Effect of Chromium on Corrosion Behavior of P110 Steels in CO2-H2S Environment with High Pressure and High Temperature

The novel Cr-containing low alloy steels have exhibited good corrosion resistance in CO2 environment, mainly owing to the formation of Cr-enriched corrosion film. In order to evaluate whether it is applicable to the CO2 and H2S coexistence conditions, the corrosion behavior of low-chromium steels in CO2-H2S environment with high pressure and high temperature was investigated using weight loss measurement and surface characterization. The results showed that P110 steel suffered localized corrosion and both 3Cr-P110 and 5Cr-P110 steels exhibited general corrosion. However, the corrosion rate of 5Cr-P110 was the highest among them. The corrosion process of the steels was simultaneously governed by CO2 and H2S. The outer scales on the three steels mainly consisted of FeS1−x crystals, whereas the inner scales on Cr-containing steels comprised of amorphous FeS1−x, Cr(OH)3 and FeCO3, in contrast with the amorphous FeS1−x and FeCO3 mixture film of P110 steel. The more chromium the steel contains, the more chromium compounds the corrosion products contain. The addition of chromium in steels increases the uniformity of the Cr-enriched corrosion scales, eliminates the localized corrosion, but cannot decrease the general corrosion rates. The formation of FeS1−x may interfere with Cr-enriched corrosion scales and lowering the corrosion performance of 3Cr-P110 and 5Cr-P110 steels.


Introduction
The CO 2 /H 2 S corrosion problems of oil country tubular goods (OCTG) become increasingly prominent with the exploitation of oil and gas field under high temperature and high CO 2 and/or H 2 S pressure [1][2][3][4][5][6]. Corrosion resistant alloys (CRAs), such as stainless steels and high-nickel alloys, have been developed to mitigate the CO 2 /H 2 S corrosion long time ago, but the high cost constrains their application in oil and gas field containing CO 2 and/or H 2 S. Therefore, carbon and low alloy steel are still cost-effective materials used for tubings and pipelines despite of the shortcoming of high corrosion rate, and great efforts have been made to increase its corrosion resistance [4][5][6][7][8][9]. In recent years, the novel Cr-containing low alloy steels have been developed to balance the cost advantage and corrosion resistance between carbon steel and CRAs [5][6][7][8][9][10][11][12][13]. Many studies indicate that the low Cr alloy steels with 3-5 wt % Cr can not only remarkably reduce the CO 2 corrosion rate, but also avoid localized corrosion in CO 2 environment, mainly due to the formation of the amorphous Cr(OH) 3 in the scales on low Cr alloy steel [8][9][10][11][12][13]. Sun et al. [12] reported that, with an addition of 3% chromium in X65 steel, the corrosion rate dropped significantly from 11.59 to 1.57 mm/y and the localized corrosion was

Material and Pretreatment
The P110, 3Cr-P110, and 5Cr-P110 tube steels, with chemical compositions shown in Table 1, were used in this study. The specimens were machined into a size of 35 mmˆ15 mmˆ3 mm. Before the tests, the working surface of each specimen was abraded with silicon carbide paper of decreasing roughness (up to 800 grit), rinsed with deionized water, degreased with acetone and dried in air. The four parallel specimens for each test were weighed using an electronic balance with a precision of 0.1 mg, installed in a modified rotating polytetrafluoroethylene (PTFE) holder and then stored in a desiccator. The immersion corrosion test solution, a 3.5 wt % NaCl solution, was prepared with analytical grade reagents and deionized water.

Weight Loss Test
To investigate the corrosion rate and corrosion morphology of tube steels in CO 2 -H 2 S environment, immersion corrosion tests were carried out in a 3 L autoclave featuring high temperature and high pressure. Prior to the tests, the solution was purged with highly-purified N 2 to deoxidize for 12 h. The specimens were immersed into solution as soon as the solution was added into the autoclave, and then purging N 2 was used to remove the air for 2 h immediately after the autoclave was closed. After that, the vent valve was closed. The solution was heated to 90˝C, and then the CO 2 /H 2 S mixture gases (P CO 2 /P H 2 S = 25) were injected to autoclave to reach a total pressure of 5.2 MPa (P CO 2 = 5 MPa, P H 2 S = 0.2 MPa). The flow rate was 1 m/s at specimen surface. The tests were carried out for 360 h.
After corrosion tests, the specimens were taken out of the autoclave, rinsed in deionized water, dehydrated in alcohol and dried in air, respectively. One of the four specimens was retained for surface characterization of corrosion scales. The rest three specimens were descaled in the solution consisting of hydrochloric acid (100 mL, density is 1.19 g/mL), hexamethylene tetramine (5 g), and deionized water (900 mL) at room temperature, and then processed as above. After that, the specimens were weighed again to determine the weight loss. The corrosion rate was calculated through the following equation: where V CR is the corrosion rate, mm/y; W is the weight loss, g; S is the exposed surface area of specimen, cm 2 ; ρ is the density of specimen, g/cm 3 ; t is the corrosion time, h; 8.76ˆ10 4 is the unit conversion constant. The average corrosion rate with error bars was calculated from the three parallel specimens for each test.

Characterization of the Corrosion Scale
The surface and cross-sectional morphologies of the corrosion scales were observed using scanning electron microscope (SEM). The elemental compositions of the corrosion scales were analyzed using energy dispersive spectroscopy (EDS) with an acceleration voltage of 15 kV. The phase compositions of the corrosion scales were identified by means of X-ray diffraction (XRD) with a Cu Kα X-ray source operated at 40 kV and 150 mA, and the surface chemistry of the corrosion scales were also measured by X-ray photoelectron spectroscopy (XPS) with an Al Kα (hv = 1486.6 eV) X-ray source. Figure 1 presents the average corrosion rates of tested tube steels after immersion corrosion tests in a CO 2 -H 2 S environment. It can be seen that the corrosion rate of 5Cr-P110 (1.57 mm/y) was the highest among the three steels, and the corrosion rate of 3Cr-P110 (1.08 mm/y) was approximate to that of P110 (1.12 mm/y), taking into account the experimental error. This suggests that the corrosion resistance of P110 tube steel cannot be significantly improved by increasing Cr content in a CO 2 -H 2 S condition (P CO 2 = 5 MPa, P H 2 S = 0.2 MPa). Therefore, although the low-chromium steel has exhibited eminent corrosion resistance in CO 2 environment, especially when Cr content is 3-5 wt % in the steel [11,12], it may not be necessarily applicable to the CO 2 -H 2 S condition with high pressure. That is to say, for low Cr steel, the corrosion performance varies depending on different operational conditions. In a CO 2 environment, the protection is provided through the formation of a Cr-enriched FeCO 3 corrosion product (Cr oxi-hydroxide) [9][10][11][12][13]. The more chromium the steel contains, the more chromium compounds the corrosion products contain, and the better the protection is. However, under test condition, H 2 S might play a significant role in determining the type and properties of the corrosion scales, i.e., gradually undermining the corrosion scales and reducing their protective performance [1,14]. condition (PCO 2 = 5 MPa, PH 2 S = 0.2 MPa). Therefore, although the low-chromium steel has exhibited eminent corrosion resistance in CO2 environment, especially when Cr content is 3-5 wt % in the steel [11,12], it may not be necessarily applicable to the CO2-H2S condition with high pressure. That is to say, for low Cr steel, the corrosion performance varies depending on different operational conditions. In a CO2 environment, the protection is provided through the formation of a Cr-enriched FeCO3 corrosion product (Cr oxi-hydroxide) [9][10][11][12][13]. The more chromium the steel contains, the more chromium compounds the corrosion products contain, and the better the protection is. However, under test condition, H2S might play a significant role in determining the type and properties of the corrosion scales, i.e., gradually undermining the corrosion scales and reducing their protective performance [1,14].  Figure 2 shows the macroscopic surface morphology of the three steels before and after the removal of corrosion scales. As exhibited in the figures, the P110 steel surface were covered with tumor-like corrosion products (Figure 2a), and after removal of the corrosion scales, the positions where the "tuberculation" covered suffered shallow mesa attack ( Figure 2b). This indicates that there is probably some connection between localized corrosion and the corrosion scale at that location. In contrast, the corrosion scales were relatively flat on 3Cr-P110 and 5Cr-P110 steels surface with a non-adhesive outer layer (Figure 2c,e). After the specimens were taken out from solution, the outer layer cracked due to dehydration and partly peeled-off the 3Cr-P110 steel surface (Figure 2c), and most of the outer layer scales peeled off the 5Cr-P110 steel surface (Figure 2e). 3Cr-P110 and 5Cr-P110 steels were subject to general corrosion (Figure 2d,f), indicating that the addition of 3-5 wt % Cr into carbon steel improves the localized corrosion resistance in CO2-H2S environment, which is consistent with the case in CO2 environment [11][12][13]. However, the corrosion rates of Cr-containing steels did not decrease, suggesting that the addition of chromium could increase the uniformity of the corrosion scale but not improve the diffusion resistance to corrosive ions.  Figure 2 shows the macroscopic surface morphology of the three steels before and after the removal of corrosion scales. As exhibited in the figures, the P110 steel surface were covered with tumor-like corrosion products (Figure 2a), and after removal of the corrosion scales, the positions where the "tuberculation" covered suffered shallow mesa attack ( Figure 2b). This indicates that there is probably some connection between localized corrosion and the corrosion scale at that location. In contrast, the corrosion scales were relatively flat on 3Cr-P110 and 5Cr-P110 steels surface with a non-adhesive outer layer (Figure 2c,e). After the specimens were taken out from solution, the outer layer cracked due to dehydration and partly peeled-off the 3Cr-P110 steel surface (Figure 2c), and most of the outer layer scales peeled off the 5Cr-P110 steel surface (Figure 2e). 3Cr-P110 and 5Cr-P110 steels were subject to general corrosion (Figure 2d,f), indicating that the addition of 3-5 wt % Cr into carbon steel improves the localized corrosion resistance in CO 2 -H 2 S environment, which is consistent with the case in CO 2 environment [11][12][13]. However, the corrosion rates of Cr-containing steels did not decrease, suggesting that the addition of chromium could increase the uniformity of the corrosion scale but not improve the diffusion resistance to corrosive ions. 4 most of the outer layer scales peeled off the 5Cr-P110 steel surface (Figure 2e). 3Cr-P110 and 5Cr-P110 steels were subject to general corrosion (Figure 2d,f), indicating that the addition of 3-5 wt % Cr into carbon steel improves the localized corrosion resistance in CO2-H2S environment, which is consistent with the case in CO2 environment [11][12][13]. However, the corrosion rates of Cr-containing steels did not decrease, suggesting that the addition of chromium could increase the uniformity of the corrosion scale but not improve the diffusion resistance to corrosive ions. (c,d) 3Cr-P110 and (e,f) 5Cr-P110. Figure 3 shows the XRD spectra of the corrosion scales on the steels. It can be seen that the crystals in the scales of P110 steel mainly comprised FeS1−x and FeCO3, with a small amount of FeS and Fe1−x S. However the main crystals in the scales of 3Cr-P110 and 5Cr-P110 steels were FeS1−x with only a small amount of FeCO3 crystals detected. The content of FeCO3 crystals in corrosion products gradually reduced in the order of P110, 3Cr-P110 and 5Cr-P110. The crystalline state of FeCO3 is predominantly determined by the pH [9]. High pH value causes the formation of FeCO3 crystal while low pH results in amorphous FeCO3. Therefore, the reduction of FeCO3 crystals is probably, to some extent, related to the pH of the solution in the proximity of the solution/scale or solution/metal interface. SEM surface morphology of corrosion scales on the steels are shown in Figure 4. As exhibited in Figure 4a,c,e, the corrosion scales on the three steels showed similar morphology. The outer corrosion layers comprised tiny crystalline products, and the result of EDS analysis indicated that the outer layer scales mainly consisted of Fe and S elements (Figure 4b,d,f). It was concluded that these tiny crystalline products are mainly FeS1−x. However, the inner scales showed amorphous characteristic with cracks caused by surface dehydration, and all contained Fe, S and O elements. In addition, the inner   Figure 3 shows the XRD spectra of the corrosion scales on the steels. It can be seen that the crystals in the scales of P110 steel mainly comprised FeS 1´x and FeCO 3 , with a small amount of FeS and Fe 1´x S. However the main crystals in the scales of 3Cr-P110 and 5Cr-P110 steels were FeS 1´x with only a small amount of FeCO 3 crystals detected. The content of FeCO 3 crystals in corrosion products gradually reduced in the order of P110, 3Cr-P110 and 5Cr-P110. The crystalline state of FeCO 3 is predominantly determined by the pH [9]. High pH value causes the formation of FeCO 3 crystal while low pH results in amorphous FeCO 3 . Therefore, the reduction of FeCO 3 crystals is probably, to some extent, related to the pH of the solution in the proximity of the solution/scale or solution/metal interface. (c,d) 3Cr-P110 and (e,f) 5Cr-P110. Figure 3 shows the XRD spectra of the corrosion scales on the steels. It can be seen that the crystals in the scales of P110 steel mainly comprised FeS1−x and FeCO3, with a small amount of FeS and Fe1−x S. However the main crystals in the scales of 3Cr-P110 and 5Cr-P110 steels were FeS1−x with only a small amount of FeCO3 crystals detected. The content of FeCO3 crystals in corrosion products gradually reduced in the order of P110, 3Cr-P110 and 5Cr-P110. The crystalline state of FeCO3 is predominantly determined by the pH [9]. High pH value causes the formation of FeCO3 crystal while low pH results in amorphous FeCO3. Therefore, the reduction of FeCO3 crystals is probably, to some extent, related to the pH of the solution in the proximity of the solution/scale or solution/metal interface. SEM surface morphology of corrosion scales on the steels are shown in Figure 4. As exhibited in Figure 4a,c,e, the corrosion scales on the three steels showed similar morphology. The outer corrosion layers comprised tiny crystalline products, and the result of EDS analysis indicated that the outer layer scales mainly consisted of Fe and S elements (Figure 4b,d,f). It was concluded that these tiny crystalline products are mainly FeS1−x. However, the inner scales showed amorphous characteristic with cracks caused by surface dehydration, and all contained Fe, S and O elements. In addition, the inner   Figure 4a,c,e, the corrosion scales on the three steels showed similar morphology. The outer corrosion layers comprised tiny crystalline products, and the result of EDS analysis indicated that the outer layer scales mainly consisted of Fe and S elements (Figure 4b,d,f). It was concluded that these tiny crystalline products are mainly FeS 1´x . However, the inner scales showed amorphous characteristic with cracks caused by surface dehydration, and all contained Fe, S and O elements. In addition, the inner scales of 3Cr-P110 and 5Cr-P110 steels contained abundant amounts of Cr (19.99 wt % and 28.55 wt %, respectively). Therefore, there must be some compounds, such as Cr-compounds, which were undetected by XRD.

The Composition and Elements Distribution of Corrosion Film
To further determine the phase composition of inner scales, XPS analysis was employed to analyze the surface chemistry of amorphous compounds in the inner scales. Figures 5-7 provides the high resolution XPS spectra of inner scales on P110, 3Cr-P110 and 5Cr-P110 steels. The elements of interest were Fe, O, S and C and, in the meantime, Cr was also investigated for 3Cr-P110 and 5Cr-P110 steels. Surface charging effects were compensated by referencing the binding energy to the C 1s line of the residual carbon set at 284.6 eV. The Gaussian-Lorentzian curves were used to fit the peaks. Binding energies of Fe 2p, O 1s, S 2p, and Cr 2p for inner scales on P110, 3Cr-P110, and 5Cr-P110 samples are summarized in Table 2.  Table 2.       Referring to Figure 5 and Table 2 [2, [27][28][29], the results suggested that the inner scale of P110 steel mainly consisted of FeS 1´x and FeCO 3 . The S 2p 3/2 peak at 163.7 eV corresponded to elemental sulfur which was detected because iron sulfide got oxidized while in air [3]. In addition, considering the limitation of the corrosion environment, the S 2p 3/2 peak at a binding energy of 168.9 eV could be attributable to the adventitious SO 4 2´ [ 30]. The high resolution XPS spectra of inner scales on 3Cr-P110 and 5Cr-P110 steels are shown in Figures 6 and 7 respectively. As exhibited in Figures 6 and 7 and Table 2, the C 1s, S 2p, and Fe 2p scans for the inner scales on 3Cr-P110 and 5Cr-P110 steels reflected a similar composition of corrosion products compared to that of P110 steel, confirming the presence of FeS 1´x and FeCO 3 . A small difference was the Fe 2p 3/2 peak at 707.3 eV for 5Cr-P110 steel associated with iron sulfide [31], as shown in Figure 7e. However, a noticeable difference existed in the O spectra for 3Cr-P110 and 5Cr-P110 steels compared to that of P110 steel, along with the existence of Cr 2 O 3 and Cr(OH) 3 [9,32], except for FeCO 3 [27]. The Cr 2p peaks both revealed the existence of Cr(OH) 3 [9,32,33] in 3Cr-P110 and 5Cr-P110 scales. The Cr 2 O 3 [32] in 5Cr-P110 scale was believed to be the product of the dehydration of Cr(OH) 3 upon removal from the system [13]. The compositions of the inner scale on 3Cr-P110 and 5Cr-P110 steels were similar, both primarily consisting of FeS 1´x , FeCO 3 and amorphous Cr(OH) 3 .     Figure 8 shows the cross-sectional backscattered electron images and EDS line scanning analysis of corrosion scales on the steels. As exhibited in Figure 8a,c,e, the scales on P110 (the position where P110 steel presented general corrosion morphology), 3Cr-P110 and 5Cr-P110 steels all had a two-layer structure after 360 h corrosion tests. It can be seen that the outer scales on three steels (Figure 8a The results were highly consistent with XRD and XPS analysis. As exhibited in Figure 8a, the outer FeS 1´x scale of P110 steel was very thin, uneven, and not closely attached to the inner scale with many large pores between inner and outer scales compared with that of 3Cr-P110 steel (Figure 8c). Whereas, for 5Cr-P110 steel, the outer FeS 1´x scale was porous and loose ( Figure 8e) compared with that of 3Cr-P110 steel. This is the reason why most of the outer scale of 5Cr-P110 steel peeled off (Figure 2e). Referring to Figure 5 and Table 2 [2,[27][28][29], the results suggested that the inner scale of P110 steel mainly consisted of FeS1−x and FeCO3. The S 2p3/2 peak at 163.7 eV corresponded to elemental sulfur which was detected because iron sulfide got oxidized while in air [3]. In addition, considering the limitation of the corrosion environment, the S 2p3/2 peak at a binding energy of 168.9 eV could be attributable to the adventitious SO4 2− [30].
The high resolution XPS spectra of inner scales on 3Cr-P110 and 5Cr-P110 steels are shown in Figures 6 and 7, respectively. As exhibited in Figures 6 and 7 and Table 2, the C 1s, S 2p, and Fe 2p scans for the inner scales on 3Cr-P110 and 5Cr-P110 steels reflected a similar composition of corrosion products compared to that of P110 steel, confirming the presence of FeS1-x and FeCO3. A small difference was the Fe 2p3/2 peak at 707.3 eV for 5Cr-P110 steel associated with iron sulfide [31], as shown in Figure 7e. However, a noticeable difference existed in the O spectra for 3Cr-P110 and 5Cr-P110 steels compared to that of P110 steel, along with the existence of Cr2O3 and Cr(OH)3 [9,32], except for FeCO3 [27]. The Cr 2p peaks both revealed the existence of Cr(OH)3 [9,32,33] in 3Cr-P110 and 5Cr-P110 scales. The Cr2O3 [32] in 5Cr-P110 scale was believed to be the product of the  Referring to Figure 5 and Table 2 [2,[27][28][29], the results suggested that the inner scale of P110 steel mainly consisted of FeS1−x and FeCO3. The S 2p3/2 peak at 163.7 eV corresponded to elemental sulfur which was detected because iron sulfide got oxidized while in air [3]. In addition, considering the limitation of the corrosion environment, the S 2p3/2 peak at a binding energy of 168.9 eV could be attributable to the adventitious SO4 2− [30].
The high resolution XPS spectra of inner scales on 3Cr-P110 and 5Cr-P110 steels are shown in Figures 6 and 7, respectively. As exhibited in Figures 6 and 7 and Table 2, the C 1s, S 2p, and Fe 2p scans for the inner scales on 3Cr-P110 and 5Cr-P110 steels reflected a similar composition of corrosion products compared to that of P110 steel, confirming the presence of FeS1-x and FeCO3. A small difference was the Fe 2p3/2 peak at 707.3 eV for 5Cr-P110 steel associated with iron sulfide [31], as shown in Figure 7e. However, a noticeable difference existed in the O spectra for 3Cr-P110 and 5Cr-P110 steels compared to that of P110 steel, along with the existence of Cr2O3 and Cr(OH)3 [9,32], except for FeCO3 [27]. The Cr 2p peaks both revealed the existence of Cr(OH)3 [9,32,33] in 3Cr-P110 and 5Cr-P110 scales. The Cr2O3 [32] in 5Cr-P110 scale was believed to be the product of the Referring to Figure 5 and Table 2 [2,[27][28][29], the results suggested that the inner scale of P110 steel mainly consisted of FeS1−x and FeCO3. The S 2p3/2 peak at 163.7 eV corresponded to elemental sulfur which was detected because iron sulfide got oxidized while in air [3]. In addition, considering the limitation of the corrosion environment, the S 2p3/2 peak at a binding energy of 168.9 eV could be attributable to the adventitious SO4 2− [30].

The Effect of Cr Content on Formation of Corrosion Scale and Its Relation to Corrosion
It is well established that driving force for precipitation is the supersaturation of corrosion products in CO2-H2S environment, which depends on both the carbon steel characteristics (microstructure, heat treatment history, alloying elements) and environmental variables (solution pH, temperature, solution composition, flow rate, etc.) [1,35]. As mentioned earlier, FeCO3, FeySx and Cr(OH)3 were all detected in the corrosion scales. The competitive deposition of FeCO3, FeySx, and Cr(OH)3 results in the mixed films which play an important role in determining the corrosion form and corrosion rate.
In this study, H2S can increase CO2 corrosion by promoting anodic dissolution through sulfide adsorption and lowering pH and chromium can provide additional anodic reaction for Cr-containing According to the characteristics of corrosion scales on the tested steels, it is reasonable to believe that the corrosion process of the steels is simultaneously governed by CO 2 and H 2 S under the test conditions (P CO 2 /P H 2 S = 25, 90˝C). It is well established that corrosion scales containing the same components could be extremely protective, very little so, or even corrosive depending on the location of these components [9,12,[34][35][36][37]. The uneven formation and local damage of CO 2 corrosion product film is the main reason for localized corrosion [38]. The mesa corrosion was observed from Figures 2b and 8a, and the outer FeS 1´x scale disappeared in the regions where localized corrosion occurred. The local damage of corrosion scales on P110 steel provides the pathway for mass transfer of corrosive ions such as Cl´, HCO 3´, HS´, and H + , which are sufficient to cause the onset of internal acidification [34], thus accelerating the dissolution of metal at the location. As the corrosion proceeds, a mass of corrosion products is formed in corrosion pits, thus forming the tumor-like corrosion products on the steel (Figures 2a and 8a). However, for the Cr-containing steel, the formation of uniform corrosion scales inhibits the localized corrosion, which may be related to the high Cr content in corrosion films.

The Effect of Cr Content on Formation of Corrosion Scale and Its Relation to Corrosion
It is well established that driving force for precipitation is the supersaturation of corrosion products in CO 2 -H 2 S environment, which depends on both the carbon steel characteristics (microstructure, heat treatment history, alloying elements) and environmental variables (solution pH, temperature, solution composition, flow rate, etc.) [1,35]. As mentioned earlier, FeCO 3 , Fe y S x and Cr(OH) 3 were all detected in the corrosion scales. The competitive deposition of FeCO 3 , Fe y S x , and Cr(OH) 3 results in the mixed films which play an important role in determining the corrosion form and corrosion rate.
In this study, H 2 S can increase CO 2 corrosion by promoting anodic dissolution through sulfide adsorption and lowering pH and chromium can provide additional anodic reaction for Cr-containing steels. Therefore, both P110 steel and Cr-containing steels can be corroded rapidly at the early stage of the immersion period. Then, the supersaturation is so high that a high nucleation rate may emerge, causing an amorphous corrosion film formation on the steel surface. It can be seen from the cross-sections (Figure 8) that all the inner layers attached directly to the steel surface were apparently denser than outer layers, indicating a better protective performance of inner layers than outer layers. The inner scales acted as a diffusion barrier restricting the diffusion of reactive species and lowering the corrosion rates with time. It can also be seen that the inner layer on 5Cr-P110 was obviously not as compact as those on P110 and 3Cr-P110, which goes against the results in sweet system [12]. It suggested [14] that Fe y S x interfered with FeCO 3 and Cr(OH) 3 mixed film and reduced its protective properties especially for 5Cr-P110. Therefore, 5Cr-P110 had the highest corrosion rate, which may be attributable to the excessive chromium in 5Cr-P110 steel. According to the EDS line scanning analysis along the vertical direction of the cross-section (Figure 8b,d,f), the Cr content increased and S content decreased obviously in the inner scale on 5Cr-P110 steel compared with those in the inner scales on P110 and 3Cr-P110 steels. Since Cr(OH) 3 is more stable than FeCO 3 and Fe y S x , thermodynamically, the enrichment of Cr in the amorphous inner layers is due to the dissolution of FeCO 3 [13,39] and Fe y S x in the scale, but it is unclear how H 2 S interferes with the Cr-enriched corrosion film formation and, hence, relatively lowers the protection properties of the films.
At the later stage of the immersion period, the outer FeS 1´x layer with a small crystalline form precipitated on the inner layers. However, the outer FeS 1´x scales of the three steels are strikingly different, such as the thickness and compactness, probably because of the the amorphous Cr(OH) 3 in the inner scales. Guo et al. [9] found that the Cr content in steel had a significant impact on the in situ pH value of solution in the proximity of the scale/solution interface. The pH value can be reduced due to the formation of Cr(OH) 3 by Equation (2) [9] or to the hydrolysis of Cr 3+ ions by Equation (3) [40]. The higher the Cr content is, the lower the pH value can be achieved at the solution/scale interface. Therefore, the pH value in the proximity of 5Cr-P110 steel is lower than that of 3Cr-P110 steel, which results in a lower precipitation rate and higher dissolution rate of the film on 5Cr-P110 steel, and a porous, loosely-adherent FeS 1´x scale formation on the inner layer ( Figure 8c). Additionally, the lower pH value in the proximity of Cr-containing steels may also, to some extent, be responsible for the lower content of FeCO 3 crystals in corrosion products (Figure 3). In low Cr steels, the protection is afforded through the formation of a Cr-rich FeCO 3 corrosion product (Cr oxi-hydroxide) in sweet production conditions [8][9][10][11][12][13]. On the introduction of H 2 S (sour systems), the formation of FeS 1´x may interfere with Cr-enriched corrosion scales and make the diffusion fluxes of the species involved in the electrochemical reactions easier, lowering the corrosion performance of Cr-containing steels, especially 5Cr-P110 steel. Therefore, 5Cr-P110 steel presents a higher corrosion rate.

Conclusions
P110 steel suffered localized corrosion and both 3Cr-P110 and 5Cr-P110 steels exhibited general corrosion in CO 2 -H 2 S environment with high pressure and high temperature. The corrosion rate of 3Cr-P110 (1.08 mm/y) was approximate to that of P110 (1.12 mm/y) and the corrosion rate of 5Cr-P110 (1.57 mm/y) was the highest among the three steels.
The corrosion process of the steels was governed by CO 2 and H 2 S simultaneously under the test conditions. The outer scales on the three steels mainly consisted of FeS 1´x crystals. The inner scale on P110 steel was composed of amorphous FeS 1´x and FeCO 3 . However, the inner scale on both 3Cr-P110 and 5Cr-P110 steels comprised amorphous FeS 1´x , Cr(OH) 3 , and FeCO 3 . The inner layers attached directly to the steel surface were apparently denser than outer layers. The more chromium the steel contains, the more chromium compounds the corrosion products contain. The addition of chromium in steels increases the uniformity of the Cr-enriched corrosion scales, eliminates the localized corrosion, but cannot decrease the general corrosion rates. Under test condition, the formation of FeS 1´x may interfere with Cr-enriched corrosion scales and reduce their protective properties.
However, it is still unclear how H 2 S interferes with the Cr-enriched corrosion film formation. As a consequence of the complexity of CO 2 /H 2 S corrosion, there is a need to carry out further study in order to clarify the mechanism involved in the chromium effect on the corrosion behavior and to develop novel low-chromium steels of both good CO 2 /H 2 S corrosion resistance and low cost.