution in CO2 corrosion [27]. When the [Fe 2+ ] × [CO3 2− ] in the medium exceeds FeCO3 solubility product Ksp (FeCO3), that is, when the FeCO3 supersaturation in the medium is S = Fe 2+ × CO 3 2-/ K sp FeCO 3 > 1, the FeCO3 would be deposited on the metal surface. As shown in the following form [27]: The Ca 2+ in the solution to the replacement in FeCO3 crystal Fe 2+ and the formation of the Fe (Ca) CO3 complex can be expressed as  Figure 11 shows that the maximum corrosion depth of the cleaned sample surface exposed to 30% crude oil/brine condition at Pco 2 = 1.0 MPa and 65 °C was 237.753 μm. The maximum corrosion depths measured in the seven other regions were compared, and the maximum corrosion depth of the cleaned sample surface exposed to 30% crude oil/brine condition at Pco 2 = 1.0 MPa and 65 °C was 382.742 μm. As shown in Figure 12, the maximum corrosion depth of the cleaned sample surface exposed to 30% crude oil/brine condition at Pco 2 = 1.5 MPa and 65 °C, where the average corrosion rate was the highest, was 90.395 μm. The type of corrosion damage changed from localized corrosion to mesa corrosion. Figure 13 shows the maximum corrosion depth and penetration rate/average corrosion rate ratio of the J55 carbon steel surface after removal of the corrosion product layers by using acid cleaning solution as a function of CO2 partial pressure in the 30% crude oil/brine mixtures. The maximum corrosion depth varied with the increase in CO2 partial pressure possibly because of the protection conferred by the corrosion product layer. The variation trend of the penetration rate/average corrosion rate ratio of the J55 carbon steel surface was the same as that of the maximum corrosion depth as the CO2 partial pressure was increased. The penetration rate/average corrosion rate ratios were greater than 4, indicating that local corrosion occurred on the  Figure 11 shows that the maximum corrosion depth of the cleaned sample surface exposed to 30% crude oil/brine condition at Pco 2 = 1.0 MPa and 65 • C was 237.753 µm. The maximum corrosion depths measured in the seven other regions were compared, and the maximum corrosion depth of the cleaned sample surface exposed to 30% crude oil/brine condition at Pco 2 = 1.0 MPa and 65 • C was 382.742 µm. As shown in Figure 12, the maximum corrosion depth of the cleaned sample surface exposed to 30% crude oil/brine condition at Pco 2 = 1.5 MPa and 65 • C, where the average corrosion rate was the highest, was 90.395 µm. The type of corrosion damage changed from localized corrosion to mesa corrosion. Figure 13 shows the maximum corrosion depth and penetration rate/average corrosion rate ratio of the J55 carbon steel surface after removal of the corrosion product layers by using acid cleaning solution as a function of CO 2 partial pressure in the 30% crude oil/brine mixtures. The maximum corrosion depth varied with the increase in CO 2 partial pressure possibly because of the protection conferred by the corrosion product layer. The variation trend of the penetration rate/average corrosion rate ratio of the J55 carbon steel surface was the same as that of the maximum corrosion depth as the CO 2 partial pressure was increased. The penetration rate/average corrosion rate ratios were greater than 4, indicating that local corrosion occurred on the surface of the carbon steel [8,18]. At Pco 2 = 1.0 MPa, the corrosion depth was the largest at 382.742 µm, which corresponded to 69.8504 mm/a. This penetration rate was considerably greater than the weight-loss corrosion rate (3.3058 mm·year −1 ) shown in Figure 3, thereby confirming local attack.   Figure 13. Maximum corrosion depth and penetration rate/average corrosion rate ratio of J55 carbon steel surface as a function of CO2 partial pressure in 30% crude oil/brine mixtures.

Variation of pH with CO2 Partial Pressure
As the CO2 partial pressure was increased, the corrosion rate of the J55 carbon steel initially increased, decreased, increased again, and then stabilized. This result is different from the report of surface of the carbon steel [8,18]. At Pco 2 = 1.0 MPa, the corrosion depth was the largest at 382.742 μm, which corresponded to 69.8504 mm/a. This penetration rate was considerably greater than the weight-loss corrosion rate (3.3058 mm·year −1 ) shown in Figure 3, thereby confirming local attack.
(a) (b) (c) Figure 11. Corrosion depth analysis on cleaned surface of the sample exposed to 30% crude oil/brine condition at Pco 2 = 1.0 MPa and 65 °C: (a) corrosion morphology; (b) corrosion depth distribution contour diagram; and (c) corrosion depth distribution 3D diagram.
(a) (b) (c) Figure 12. Corrosion depth analysis on the cleaned surface of the sample exposed to 30% crude oil/brine condition at Pco 2 = 1.5 MPa and 65 °C: (a) corrosion morphology; (b) corrosion depth distribution contour diagram; and (c) corrosion depth distribution 3D diagram.  Figure 13. Maximum corrosion depth and penetration rate/average corrosion rate ratio of J55 carbon steel surface as a function of CO2 partial pressure in 30% crude oil/brine mixtures.

Variation of pH with CO2 Partial Pressure
As the CO2 partial pressure was increased, the corrosion rate of the J55 carbon steel initially increased, decreased, increased again, and then stabilized. This result is different from the report of surface of the carbon steel [8,18]. At Pco 2 = 1.0 MPa, the corrosion depth was the largest at 382.742 μm, which corresponded to 69.8504 mm/a. This penetration rate was considerably greater than the weight-loss corrosion rate (3.3058 mm·year −1 ) shown in Figure 3, thereby confirming local attack.
(a) (b) (c) Figure 11. Corrosion depth analysis on cleaned surface of the sample exposed to 30% crude oil/brine condition at Pco 2 = 1.0 MPa and 65 °C: (a) corrosion morphology; (b) corrosion depth distribution contour diagram; and (c) corrosion depth distribution 3D diagram.
(a) (b) (c) Figure 12. Corrosion depth analysis on the cleaned surface of the sample exposed to 30% crude oil/brine condition at Pco 2 = 1.5 MPa and 65 °C: (a) corrosion morphology; (b) corrosion depth distribution contour diagram; and (c) corrosion depth distribution 3D diagram.  Figure 13. Maximum corrosion depth and penetration rate/average corrosion rate ratio of J55 carbon steel surface as a function of CO2 partial pressure in 30% crude oil/brine mixtures.

Variation of pH with CO2 Partial Pressure
As the CO2 partial pressure was increased, the corrosion rate of the J55 carbon steel initially increased, decreased, increased again, and then stabilized. This result is different from the report of Figure 13. Maximum corrosion depth and penetration rate/average corrosion rate ratio of J55 carbon steel surface as a function of CO 2 partial pressure in 30% crude oil/brine mixtures.
Wiebe et al. [31,32] calculated the solubility of CO 2 in water at 12-100 • C, and assumed that the solubility of CO 2 in water was not related to the concentration of solution and pH value. Ziegler [33] fitted the solubility data of Wiebe to the following empirical formula: where S CO 2 is the total concentration of CO 2 dissolved in water, mol·kg −1 .
The first ionization of the carbonated solution is the main reaction, and K a1 is much larger than K w and K a2 . When estimating the pH of the CO 2 -H 2 O system, the ionization of water and the secondary ionization of carbonic acid can be neglected. Carbonic acid is a weak acid, and its x CO 2 is much larger than its x H + . In carbonate solution, except for CO 2(aq) , the concentration of other ions is negligible. Carbonic acid is a dilute solution with an ionic activity coefficient of about 1. The density of carbonic acid is similar to that of pure water, and the density of pure water slightly varies with temperature, then x H + ≈ S CO 2 . In this way, x H + and pH are calculated as follows: Figure 14 shows the average corrosion rate and estimated pH value as a function of CO 2 partial pressure in 30% crude oil/brine mixtures. The system pH decreased with increasing CO 2 partial pressure. When the CO 2 partial pressure was small, the system pH decreased significantly with increasing CO 2 partial pressure. When the CO 2 partial pressure was high, the pH of the system decreased insignificantly with increasing CO 2 partial pressure. The dissolution of CO 2 in water to achieve equilibrium continued to increase the CO 2 partial pressure but the pH value almost no longer increased [25]. Therefore, the different average corrosion rates in 30% crude oil/brine with CO 2 partial pressure were not only caused by pH changes but by a series of chemical changes. much larger than its x H + . In carbonate solution, except for CO2(aq), the concentration of other ions is negligible. Carbonic acid is a dilute solution with an ionic activity coefficient of about 1. The density of carbonic acid is similar to that of pure water, and the density of pure water slightly varies with temperature, then x H + ≈ S CO 2 . In this way, x H + and pH are calculated as follows: Figure 14 shows the average corrosion rate and estimated pH value as a function of CO2 partial pressure in 30% crude oil/brine mixtures. The system pH decreased with increasing CO2 partial pressure. When the CO2 partial pressure was small, the system pH decreased significantly with increasing CO2 partial pressure. When the CO2 partial pressure was high, the pH of the system decreased insignificantly with increasing CO2 partial pressure. The dissolution of CO2 in water to achieve equilibrium continued to increase the CO2 partial pressure but the pH value almost no longer increased [25]. Therefore, the different average corrosion rates in 30% crude oil/brine with CO2 partial pressure were not only caused by pH changes but by a series of chemical changes.

Formation Mechanism of Localized Corrosion
With the change in CO2 partial pressure, the corrosion rate of the J55 carbon steel changed significantly in 30% crude oil/brine mixtures, which indicated that the corrosion mechanism also changed. Figure 13 shows the partition graph of the corrosion rate in CO2/30% crude oil/brine mixtures as the change in CO2 partial pressure. Water-in-oil and oil-in-water emulsions coexist in 30% crude oil/brine mixtures, and the ratio of oil-in-water emulsion is large. Crude oil and water can all moisten the metal surface. The crude oil with corrosion inhibition was not evenly adsorbed on the metal surface, causing localized corrosion. As shown in Figure 14, corrosion models were proposed to clarify the influence of CO2 partial pressure on the mechanism of localized corrosion. The four stages describing the formation of localized corrosion are as follows: Model I (shown in Figure 15a): At Pco 2 = 0-1.5 MPa, the corrosion rate increased rapidly with the increase in CO2 partial pressure. When the CO2 partial pressure was small, the system pH

Formation Mechanism of Localized Corrosion
With the change in CO 2 partial pressure, the corrosion rate of the J55 carbon steel changed significantly in 30% crude oil/brine mixtures, which indicated that the corrosion mechanism also changed. Figure 13 shows the partition graph of the corrosion rate in CO 2 /30% crude oil/brine mixtures as the change in CO 2 partial pressure. Water-in-oil and oil-in-water emulsions coexist in 30% crude oil/brine mixtures, and the ratio of oil-in-water emulsion is large. Crude oil and water can all moisten the metal surface. The crude oil with corrosion inhibition was not evenly adsorbed on the metal surface, causing localized corrosion. As shown in Figure 14, corrosion models were proposed to clarify the influence of CO 2 partial pressure on the mechanism of localized corrosion. The four stages describing the formation of localized corrosion are as follows: Model I (shown in Figure 15a): At Pco 2 = 0-1.5 MPa, the corrosion rate increased rapidly with the increase in CO 2 partial pressure. When the CO 2 partial pressure was small, the system pH decreased significantly with increasing CO 2 partial pressure, similar to the estimated pH value shown in Figure 14. The concentration of H 2 CO 3 increased with increasing CO 2 partial pressure, which accelerated the cathodic reactions, increased the corrosion rate [13][14][15][16], and finally increased the Fe 2+ content. The corrosion scale precipitated and the scattered corrosion scale appeared on the surface, but the solubility products of FeCO 3 and CaCO 3 increased because the pH decreased, as shown in Figure 6b. Pitting may also occur locally due to the presence of crude oil and Cl − [9].
Model III (shown in Figure 15c): At Pco 2 = 5.0-9.0 MPa CO2, the corrosion rate increased with increasing CO2 partial pressure, which is in good agreement with the literature [28]. With the increase in CO2 partial pressure from 5 to 9.0 MPa, CO2 phase changed to a supercritical state, and the solubility of CO2 in crude oil increased rapidly [35]. When the decrease in pH value promoted the dissolution of protective layers, CO2 possibly transferred from crude oil to the aqueous phase, supplemented with consumed H + dissolving the product layer [36]. Thus, the surface of carbon steel was exposed to corrosive medium, and corrosion reaction was promoted [13,16]. The dissolution rate of the corrosion product layer was greater than the precipitation rate as the CO2 partial pressure was increased.
Model IV (shown in Figure 15d): At Pco 2 = 9.0~15.0 MPa, the corrosion rate was almost constant with the increase in CO2 partial pressure. This result can be attributed to the almost-constant system pH value (about 3.10-3.14), as shown in Figure 14. At the initial stage of the experiment, the metal surface suffered strong localized corrosion. The production and dissolution of corrosion scale were carried out simultaneously. Finally, a dense, complete, and protective corrosion product layer formed rapidly on the metal surface. Yoon-Seok Choi et al. [18] also obtained similar results, i.e., the corrosion rates of L80 in 25 wt.% NaCl solution started out high but ended up being very low at 90 °C and 12 MPa CO2 pressure. Pitting may occur under dense protective layer, as shown in Figure 13.  Model II (shown in Figure 15b): At Pco 2 = 1.5-5.0 MPa, the corrosion rate decreased with increasing CO 2 partial pressure. At the initial stage of the experiment, the metal surface suffered strong localized corrosion. This result is consistent with the conclusions of many researchers [5,[13][14][15][16]. The Fe 2+ concentration increased and acidic concentration decreased rapidly on the steel surface. The nucleation and growth of FeCO 3 typically start on the steel surface where the pH and FeCO 3 saturation values are the highest [34]. The FeCO 3 layer restricted the transport of H + in and Fe 2+ out; thus, the corrosion rate decreased with the increase in CO 2 partial pressure. The reduction of the system pH value also dissolved the corrosion scale, but the dissolution rate of the corrosion product layer was lower than the precipitation rate as the CO 2 partial pressure was increased.
Model III (shown in Figure 15c): At Pco 2 = 5.0-9.0 MPa CO 2 , the corrosion rate increased with increasing CO 2 partial pressure, which is in good agreement with the literature [28]. With the increase in CO 2 partial pressure from 5 to 9.0 MPa, CO 2 phase changed to a supercritical state, and the solubility of CO 2 in crude oil increased rapidly [35]. When the decrease in pH value promoted the dissolution of protective layers, CO 2 possibly transferred from crude oil to the aqueous phase, supplemented with consumed H + dissolving the product layer [36]. Thus, the surface of carbon steel was exposed to corrosive medium, and corrosion reaction was promoted [13,16]. The dissolution rate of the corrosion product layer was greater than the precipitation rate as the CO 2 partial pressure was increased.
Model IV (shown in Figure 15d): At Pco 2 = 9.0~15.0 MPa, the corrosion rate was almost constant with the increase in CO 2 partial pressure. This result can be attributed to the almost-constant system pH value (about 3.10-3.14), as shown in Figure 14. At the initial stage of the experiment, the metal surface suffered strong localized corrosion. The production and dissolution of corrosion scale were carried out simultaneously. Finally, a dense, complete, and protective corrosion product layer formed rapidly on the metal surface. Yoon-Seok Choi et al. [18] also obtained similar results, i.e., the corrosion rates of L80 in 25 wt.% NaCl solution started out high but ended up being very low at 90 • C and 12 MPa CO 2 pressure. Pitting may occur under dense protective layer, as shown in Figure 13.

Conclusions
Based on the observed corrosion behavior of J55 carbon steel in different CO 2 /30% crude oil/brine mixtures at 65 • C, we conclude the following: (1) The corrosion rate sharply increased as the CO 2 partial pressure was increased from 0 to 1.5 MPa, decreased from Pco 2 = 1.5 MPa to Pco 2 = 5.0 MPa, increased again at Pco 2 = 5.0 MPa, and then reached a constant value after Pco 2 = 9.0 MPa.
(2) In 30% crude oil/brine mixtures, the surface of J55 carbon steel was covered by FeCO 3 and CaCO 3 . The surface of the J55 carbon steel suffered localized corrosion in different CO 2 /30% crude oil/brine mixtures at 65 • C.
(3) The system pH initially decreased, rapidly increased, and then stabilized as CO 2 partial pressure was increased. The CO 2 partial pressure changed the system pH and CO 2 solubility in crude oil, which further affected the formation and protection performance of the corrosion product layer.