Corrosion Behavior of Oxide Scale of 5Cr Steel in CO₂ Flooding Environment

Abstract

Medium-Cr steel is considered as the most economical and effective pipeline material in a CO₂ flooding environment; however, the effects of iron oxides on the corrosion resistance of medium-Cr steel are still not clear. Therefore, in the present study, a high-temperature, high-pressure reactor was used to accelerate the corrosion of a rolled oxidized sheet of 5Cr steel. The corrosion behavior of oxide scale leaching was studied by comparing the mass loss and gain rate with bare steel, through using XRD, Raman, XPS, SEM, EDS, EPMA and other methods. It was found that the 5Cr steel rolled scale was composed of three layers: the innermost layer contained Fe₃O₄, the middle layer consisted of Fe₂O₃, and the outer layer contained Fe₃O₄ and a small amount of Fe₂O₃. In the CO₂ environment, the oxide scale mainly dissolved in the initial stage, and corrosion products were deposited in the later stage. The oxide scale delayed the occurrence of corrosion, and eventually, similar to the bare steel, a double-layer product film was formed. Also, the possible corrosion mechanism of the medium-Cr steel with the oxide scale was given.

1. Introduction

Medium-Cr, low-C steel is considered as the most suitable pipeline material in a CO₂ flooding environment.¹⁾ During corrosion, an amorphous inner film of Cr(OH)₃ and an outer film of FeCO₃ crystals are formed on the steel substrate surface,^2,3,4) and they hindered the ion exchange process between the solution and the substrate to prevent further corrosion. However, in an actual service process of pipeline steel, oxide scales are not cleanly removed, and oxidation occurs during storage and transportation. There are currently few studies on the role of the oxide layer in the service process.

From the only reports related to iron oxides in CO₂ environment, we found that iron oxides mainly appear in the resultant corrosion product film at high temperatures. Yuan⁵⁾ et al. found that FeCO₃ was decomposed at high temperature (above 240°C), and the corrosion products of medium-Cr steel were composed of FeCO₃, (Fe, Cr)₃O₄ and Fe₂O₃. Moreover, (Fe, Cr)₃O₄ was prone to crack, decreasing the corrosion resistance of the steel. Chen⁶⁾ et al. observed that at 90°C, the formation of dense Fe₂O₃ decreased the corrosion rate and inhibit overall corrosion. Hua⁷⁾ et al. reported that Fe₃O₄ formed at higher temperatures (250°C) had excellent corrosion resistance.

Luo⁸⁾ et al. found that the introduction of O₂ in the CO₂ flooding environment resulted in the appearance of Fe₂O₃ in various forms. Hua⁹⁾ et al. reported that in the O₂ environment, in addition to Fe₂O₃, FeOOH and Fe₃O₄ were also formed; thus, uniform corrosion was inhibited, and local corrosion was aggravated. Lin¹⁰⁾ et al. found that in the CO₂ environment containing O₂, the oxide film of corrosion products formed on the surface of 3Cr steel was loose and porous, and mainly consisted of FeCO₃, Fe₃O₄, and Fe₂O₃. No obvious Cr element enrichment in corrosion products was noticed, and the 3Cr steel accelerated pitting corrosion.

Forero¹¹⁾ et al. described the transformation mechanism of different iron oxides at 80°C. FeCO₃ decomposes into FeO+CO₂. FeO and O₂ form Fe₂O₃ in an aerobic environment, whereas FeO can form Fe₃O₄+CO with CO₂ in an anaerobic environment. Han¹²⁾ et al. confirmed the existence of Fe₃O₄ in the passivation product of carbon steel in CO₂ water environment, also indicated that the Fe₃O₄ formed predominantly at the boundaries of FeCO₃ crystals.

The above-mentioned literatures only found the presence of iron oxides in the CO₂ flooding environment, however, whether the presence of oxides has a positive or negative impact on the corrosion resistance of pipeline steel is still unclear. It is generally believed that dense oxide layers are better than loose ones. In the present study, a dense 5Cr steel (wt.%) rolled oxide scale was used to study its corrosion behavior in a CO₂ environment to reveal the influence of oxides on the corrosion resistance of medium-Cr steel.

2. Experimental

The steel used in this experiment was smelted in a vacuum smelting furnace, and its main chemical composition (wt.%) was: 0.035C, 5Cr, 0.25Si, 0.15Mn, <0.005P, <0.008S, <0.01Ti, and balance Fe. The starting and finishing forging temperatures were 1150°C and 900°C, respectively, and the forged sample size was 80 mm×90 mm×120 mm. The rolling process was carried out in two stages. Prior to rolling, the steel was heated to 1200°C and then isothermally held for two hours.

First stage:

Second stage:

A laser confocal microscope (Model: LEXT OLS4000) was used to detect the morphology of the oxide scale. A Zeiss EVO-18 scanning electron microscope (SEM) was used to characterize the micro-morphology of the oxide scale, and an energy dispersive spectrometer (EDS) was employed to measure the element content of corrosion products. A SHIMADZU EPMA-1720 electronic probe microanalyzer (EPMA) was used to accurately measure the elemental distribution in the oxide scale. Due to the poor conductivity of oxide scale and corrosion products, carbon was sprayed on the sample surface to improve the image quality. Therefore, the C element content was not considered in EDS and EPMA. An X-ray diffractometer (XRD), (Model: (Rigaku) D/MAX-RB12KW) was used to reveal the phase structure of the oxide scale.

The corrosion experiment was carried out in a 3L high-temperature autoclave with CO₂. The specimen size of the coupon was 30 mm × 10 mm × 3 mm. A small hole of Φ4 mm was cut in the upper part of the coupon. Five parallel samples were used for each group of experimental steel: three for corrosion rate measurements, one for surface morphology detection, and one for cross-sectional morphology detection. The samples without the oxide scale were used as comparison samples (three for each group in order to ensure reproducibility). After processing, all the six sides of the sample without oxide scale were sanded to 800# step by step, and the five sides of the sample with oxide scale were sanded except the oxide scale side. After sanding, anhydrous ethanol was used to clean and remove moisture. An electronic balance with a precision of 0.1 mg was used to weigh sample mass before and after the experiment.

The corrosion solution refers to the design of an oil field, and its composition is shown in Table 1. Before the experiment, CO₂ was injected for 30 min to remove the air in the autoclave and then continuously injected to maintain the partial pressure of carbon dioxide in the autoclave at 0.6 MPa. The temperature in the autoclave was set to 70°C. The corrosion experiment was carried out for different corrosion times of 6 h, 24 h, 48 h, 96 h, and 216 h. The samples were applied a rotation at a speed of 2.5 m/s.

Table 1. Composition of the corrosion solution (mg/L).

NaHCO3	MgSO4	CaSO4	NaCl
4140	750	537.5	16480

As the oxide scale only existed on one side of the sample, during the calculation of the mass loss and gain (rate) of the oxide scale sample, the mass loss and gain (rate) of the non-oxidized scale sample was used to calibrate the other sides. During the test, the non-oxidized scale samples were put into the reactor together with the oxide scale samples. After the test, corrosion products were removed by 50 mL of concentrated hydrochloric acid, 50 mL of deionized water, and 0.35 g of hexamethylenetetramine solution.

The corrosion mass change was calculated by the following equation:   

$$\text{Mass loss }\left({\text{g/cm}}^{\text{2}}\right)=\left(m_1-m_0-\mathrm{\Delta }{m}_1\right)/S_1$$(1)

  

$$\text{Mass loss rate }\left({\text{g/cm}}^{\text{2}}\text{h}\right)=\left(m_1-m_0-\mathrm{\Delta }{m}_1\right)/S_{1}t$$(2)

  

$$\text{Mass gain }\left({\text{g/cm}}^{\text{2}}\right)=\left(m_0-m_2-\mathrm{\Delta }{m}_2\right)/S_1$$(3)

  

$$\text{Mass gain rate }\left({\text{g/cm}}^{\text{2}}\text{h}\right)=\left(m_0-m_2-\mathrm{\Delta }{m}_2\right)/S_{1}t$$(4)

where m₀ is the original mass of oxide scale sample (g), m₁ is the mass after corrosion oxide scale sample (g), m₂ is the mass after the removal of corrosion products oxide scale sample (g), Δm₁ is the mass loss caused by the bare matrix of oxide scale sample (g), Δm₂ is the mass gain caused by the bare matrix of oxide scale sample (g), S₁ is the area covered by oxide scale (cm²), and t is corrosion time (h).   

$$\mathrm{\Delta }{m}_1=(m_1\prime -m_0\prime )S_2/S$$(5)

  

$$\mathrm{\Delta }{m}_2=(m_0\prime -m_2\prime )S_2/S$$(6)

where m₀′ is the original mass of non-oxidized scale sample (g), m₁′ is the mass after corrosion non-oxidized scale sample (g), m₂′ is the mass after the removal of corrosion products non-oxidized scale sample (g), S₂ is the bare area of the matrix of oxide scale sample (cm²), S is the surface area of non-oxidized scale sample (cm²).

The cross-sectional morphology of corrosion products was analyzed by a Horiba LabRAM HR Evolution Raman spectrometer. The analysis position was the center of the cross-section of corrosion products (about 10 μm away from the matrix). The XPS test was performed by an AXISULTRA-DLD X-ray photoelectron spectrometer (Kratos), and the powder used in the test was scraped off the surface of the corrosion sample.

3. Results

3.1. Microstructure and Phase Analysis of the Oxide Scale

The laser confocal surface image in Fig. 1(a) reveals that the rolled oxide scale of 5Cr steel had a little surface fluctuation; however, it was fragmentary and prone to be broken. The height difference of the surface fluctuation was about 50 μm (Fig. 1(b)). It is noticeable from the SEM image of the cross-section of the oxide scale that it had a thickness range of 15–20 μm and was composed of multiple layers. The innermost layer near the matrix was thin and dense, the middle layer was composed of bulk material, and the outermost layer was sparse and fragile. The EDS analysis revealed that (Figs. 1(d)–1(f)) the contents of Cr and O in the oxide scale were significantly higher than those in the matrix, whereas the Fe content was lower than those in the matrix. The oxide scale consisted of Fe and O and was also enriched with Cr.

Fig. 1.

Morphology and elemental distribution of the 5Cr steel rolled oxide scale: (a) Surface morphology, (b) Surface height difference, (c) Cross-sectional morphology, (d) Fe elemental distribution, (e) Cr elemental distribution, (f) O elemental distribution. (Online version in color.)

The finer cross-sectional elemental distribution in the oxide scale was detected by EPMA (Fig. 2). Figure 2(b) reveals that the oxide scale of 5Cr steel was divided into three layers. The innermost layer was thinner and continuously distributed on the outside of the matrix and had a low oxygen content. The middle layer was thicker and contained a higher O content, and the outer layer was loose. It can be seen from the content of O element, the outer layer was composed of a mixture of innermost and middle layers.

Fig. 2.

Elemental distribution in the cross-section of the oxide scale: (a) Location morphology, (b) O, (c) Cr, (d) Fe. (Online version in color.)

The Cr element distribution in Fig. 2(c) shows that the innermost layer and a part of the middle layer close to the matrix had a high Cr content. In addition, substances with a higher Cr content were distributed between the middle layer and the innermost layer, in the center of the middle layer, and on the outer layer, and it can be speculated that these substances were composed of Cr₂O₃.^13,14)

The XRD analysis expresses that the iron oxide scale of 5Cr steel was composed of Fe₂O₃, Fe₃O₄, and a small amount of FeCr (shown in Fig. 3). The EDS result in Fig. 1(e) reveals that a certain amount of Cr existed in the oxide scale; however, the XRD analysis depicts that the lattice constant of the iron oxide did not change significantly due to the presence of Cr. It can be presumed that Cr replaced a part of Fe in Fe₂O₃ or Fe₃O₄. Therefore, the oxide scale of 5Cr steel was mainly composed of Fe₂O₃ (or (Fe_0.6Cr_0.4)₂O₃), Fe₃O₄ (or FeCr₂O₄), and Cr₂O₃ (Fig. 2).

Fig. 3.

XRD test results of oxide scale surface before corrosion.

Combined with the XRD data and the elemental distribution in Fig. 2. As the O content of Fe₂O₃ was higher than that of Fe₃O₄, the thin innermost layer was composed of Fe₃O₄, the thicker middle layer contained Fe₂O₃, and the outer layer was composed of Fe₃O₄ and Fe₂O₃. Cr₂O₃ existed between the inner layer and the center of the middle layer and on the outer layer.

3.2. Corrosion Behavior of Oxide Scale

3.2.1. Mass Change during Corrosion

The mass change curve of the scale sample and the matrix sample after corrosion different times is shown in Fig. 4. The mass-loss rates of the sample with or without the oxide scale in the late stage of corrosion were close (Figs. 4(a), 4(b)), indicating that the existence of the oxide scale had no significant effect on the corrosion resistance of 5Cr steel. In the initial stage of corrosion, the mass loss of the oxide scale sample was significantly higher than that of the matrix sample. After 24 h, the mass loss reached a certain level then it did not change greatly. In contrast, the mass loss of the matrix sample increased with the extension of the corrosion time. The mass loss rate of the matrix sample was noticeably lower than that of the oxide scale sample in the early stage of corrosion. With the prolonged corrosion time, the mass loss rate of the matrix sample decreased slowly. The mass loss rate of the oxide scale sample was much higher in the early stage of corrosion; however, it decreased rapidly. The decreasing speed decelerated with the extension of the corrosion time, and the final corrosion rate was close to that of the matrix sample.

Fig. 4.

Mass change curves of corrosion samples: (a) Mass loss, (b) Mass loss rate, (c) Mass gain, (d) Mass gain rate. (Online version in color.)

It is noticeable from Fig. 4(c) that the mass gains of the matrix sample and the oxide scale sample in the initial stage of corrosion were negative, indicating that they were in a dissolved state. The rate of the oxide scale flaking off was greater than the dissolution rate of the matrix. Before 48 h, with the extension of the corrosion time, the difference between the mass gains of the matrix sample and the oxide scale sample gradually increased. However, after 48 h, the oxide scale changed from a massless state to a mass gain state (the mass gain became positive), whereas the matrix sample was still in the dissolved state.

With the extension of the corrosion time, the mass gain rates of the two kinds of samples increased continuously (Fig. 4(d)). The reasons for the increase in mass gain rate are composed of two aspects. On the one hand, the deposition of corrosion products compensated for the mass loss; on the other hand, corrosion products played a protective role and delayed the dissolution (or flaking) process. It is worth noting that the mass gain rate of the oxide scale sample at 48 h was close to that of the matrix sample. With a further extension of corrosion time, the mass gain rate of the oxide scale sample became positive, and the deposition amount of corrosion products exceeded the dissolved amount. Under long-time corrosion, the matrix sample was still in the dissolved state, although the dissolution rate slowed down.

3.2.2. Raman Analysis

The Raman spectroscopy analysis was performed in the cross-section of the product film of the matrix sample and oxide scale sample formed after 48 h of corrosion, and the analysis position was at the center of the product film (about 10 μm away from the matrix). In Fig. 5, for the matrix sample, the peak at 552 cm⁻¹ appeared from Cr₂O₃,¹⁵⁾ and the peaks at 731 cm^{−1 16)} and 1092 cm^{−1 17)} appeared from FeCO₃. For the oxide scale sample, the peak at 552 cm⁻¹ appeared from Cr₂O₃,¹⁵⁾ the peak at 670 cm⁻¹ appeared from Fe₃O₄,¹⁸⁾ and the peak at 1092 cm⁻¹ was weaker than that of the matrix sample.

Fig. 5.

Cross-sectional Raman spectra of products after 48 h corrosion. (Online version in color.)

Previous work showed that amorphous Cr(OH)₃ and FeCO₃ crystals were formed on the Cr steel surface in CO₂ environment, and they protected the steel from further corrosion.¹⁹⁾ The matrix sample product was composed of Cr₂O₃ and FeCO₃, in which Cr₂O₃ was dehydrated from Cr(OH)₃. Except for Cr₂O₃, the oxide scale sample had the strongest characteristic peak of Fe₃O₄. In comparison, the FeCO₃ peak of the oxide scale sample was weaker than that of the matrix sample.

3.2.3. XRD Test

The oxide scale XRD test result after 96 h of corrosion is displayed in Fig. 6. After the reaction, the strongest peak was replaced by FeCO₃, and the FeCr phase signal became stronger. It indicates that the oxide scale was involved in the reaction and became consumed, resulting in the exposure of more matrix phases and the production of a large amount of FeCO₃.

Fig. 6.

XRD test results of oxide scale surface after 96 h corrosion.

3.2.4. XPS Analysis

Figures 7, 8, 9, 10, 11 display the high-resolution XPS spectra of the corrosion products of the oxide scale (C, Ca, Cr, Fe, O) at different stages. It is noticeable from the C spectra in Fig. 7(a) that a small peak with higher energy moved to the high-energy direction with the extension of the corrosion time. The peak at 284.8 eV that appeared after six hours of corrosion was associated with C–C and C–H bonds (Fig. 7(b)), and it was the impurity peak caused by the cleaning and drying of the sample.

Fig. 7.

XPS spectra of C in corrosion products of the oxide scale: (a) The trend of spectra changes with corrosion times, (b) Curve fitted C 1s spectra of 6 h and 216 h. (Online version in color.)

Fig. 8.

XPS spectra of Ca in corrosion products of the oxide scale: (a) The trend of spectra changes with corrosion times, (b) Curve fitted Ca 2p spectra of 6 h and 216 h. (Online version in color.)

Fig. 9.

XPS spectra of Cr in corrosion products of the oxide scale: (a) The trend of spectra changes with corrosion times, (b) Curve fitted Cr 2p spectra of 6 h and 216 h. (Online version in color.)

Fig. 10.

XPS spectra of Fe in corrosion products of the oxide scale: (a) The trend of spectra changes with corrosion times, (b) Curve fitted Fe 2p spectra of 6 h and 216 h. (Online version in color.)

Fig. 11.

XPS spectra of O in corrosion products of the oxide scale: (a) The trend of spectra changes with corrosion times, (b) Curve fitted O 1s spectra of 6 h and 216 h. (Online version in color.)

The peak at 286.4 eV was associated with the combination of carbon and oxygen, and the peak at 288.4 eV appeared from carbon in carbonate.²⁰⁾ The peak at 289.6 eV that appeared after 216 h was also associated with carbonate,²¹⁾ and the peak at 285.9 eV was related to the deposition of Cr on the C surface.²²⁾ In summary, Carbonate precipitates existed in initial corrosion products; however, the peak position of carbonate changed with time. It indicates that although carbonate existed in corrosion products in each stage, the carbonate type changed slightly.

Figure 8 displays the fine XPS spectra of Ca, and it was composed of a pair of 2p peaks in the same position after different corrosion times (Fig. 8(a)). The peaks at 347.4 eV and 351.1 eV appeared from CaCO₃,²³⁾ implying that CaCO₃ existed in corrosion products at any time of corrosion.

It is observable from Fig. 9(a) that the XPS spectra of Cr after different corrosion times were composed of a pair of 2p peaks. Although the position of these peaks changed slightly with the extension of the corrosion time, through fitting (Fig. 9(b)), it was found that all of them correspond to Cr₂O₃ or Cr(OH)₃.²¹⁾ If these peaks correspond to Cr(OH)₃, then it must came from the products formed during the corrosion process. However, if these peaks correspond to Cr₂O₃, it might have two sources. One source is Cr₂O₃ in the original oxide scale, and the other source is produced by dehydration of Cr(OH)₃. From a single XPS result, the state and source of the Cr compound cannot be inferred.

Figure 10 exhibits the fine XPS spectra of Fe in corrosion products, and it was composed of a pair of 2p peaks. As shown in Fig. 10(a), with the extension of the corrosion time, the high-energy peaks moved to the low-energy direction, whereas the low-energy peaks moved to the high-energy direction. After six hours of corrosion, the 2p_3/2 peak at 711.3 eV appeared from Fe₃O₄,²⁴⁾ Fe₂O₃,²⁵⁾ or FeCO₃,²¹⁾ and the peak at 724.7 eV appeared from Fe₂O₃.²⁴⁾ At 216 h, the 2p_3/2 peak at 711.4 eV appeared from Fe₃O₄,²⁴⁾ Fe₂O₃²⁶⁾ or FeCO₃,²⁷⁾ and the peak at 724.4 eV appeared from FeCO₃.^27,28) Based on the results of XRD and Raman, Fe₂O₃ and Fe₃O₄ were present in initial corrosion products, and there might be a small amount of FeCO₃, Fe₂O₃ occupied the majority of these three substances. In the later stage of corrosion, FeCO₃ became the major element of corrosion products, while Fe₂O₃ and Fe₃O₄ became the minority.

As shown in Fig. 11(a), in the XPS spectra of O, a small peak was detected at low energies in the early stage of corrosion. As the corrosion progressed, this peak decreased and, finally, disappeared. Larger peaks gradually shifted toward higher energies as the reaction progressed. After six hours of corrosion, the peaks at 529.8 eV, 531.3 eV, and 532.4 eV appeared from Fe₂O₃,²⁵⁾ CaCO₃,²⁹⁾ and Cr₂O₃,³⁰⁾ respectively. After 216 h, the peaks at 530.2 eV appeared from FeCr₂O₄ or Fe₃O₄,³¹⁾ 531.4 eV appeared from FeCO₃,³²⁾ 532.2 eV appeared from Cr(OH)₃ or Cr₂O₃,²⁸⁾ and 533.2 eV appeared from Cr₂O₃.^21,33)

In summary, the XPS results show that the content of Fe₂O₃ decreased continuously, whereas Fe₃O₄ was retained in the later stage of corrosion. In the early stage of corrosion, CO₃²⁻ was combined with Ca²⁺ and was deposited on the steel surface, whereas in the later stage of corrosion, it mainly formed FeCO₃ with Fe²⁺. At the same time, the amount of Cr(OH)₃ and Cr₂O₃ increased continuously.

3.2.5. Analysis of Microstructure and Elemental Distribution of Corrosion Products

Figures 12, 13, and 14 respectively show corrosion morphology (and element distribution) of the surface, cross section, and the removal scale surface of 5Cr steel oxide scale samples at different corrosion stages.

Fig. 12.

Surface corrosion morphology and elemental composition of 5Cr steel oxide scale samples at different stages (a) 6 h, (b) 24 h, (c) 48 h, (d) 96 h, (e) 216 h. (Online version in color.)

Fig. 13.

Cross-section corrosion morphology and elemental distribution of 5Cr steel oxide scale samples at different stages (a) 6 h, (b) 24 h, (c) 48 h, (d) 96 h, (e) 216 h. (Online version in color.)

Fig. 14.

Surface corrosion morphology after product removal of 5Cr steel oxide scale samples at different stages (a) 6 h, (b) 24 h, (c) 48 h, (d) 96 h, (e) 216 h.

Figure 12(a) reveals that the surface of the sample was piled up by loose fragments of different sizes in the initial stage of corrosion. EDS shows that these fragments are iron oxides. The cross-sectional view shows that these oxides were partially flaked off (Fig. 13(a)). And after the removal of the scale, no obvious corrosion occurred (Fig. 14(a)).

Figure 12(b) expresses that the oxide scale continued to decompose after 24 h of corrosion. The fine iron oxide had been peeled off, and the large internal object remained, which was also an iron oxide. Due to the falling off of some oxides, the oxide scale became separated from the matrix at some locations and the Ca element in the solution had infiltrated into the oxide scale (Fig. 13(b)). Figure 14(b) reflects that pitting corrosion had occurred on the matrix under the scale.

A large amount of Ca was deposited on the sample surface in the form of coarse particles after 48 h (Fig. 12(c)). In addition, fine corrosion products were deposited in the gaps between large particles. Figure 13(c) reflects that the oxide scale was further separated from the matrix. The scale was mainly composed of O, Cr, Ca, but Fe content was low. Moreover, Cr and O accumulated on the side of the matrix. According to Fig. 14(c), the pitting corrosion was not further deepened. The matrix was dissolved and became smooth after dissolution.

After 96 h of corrosion, a new, smaller cube began to deposit heavily on the sample Fig. 12(d). EDS shows that these cubes had a higher Fe content but lower Ca. According to the results of the phase test, the smaller cubes were FeCO₃, and the bigger cubes were CaCO₃. At this stage, Cr-containing products were detected on the sample surface, a variety of products were deposited in the product film (Fig. 13(d)). And the bottom matrix became relatively flat and smooth (Fig. 14(d)).

Figure 12(e) reveals that in the later stage of corrosion, coarse CaCO₃ particles were re-deposited on the sample surface and granular ferrous carbonate particles filled the gaps of CaCO₃. The EDS analysis (Fig. 13(e)) reveals that the corrosion product film was divided into two layers: the inner layer was rich in Cr and O, whereas the outer layer was enriched with Fe and Ca. This structure was similar to the product film formed by the bare steel.¹⁹⁾ Figure 14(e) shows that after a long period of corrosion, the type of corrosion was uniform corrosion without obvious pitting corrosion.

4. Discussion

The following reactions mainly occurred in the medium-Cr steel in the CO₂ environment:^34,35)

Cathodic reaction:   

$${\mathrm{H}}_2\mathrm{C}{\mathrm{O}}_3+{\mathrm{e}}^{-}\to \mathrm{H}\mathrm{C}{\mathrm{O}}_3{}^{-}+{\mathrm{H}}_{\mathrm{a}\mathrm{d}}$$(7)

  

$$\mathrm{H}\mathrm{C}{\mathrm{O}}_3{}^{-}+{\mathrm{e}}^{-}\to \mathrm{C}{\mathrm{O}}_3{}^{2-}+{\mathrm{H}}_{\mathrm{a}\mathrm{d}}$$(8)

  

$$\mathrm{H}\mathrm{C}{\mathrm{O}}_3{}^{-}+{\mathrm{e}}^{-}+{\mathrm{H}}_{\mathrm{a}\mathrm{d}}\to \mathrm{C}{\mathrm{O}}_3{}^{2-}+{\mathrm{H}}_2$$(9)

  

$${\mathrm{H}}_2\mathrm{O}+{\mathrm{e}}^{-}\to \mathrm{O}{\mathrm{H}}^{-}+{\mathrm{H}}_{\mathrm{a}\mathrm{d}}$$(10)

Anode reaction:   

$$\mathrm{F}\mathrm{e}+\mathrm{O}{\mathrm{H}}^{-}\to \mathrm{F}\mathrm{e}\mathrm{O}{\mathrm{H}}_{\mathrm{a}\mathrm{d}}+{\mathrm{e}}^{-}$$(11)

  

$$\mathrm{F}\mathrm{e}\mathrm{O}{\mathrm{H}}_{\mathrm{a}\mathrm{d}}\to \mathrm{F}\mathrm{e}\mathrm{O}{\mathrm{H}}^{+}+{\mathrm{e}}^{-}$$(12)

  

$$\mathrm{F}\mathrm{e}\mathrm{O}{\mathrm{H}}^{+}\to \mathrm{F}{\mathrm{e}}^{2+}+\mathrm{O}{\mathrm{H}}^{-}$$(13)

  

$$\mathrm{C}\mathrm{r}\to \mathrm{C}{\mathrm{r}}^{3+}+3\mathrm{e}$$(14)

Additional reactions:   

$$\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}\to {\mathrm{H}}_2\mathrm{C}{\mathrm{O}}_3$$(15)

  

$$2{\mathrm{H}}_{\mathrm{a}\mathrm{d}}\to {\mathrm{H}}_2$$(16)

  

$$\mathrm{F}{\mathrm{e}}^{2+}+\mathrm{C}{\mathrm{O}}_3{}^{2-}\to \mathrm{F}\mathrm{e}\mathrm{C}{\mathrm{O}}_3$$(17)

  

$$\mathrm{C}{\mathrm{r}}^{3+}+3{\mathrm{H}}_2\mathrm{O}\to \mathrm{C}\mathrm{r}{\left(\mathrm{O}\mathrm{H}\right)}_3+3{\mathrm{H}}^{+}$$(18)

  

$$\mathrm{C}{\mathrm{a}}^{2+}+\mathrm{C}{\mathrm{O}}_3{}^{2-}\to \mathrm{C}\mathrm{a}\mathrm{C}{\mathrm{O}}_3$$(19)

  

$$2\mathrm{C}\mathrm{r}{\left(\mathrm{O}\mathrm{H}\right)}_3\to \mathrm{C}{\mathrm{r}}_2{\mathrm{O}}_3+3{\mathrm{H}}_2\mathrm{O}$$(20)

The cathodic reaction of steel in the CO₂ aqueous solution was mainly the ionization reaction of CO₂ in the aqueous solution to produce CO₃²⁻ and HCO₃⁻ (Eqs. (7), (8), (9)). In addition, there was also ionization of water (Eq. (10)). The anodic reaction was mainly the oxidation process of Fe (Eqs. (11), (12), (13)). And, the oxidation process of Cr also existed in medium-Cr steel (Eq. (14)). When the concentration of FeCO₃ and Cr(OH)₃ in the solution exceeded the critical supersaturation, the corrosion product film would be precipitated. Corrosion products of the medium-Cr steel in the CO₂ environment were composed of two layers. A Cr-rich amorphous inner layer was formed in the early stage of corrosion, and crystal grains were then deposited on top of the inner amorphous layer; hence, an outer layer began to form with the prolonged corrosion time.^28,36) With reference to the corrosion product film formation process of medium-Cr steel, Fig. 15 schematically presents the possible corrosion mechanism of the medium-Cr steel with the oxide scale.

Fig. 15.

Schematic representation of the corrosion behavior of the rolled oxide scale of 5Cr steel in the CO₂ environment: (a), (b), (c), early stage, (d) medium stage, (e) late stage. (Online version in color.)

When the corrosion began, the oxide scale sample was in a state of rapid mass loss. The reason for this mainly comes from two aspects. One is that the hydrogen bubbles flake off the oxide scale from the surface of the sample, and the other is that the H⁺ in the solution reacts with the oxide scale to dissolve it. The hydrogen bubbles come from the H_ad (Eqs. (9), (16)) which is produced by cathode reaction (Eqs. (7), (8), and (10)). Hydrogen ions are derived from the ionization of carbonic acid and bicarbonate, and the anode reaction (Eq. (18)) between exposed matrix and CO₂ aqueous solution. As the corrosion progressed, the mass loss rate decreased rapidly. The flaking of the oxide scale mainly occurred in the middle layer of Fe₂O₃, whereas Fe₃O₄ was retained to a certain extent (Fig. 15(a)).

The decomposition of the oxide scale occurred preferentially in the early stage of corrosion, which delayed the corrosion of the matrix under the oxide scale. At the same time, the gap formed by the flaking of the oxide scale provided the space for the deposition of CaCO₃, as Fig. 15(b), Ca²⁺ and CO₃²⁻ in the solution formed large particles and were deposited on the steel surface. With the flaking of iron oxides and the deposition of coarse CaCO₃, the oxide sheet became separated from the matrix, the bottom matrix got in contact with the corrosive solution, and finally, the bottom matrix began to dissolve.

The dissolution of the partially exposed bottom matrix caused pitting corrosion, and the dissolved matrix released Fe²⁺ and Cr³⁺ ions (Eqs. (11), (12), (13), (14)), which formed FeCO₃ and Cr(OH)₃ in the end, respectively (Eqs. (17)–(18)). Cr(OH)₃ precipitated first. The production of Cr(OH)₃ prevented the further progress of pitting corrosion (Fig. 15(c)); thus, the corrosion only occurred along with the interface between the oxide scale and the matrix. This phenomenon further separated the oxide scale and the matrix, and eventually, the interface became relatively flat.

The deposition of Cr(OH)₃ decreased the pH of the nearby solution.³⁷⁾ A part of the CaCO₃ dissolved, thus, CO₃²⁻ ions had an opportunity to redistribute between Fe²⁺ and Ca²⁺. As shown in Fig. 15(d), when the concentration of these three ions reached the supersaturation stage, FeCO₃, and CaCO₃ (or Fe_xCa_yCO₃ (x+y=1)³⁸⁾) nucleated and grew on the surface of residual Fe₃O₄ (or FeCr₂O₄). An outer product film enriched with Fe, Ca, O and an inner film enriched with Cr, O were gradually formed.

With the extension of the corrosion time, an outer film composed of a large amount of FeCO₃ with a small amount of CaCO₃ (or Fe_xCa_yCO₃ (x+y=1)) and an inner film composed of Cr(OH)₃ were formed (Fig. 15(e)), and this structure was similar to the product film formed on the bare steel. Therefore, the dense oxide layer delayed the occurrence of corrosion, and under long-term corrosion, similar to the bare steel, it eventually formed a double-layer product film.

5. Conclusion

The rolled oxide scale of 5Cr steel was divided into three layers: the innermost layer contained thin Fe₃O₄, the middle layer was composed of thick Fe₂O₃, and the outer layer contained Fe₃O₄ and a small amount of Fe₂O₃. Cr₂O₃ existed between the inner layer and the center of the middle layer and on the outer layer. The reaction of the oxide scale of 5Cr steel in the CO₂ environment occurred in two stages. When the corrosion time was less than 48 h, the flaking of the oxide scale occurred. When the corrosion time was longer than 48 h, the deposition of corrosion products occurred. The dense oxide layer delayed the occurrence of corrosion, and under long-term corrosion, similar to the bare steel, it formed a double-layer product film.

Acknowledgments

This work was financially supported by the National Key R and D Program of China (No. 2017YFB0304900) and Technology Development Program of Weifang (No. 2019GX077).

References

• 1)   M. H.  Sk,  A. M.  Abdullah,  J.  Qi,  M.  Ko,  B.  Ingham,  N.  Laycock,  M. P.  Ryan and  D. E.  Williams: J. Electrochem. Soc., 165 (2018), C278. https://doi.org/10.1149/2.1011805jes
• 2)   Z.  Liu,  X.  Gao,  J.  Li,  L.  Du,  C.  Yu,  P.  Li and  X.  Bai: Electrochim. Acta, 213 (2016), 842. https://doi.org/10.1016/j.electacta.2016.08.024
• 3)   C.  Tan,  X.  Xu,  L.  Xu,  C.  Yin and  L.  Qiao: Mater. Corros., 70 (2019), 366. https://doi.org/10.1002/maco.201810420
• 4)   S.  Li,  Z.  Zeng,  M. A.  Harris,  L. J.  Sánchez and  H.  Cong: Front. Mater., 6 (2019), 10. https://doi.org/10.3389/fmats.2019.00010
• 5)   J.  Yuan,  K.  Zhu,  J.  Jiang,  H.  Zhang,  A.  Fu,  W.  Li,  C.  Yin and  F.  Li: Mater. Res. Express, 7 (2020), 106518. https://doi.org/10.1088/2053-1591/abbfbc
• 6)   X.  Chen,  Z.  Guan,  M.  Du and  C.  Lin: Int. J. Electrochem. Sci., 14 (2019), 7332. https://doi.org/10.20964/2019.08.87
• 7)   Y.  Hua,  S.  Xu,  Y.  Wang,  W.  Taleb,  J.  Sun,  L.  Zhang,  R.  Barker and  A.  Neville: Corros. Sci., 157 (2019), 392. https://doi.org/10.1016/j.corsci.2019.06.016
• 8)   B.  Luo,  J.  Zhou,  P.  Bai,  S.  Zheng, T. An and X. Wen: Int. J. Miner. Metall. Mater., 24 (2017), 646. https://doi.org/10.1007/s12613-017-1447-9
• 9)   Y.  Hua,  R.  Barker and  A.  Neville: Appl. Surf. Sci., 356 (2015), 499. https://doi.org/10.1016/j.apsusc.2015.08.116
• 10)   X. Q.  Lin,  W.  Liu,  J.  Zhang,  S.  Dong,  H. L.  Zhang,  X. B.  Li,  C. C.  Xu and  M. X.  Lu: Acta Phys. Chim. Sin., 29 (2013), 2405 (in Chinese). https://doi.org/10.3866/PKU.WHXB201309171
• 11)   A. B.  Forero,  M. M. G.  Núñez and  I. S.  Bott: Mater. Res., 17 (2014), 461. https://doi.org/10.1590/S1516-14392013005000182
• 12)   J.  Han,  D.  Young,  H.  Colijn,  A.  Tripathi and  S.  Nešić: Ind. Eng. Chem. Res., 48 (2009), 6296. https://doi.org/10.1021/ie801819y
• 13)   L.  Wei,  L.  Han,  L.  Chen and  Y.  Zhao: Procedia Manuf., 15 (2018), 1588. https://doi.org/10.1016/j.promfg.2018.07.308
• 14)   B.  Hua,  Y.  Kong,  W.  Zhang,  J.  Pu,  B.  Chi and  L.  Jian: J. Power Sources, 196 (2011), 7627. https://doi.org/10.1016/j.jpowsour.2011.05.007
• 15)   L.  Xin,  B.  Yang,  J.  Li,  Y.  Lu and  T.  Shoji: Corros. Sci., 123 (2017), 116. https://doi.org/10.1016/j.corsci.2017.04.009
• 16)   M.  Hanesch: Geophys. J. Int., 177 (2009), 941. https://doi.org/10.1111/j.1365-246X.2009.04122.x
• 17)   Z.  Tan,  D.  Zhang,  L.  Yang,  Z.  Wang,  F.  Cheng,  M.  Zhang, Y. Jin and S. Zhu: Tribol. Int., 146 (2020), 106145. https://doi.org/10.1016/j.triboint.2019.106145
• 18)   B.  David,  O.  Schneeweiss,  N.  Pizúrová,  Šantavá,  V. Kudrle, P. Synek and O. Jašek: Phys. Procedia, 44 (2013), 206. https://doi.org/10.1016/j.phpro.2013.04.025
• 19)   Y.  Gu,  H.  Wu,  R.  Yuan,  X.  Wang,  T.  Wu and  S.  Shuai: Mater. Res. Express, 6 (2019), 046511. https://doi.org/10.1088/2053-1591/aaf969
• 20)   A. F.  Avelino,  W. S.  Araújo,  D. F.  Dias,  L. P. M.  dos Santos,  A. N.  Correia and  P.  de Lima-Neto: Electrochim. Acta, 286 (2018), 339. https://doi.org/10.1016/j.electacta.2018.08.042
• 21)   S.  Guo,  L.  Xu,  L.  Zhang,  W.  Chang and  M.  Lu: Corros. Sci., 110 (2016), 123. https://doi.org/10.1016/j.corsci.2016.04.033
• 22)   J. L.  Jordan,  P. N.  Sanda,  J. F.  Morar,  C. A.  Kovac,  F. J.  Himpsel and  R. A.  Pollak: J. Vac. Sci. Technol. A, 4 (1986), 1046. https://doi.org/10.1116/1.573451
• 23)   M. I.  Sosulnikov and  Y. A.  Teterin: Dokl. Akad. Nauk SSSR, 317 (1991), 418. https://www.researchgate.net/publication/285423985_XPS_Study_of_Calcium_Strontium_Barium_and_Their_Oxides
• 24)   F. F.  Eliyan and  A.  Alfantazi: Corros. Sci., 85 (2014), 380. https://doi.org/10.1016/j.corsci.2014.04.055
• 25)   D. A.  López,  W. H.  Schreiner,  S. R.  De Sánchez and  S. N.  Simison: Appl. Surf. Sci., 207 (2003), 69. https://doi.org/10.1016/S0169-4332(02)01218-7
• 26)   L.  Wei,  X.  Pang,  C.  Liu and  K.  Gao: Corros. Sci., 100 (2015), 404. https://doi.org/10.1016/j.corsci.2015.08.016
• 27)   S.  Guo,  L.  Xu,  L.  Zhang,  W.  Chang and  M.  Lu: Corros. Sci., 63 (2012), 246. https://doi.org/10.1016/j.corsci.2012.06.006
• 28)   B.  Dong,  D.  Zeng,  Z.  Yu,  L.  Cai,  S.  Shi,  H.  Yu,  H.  Zhao and  G.  Tian: J. Mater. Eng. Perform., 28 (2019), 1030. https://doi.org/10.1007/s11665-018-3847-z
• 29)   B.  Demri and  D.  Muster: J. Mater. Process. Technol., 55 (1995), 311. https://doi.org/10.1016/0924-0136(95)02023-3
• 30)   X.  Yue,  L.  Zhang,  C.  Sun,  S.  Xu,  C.  Wang,  M.  Lu,  A.  Neville and  Y.  Hua: Corros. Sci., 169 (2020), 108640. https://doi.org/10.1016/j.corsci.2020.108640
• 31)   M. C.  Biesinger,  B. P.  Payne,  A. P.  Grosvenor,  L. W. M.  Lau,  A. R.  Gerson and  R. S. C.  Smart: Appl. Surf. Sci., 257 (2011), 2717. https://doi.org/10.1016/j.apsusc.2010.10.051
• 32)   Y.  Tang,  X. P.  Guo and  G. A.  Zhang: Corros. Sci., 118 (2017), 118. https://doi.org/10.1016/j.corsci.2017.01.028
• 33)   J.  Sun,  C.  Sun,  X.  Lin,  X.  Cheng and  H.  Liu: Materials, 9 (2016), 200. https://doi.org/10.3390/ma9030200
• 34)   Z.  Liu,  X.  Gao,  L.  Du,  J.  Li,  X.  Zhou,  X.  Wang,  Y.  Wang,  C.  Liu,  G.  Xu and  R. D. K.  Misra: Appl. Surf. Sci., 440 (2018), 974. https://doi.org/10.1016/j.apsusc.2018.01.223
• 35)   T.  Wu,  H.  Wu,  G.  Niu,  T.  Li,  R.  Sun,  Y.  Gu and  R.  Yuan: Corrosion., 74 (2018), 757. https://doi.org/10.5006/2595
• 36)   L.  Xu,  S.  Guo,  W.  Chang,  T.  Chen,  L.  Hu and  M.  Lu: Appl. Surf. Sci., 270 (2013), 395. https://doi.org/10.1016/j.apsusc.2013.01.036
• 37)   J.  Zhu,  L.  Xu,  M.  Lu and  W.  Chang: Int. J. Miner. Metall. Mater., 26 (2019), 1405. https://doi.org/10.1007/s12613-019-1861-2
• 38)   R.  Rizzo and  R.  Ambat: Mater. Corros., 72 (2021), 1076. https://doi.org/10.1002/maco.202011822