Effect of Fe ion concentration on corrosion of carbon steel in CO 2 environment

Abstract

In this work, the corrosion behaviour of steel wires in solutions containing different concentrations of Fe²⁺ was investigated by the linear polarisation resistance method, while the evolution of pH was monitored in situ and changes of the Fe²⁺ concentration were monitored ex situ. Characterisation of the corrosion scales was performed using microscopic and diffraction techniques. Scale analysis revealed that the passivation of samples, exposed to initially highly Fe²⁺ supersaturated solution, occurred when a formation of a double layer took place, resulting in 30 times lower corrosion rate compared to samples tested in solutions without initial Fe²⁺.

Keywords

Carbon steel Carbon dioxide Supersaturation Linear polarisation resistance Scanning electron microscopy X-ray diffraction Passivation

Introduction

Flexible pipes consist of concentric layers of polymers and steel, which are not bonded together to preserve the flexibility of the pipe.¹ The fatigue life of flexible pipes, which is determined by the fatigue performance of steel armours, is an important issue. The fatigue resistance of tensile armours is determined by the applied stresses and operating corrosive environment.² As a result of diffusion of small molecules such as methane (CH₄), carbon dioxide (CO₂), hydrogen sulphide (H₂S) and water (H₂O) through the inner liner into the annulus, corrosion of steel armours might occur.^1–3 The confined environment, which steel armours are exposed to in the annulus region, is characterised by a low ratio of free water volume (V) to steel surface area (S), typically < 0.1 mL cm^{− 2}.³ Under corrosive conditions, the low V/S ratio results in a fast supersaturation of Fe²⁺ in the liquid and corrosion rates are typically < 10 μm year^{− 1}.³ In the annulus region, unlike bore conditions, there is no vigorous flow of water, and even though the pipe moves due to waves and water current, it is believed that conditions are relatively stagnant.

In the aqueous CO₂ environment, the corrosion process of carbon steel occurs through an electrochemical process, which involves dissolution of iron (equation (1)) and cathodic evolution of hydrogen.⁴ In the presence of CO₂, the corrosion rate of carbon steel is increased by a corresponding increase of the rate of the hydrogen evolution reaction.^4–6 It has been reported^4,5,7 that cathodic reactions, which take place in aqueous CO₂ environment, contain two independent reactions, such as hydrogen reduction (equation (2)) and carbonic acid reduction (H₂CO₃, equation (3)). It is, however, still not certain whether H₂CO₃ is directly reduced or whether it dissociates to form bicarbonate (HCO₃ ⁻). (1) (2) (3) The most common corrosion product that formed in an aqueous CO₂ environment is iron carbonate (FeCO₃). When the concentration of Fe²⁺ and CO₃ ^{2 −} exceeds the solubility limit, a formation of FeCO₃ takes place^4,8–10 according to the reaction presented in equation (4). (4) There are many factors that influence the formation of FeCO₃, and one of the most important is the chemistry of the water.⁴ The rate of FeCO₃ precipitation is controlled by the crystal growth rate,⁹ and in order to ensure fast scale precipitation and, consequently, low corrosion rates, high supersaturation (SR) with respect to FeCO₃ must be provided as presented by equation (5), where R _gr is the crystalline film growth rate and k _gr is the crystal growth constant.^6,9,11 (5) Supersaturation (equation (6)) is often described as a driving force for crystal growth,¹² and in order to have a precipitation of FeCO₃, it must be much larger than unity and the growth rate constant (k _gr) must be large as well.⁹ SR, as shown from equation (6), is influenced by Fe²⁺ and CO₃ ^{2 −} concentrations. The solubility product (K _sp) of FeCO₃, which is also present in equation (6), depends on the temperature and ionic strength.¹⁰ (6) In order to ensure high SR and formation of protective FeCO₃ scale, the concentration of Fe²⁺ and pCO₂ must be relatively high, and the pH should be larger than 5.9 It has been reported^13,14 that the Fe²⁺ concentration in the solution is one of the main factors that determine the kinetics of FeCO₃ formation and resulting corrosion rates. Results show^13,14 that exposing samples to solutions saturated or nearly saturated with Fe²⁺ lead to formation of protective FeCO₃, while no development of FeCO₃ was seen in solutions with Fe²⁺ content below the solubility limit of FeCO₃. However, the kinetics of FeCO₃ formation are also dependent on the exposing temperature. At room temperature, the kinetics of scale formation are very slow, and longer times are needed to develop protective scales, or a non-protective scale might be formed even when ensuring high SR of the solution.^9,15–17 Owing to the high precipitation rate of FeCO₃ at temperatures >60°C, dense and protective scales are formed even at low supersaturations and, consequently, low corrosion rates can be achieved.^4,8,12,15,16 It has been reported^7,9,18–20 that the pH of the solution also has a big impact on the scale formation and it influences the solubility limit of FeCO₃ and resulting corrosion rates. Nešić et al. ⁹ have reported that in order for a FeCO₃ precipitation to occur, the pH has to exceed a critical value, which depends on Fe²⁺ concentration, temperature and ionic strength. When increasing the pH, the concentration of H⁺ is decreased and the concentration of CO₃ ^{2 −} is increased. It results in a lower solubility limit of FeCO₃ (Ref. 20) and consequently higher supersaturation, which leads to faster and more protective scale precipitation.^21–23 Videm and Dugstad¹⁴ have shown that by increasing pH from 4.2 to 7, the corrosion rate was decreased by a factor of 3 in a solution saturated with Fe²⁺. It is in clear contrast to solutions with low pH, where the solubility limit of FeCO₃ is higher and more Fe²⁺ is needed for forming FeCO₃.²² When uniform corrosion of carbon steel takes place, the anodic and cathodic current densities are equal, which indicates that Fe²⁺ and HCO₃ ⁻ are produced simultaneously in the same amounts and in the same place.²⁴ In the non-buffered solutions, where pH can spontaneously evolve, the pH is directly related to the alkalinity of the solution, i.e. to the concentration of HCO₃ ⁻, which is associated with the amount of Fe²⁺ present in the solution.

The other very important factor that influences the scale formation is the volume to steel area ratio (V/S ratio). The V/S ratio has a significant impact on the time needed to obtain high supersaturation of Fe²⁺ in the solution. It has been reported^3,15,17 that for tests performed with low V/S ratio, the time needed to obtain high supersaturation of Fe²⁺, and thus to exceed the solubility limit of FeCO₃, was shorter compared to tests performed with high V/S ratio. Nevertheless, Rubin et al. ³ have reported that by exposing samples to V/S ratio of 40 mL cm^{− 2} and a solution initially supersaturated with Fe²⁺, very low corrosion rates of 20 μm year^{− 1} were observed, which were comparable to corrosion rates (10 μm year^{− 1}) obtained for samples exposed to low V/S ratio of 0.22 mL cm^{− 2}. These results are particularly important when performing corrosion fatigue testing on a laboratory scale. In the corrosion fatigue tests, the low V/S ratio observed in the annulus is difficult to obtain due to space limitations in the test vessel. Consequently, in order to simulate high supersaturation of Fe²⁺, which is present in the annulus of a flexible pipe, samples are exposed to solutions initially supersaturated with Fe²⁺. It is believed that the correct simulation of a confined environment has a significant impact on the corrosion and, consequently, the fatigue behaviour of steel wires; therefore, it is important to know the influence of the initial concentration of Fe²⁺ on the fatigue resistance of steel wires. The work presented in this paper is part of a study on corrosion fatigue mechanisms in aqueous CO₂ environment. In order to distinguish the difference between the impacts of corrosion processes from the impact of dynamic loading on the fatigue life of steel wires, the first experiments were performed in load free conditions and are presented in this paper. Further results from the experiments performed in the same corrosive media, while testing samples dynamically, will be published elsewhere.

The focus of this paper was to investigate the influence of the initial Fe²⁺ concentration and corresponding alkalinity of the solution (HCO₃ ⁻), together with the resulting pH value on the scale morphology and the corrosion resistance of steels used in flexible pipes. Experiments were performed with different initial Fe²⁺ and HCO₃ ⁻ concentrations under high supersaturated conditions. Variation of corrosion rates as a function of exposure time was measured using linear polarisation resistance (LPR), which was performed during the entire corrosion tests in solutions saturated with CO₂ under atmospheric pressures at 20°C. The pH evolution was measured in situ and supported with measurements of Fe²⁺ concentrations. Electrochemical measurements were supported with the analysis of corrosion scales using scanning electron microscopy (SEM) and X-ray diffraction (XRD) studies. The commercial Multiscale model was used for thermodynamic calculations to understand the limits of solubility of FeCO₃ in the presence of CO₂ and compared with levels determined experimentally.

Experimental

Experimental set-up for corrosion testing

The experimental set-up consisted of a de-aeration vessel, a pre-corrosion vessel and an electrochemical cell (Fig. 1). All vessels were made of glass and equipped with 316L steel tubing to avoid any ingress of oxygen. The electrochemical cell contained a saturated calomel electrode (SCE) (REF 421 Reference Electrode, Calomel) as a reference electrode, a platinum counter electrode with a surface area of 310 cm² and a rectangular carbon steel specimen as a working electrode with a surface area of 6 cm². Apart from the working electrode, the electrochemical cell had three additional samples (each 49.5 cm²), which were used for the scale analysis, and one weight loss coupon used for mass loss determination (6 cm²). Additionally, the set-up was equipped with a pH electrode to monitor the pH changes during the entire test.

a de-aeration chamber; b pre-corrosion chamber; c electrochemical cell (1: pH electrode, 2: working electrode, 3: counter electrode, 4: reference electrode, 5: wash bottle)

Material and specimen preparation

The material used for the electrochemical experiments was carbon steel with a composition of 0.77 wt-% C, 0.24 wt-% Si, 0.65 wt-% Mn, 0.17 wt-% P, 0.06 wt-% S, Fe as the balance. The yield strength of the material is 1350 MPa. As received, from cold rolling process (R _a = 0.38 ± 0.05 μm), surface conditions are used for the experiment to simulate material conditions under use in flexible pipes. Before each experiment, the electrode surface was thoroughly degreased with acetone, cleaned with ethanol and dried in a hot air flow.

Electrolyte solution

Electrochemical experiments were conducted in artificial sea water with a chemical composition according to ASTM D1141 (Table 1), under 1 bar CO₂ conditions. The flowrate of CO₂ gas was 250 mL min^{− 1}. The temperature of the electrolyte solution was set to 20 ± 1°C. The solution was de-aerated in the de-aeration vessel by purging nitrogen gas until the oxygen content was < 5 ppb. The dissolved oxygen was measured using optical oxygen sensor (Visiferm DO 225).

Table 1

Chemical composition of sea water (ASTM D1141)

	NaCl	MgCl₂	Na₂SO₄	CaCl₂	KCl	NaHCO₃	KBr	H₃BO₃	SrCl₂	NaF
Composition/mg L^{− 1}	24 530	5200	4090	1160	695	201	101	27	25	3

The whole set-up was flushed with nitrogen in order to remove any oxygen. Electrochemical experiments were carried out with four starting Fe²⁺ concentrations, namely, 0, 370, 514 and 1300 mg L^{− 1} (the last concentration refers to the high supersaturation condition). The corresponding alkalinity (HCO₃ ⁻ concentration) of the test solutions is shown in Table 3. In order to generate electrolytes containing different concentrations of Fe²⁺, steel wool was exposed to de-aerated sea water under 1 bar CO₂ in the pre-corrosion vessel for different periods of time. Owing to corrosion of the steel wool, the generation of Fe²⁺ together with HCO₃ ⁻ took place. When the solution obtained a specific Fe²⁺ concentration, it was transported to the test vessel. The Fe²⁺ contents were measured by UV–Vis spectrophotometry method using a Shimadzu UV-1800 spectrophotometer, with an error margin of 1%. The pH of the test solutions was measured using a pH meter (Mettler InPro4800i). The water free volume to steel area ratio was 24 mL cm^{− 2} during all the experiments. The volume of the test solution was 3.8 L. In order to ensure the homogenous bulk solution mixing, the electrolyte was stirred at 350 rev min^{− 1}. Details of test solutions are listed in Table 2.

Table 2

Details of test solutions and conditions

Base solution	Fe²⁺/mg L^{− 1}	HCO₃ ⁻/mg L^{− 1}	pH	Temperature/°C	Oxygen/ppb	Exposure time/h
Artificial sea water	0	61*	4.7	20	<5	242
	370	869	5.6			242
	514	1184	5.7			4485
	1300	2901	6.2			242, 1613

* HCO₃ ⁻ concentration in pure sea water.

Electrochemical measurements

In order to perform electrochemical measurements, the CH Instruments Electrochemical Analyzer potentiostat, together with CHI602b software, was used. LPR measurements were carried out as a function of exposure time on the same sample during continued exposure. Potentials were scanned between (20 and 20 mV versus open circuit potential at a scan rate of 0.166 mV s^{− 1}. All potentials were measured with respect to SCE. The intersection of the anodic and cathodic part was measured as E _corr.

The area averaged corrosion current density is calculated according to equation (7), where B is the Stearn–Geary constant, R _p is the polarisation resistance and A is the area of the specimen.²⁵ For each tested sample, the B value was calibrated so that the integral material loss, found from the LPR measurements, fitted with the weight loss obtained experimentally (see Table 3). Depending on the exposure conditions, the B values varied from 20 to 27 mV. (7)

Table 3

Results of corrosion rates from weight loss measurements.

Initial Fe²⁺ in experiments/mg L^{− 1}	Test exposure/h	Corrosion rate/μm year^{− 1}
0	242	1339
370	242	788
1300	242	373
514	4485	165
1300	1613	33

Scale analysis

To assess the morphology of the resulting corrosion scales, cross-sections of the exposed samples were metallographically prepared and investigated by means of SEM (Zeiss EVO MS10) using secondary electrons and backscattered electrons. Qualitative phase analysis of the corrosion products was performed by means of XRD (Bruker AXS Diffractometer D8 Discover). Measurements were performed with Cr K _α radiation and carried out on the surface of the samples, resulting in a penetration depth of up to 17 μm.

Thermodynamic calculations

Thermodynamic calculations were performed with the commercially available Multiscale model (version 8.1) to calculate the solubility limit of FeCO₃ and saturation ratio (SR) with respect to FeCO₃ (see equation (6)) for different Fe²⁺ and corresponding HCO₃ ⁻ concentrations (see Table 4). Calculations were performed for artificial sea water with a chemical composition according to the ASTM D1141 standard (see Table 1) at 1 bar CO₂ at 20°C.

Table 4

Concentrations of Fe²⁺ and HCO₃ ⁻ used for Multiscale calculations

Fe²⁺/mg L^{− 1}	0	100	200	370	500	700	900	1100	1300
HCO₃ ⁻/mg L^{− 1}	61	279.5	497.9	869.3	1153.3	1590.2	2027.1	2464.1	2901

Results

Thermodynamic calculations

Figure 2 presents thermodynamic calculations of pH and SR with respect to FeCO₃ as a function of the initial Fe²⁺ and HCO₃ ⁻ concentration together with the experimental data obtained from the solution analysis of the electrolyte used for electrochemical testing. The solubility limit of FeCO₃, calculated by the Multiscale model, in 1 bar CO₂ at 20°C was equal to 202 mg L^{− 1} of Fe²⁺ and the pH 5.5. Results revealed that for all experiments of the present work, supersaturated conditions existed except for the experiment with initial pH of 4.7 and 0 mg L^{− 1} of Fe²⁺. Calculated results also showed that by increasing the initial Fe²⁺ and HCO₃ ⁻ content in the electrolyte, the SR was similarly increased. It was shown that pH was increasing with Fe²⁺ concentration, which was directly related to increasing HCO₃ ⁻ content in the solution. Overall, the pH measured from the electrochemical experiments and the corresponding Fe²⁺ content were in a good agreement with the calculations.

Thermodynamic calculations made in Multiscale model together with initial pH and Fe²⁺ concentrations from electrochemical experiments

Electrochemical testing and electrolyte analysis

Evolution of Fe²⁺ and pH

Figure 3 presents the evolution of pH and Fe²⁺ as a function of exposure time in the solutions with various initial Fe²⁺ concentrations. The trends of pH evolutions were observed to follow changes in Fe²⁺ concentrations as a direct consequence of changing the HCO₃ ⁻ concentration in the solution. Results from the tests without initial Fe²⁺ (Fig. 3a) and with 370 mg L^{− 1} of Fe²⁺ (Fig. 3b) showed that the pH increased simultaneously with an increase in the Fe²⁺ content in the solution during the time of exposure. Results shown in Fig. 3c (highly supersaturated condition) showed that the pH increased at the beginning of the test and stabilised after 125 h. However, it should be noticed that no data are available between 25 and 125 h, so stabilisation could have occurred earlier. For the prolonged test with initially high supersaturated conditions (Fig. 3d) and 514 mg L^{− 1} of Fe²⁺ (Fig. 3e), pH increased until it reached a maximum level of 6.3 after 250 and 921 h respectively, and it decreased with further exposure time and so the Fe²⁺ concentration does as well.

Evolution of Fe²⁺ content (right axis) and pH (left axis) for tests performed in solutions with initial Fe²⁺ content of a 0 mg L^{− 1}, b 370 mg L^{− 1}, and c 1300 mg L^{− 1} for 242 h; d 1300 mg L^{− 1} for 1613 h of exposure; and e 514 mg L^{− 1} for 4485 h of exposure

Evolution of corrosion potential

To further explore the impact of Fe²⁺ concentrations and pH changes on the electrochemical behaviour of the material and its scale formation, the evolution of the corrosion potential (E _corr) was analysed. The curves in Fig. 4 clearly show that the initial E _corr values were influenced by different initial Fe²⁺ concentration in the corrosive media. The E _corr value for the experiment with highest concentration of Fe²⁺ in the solution had more positive potential in the beginning of the test compared to potentials measured in solutions without initial Fe²⁺ and with 370 mg L^{− 1} (Fig. 4a). For experiments without initial Fe²⁺ and with 370 mg L^{− 1} of Fe²⁺, E _corr increased as a function of exposure time and stabilised after ∼167 and 75 h respectively. For the sample exposed to the highly supersaturated solution (1300 mg L^{− 1}), the measured potential revealed a constant value of − 450 mV during the whole test, and after ∼440 h, it started to drift towards more positive values (Fig. 4b). Further, the corrosion potential for the sample exposed to the solution with 514 mg L^{− 1} of Fe²⁺ (Fig. 4b) increased at the beginning of the test, and after 1820 h, it began declining again, reaching the final value of ∼ − 440 mV. The corrosion potentials measured in the experiments are higher compared to the theoretical values obtained from the Nernst equation, where the potential varies from − 670 mV up to (580 mV for Fe²⁺ content of 1 and 1300 mg L^{− 1} respectively. The calculated potentials represent theoretical equilibrium conditions and, in this case, only take the impact of the Fe²⁺ concentration on the obtained potential into account. However, in the experiments presented in this paper, the equilibrium was not obtained. Additionally, the potential obtained experimentally in the present tests was not only influenced by the concentration of Fe²⁺ but also by the formation of corrosion scale and its protective properties, which could explain differences between experimentally measured and theoretical potentials.

Evolution of E _corr as function of time for exposure tests with different initial Fe²⁺ content that were running for a 242 h and b 1613 and 4485 h

Corrosion rates

Figure 5 presents the corrosion rates as a function of exposure time for samples exposed to solutions with various initial Fe²⁺ concentrations. The maximum corrosion rate was observed for samples tested in solutions containing no initial Fe²⁺(0 mg L^{− 1}), where the corrosion rate increased as a function of time and reached 1905 μm year^{− 1} (Fig. 5a). The addition of 370 mg L^{− 1} of Fe²⁺ to the solution decreased the final corrosion rate to 981 μm year^{− 1} (Fig. 5a). A decrease of the corrosion rate in CO₂ environment with time was first observed when exposing samples to highly supersaturated solution: corrosion rates of 60 μm year^{− 1} were obtained after 242 h (Fig. 5a) and 1.5 μm year^{− 1} after ∼440 h of exposure (Fig. 5b), which correlates very well with increasing corrosion potential after approximately the same time of exposure, as shown in Fig. 4b. Electrochemical experiments revealed that the time needed to obtain a decrease in corrosion rate is considerably longer for the sample exposed to the solution with 514 mg L^{− 1} (826 h) than the one with 1300 mg L^{− 1} (48 h) (Fig. 5b). It should be noticed that for samples tested in a solution with an initial Fe²⁺ content of 514 mg L^{− 1}, a decrease in corrosion rate occurred approximately at the same time when the maximum Fe²⁺ concentration and pH was reached, as presented in Fig. 3e.

Corrosion rates as function of time for tests with different initial Fe²⁺ content that were running for a 242 h and b 1613 and 4485 h

SEM and XRD analysis

The top surface and cross-sectional morphologies of the corrosion scales formed in solutions with various initial Fe²⁺ concentrations are shown in Fig. 6. The thickness of the layer decreased with increase in initial Fe²⁺ content [∼12 μm for solutions with no initial Fe²⁺ (Fig. 6a) and ∼5 μm in highly supersaturated environment (Fig. 6e)]. Depending on the exposure conditions, the corrosion scales exhibited differences in morphologies. For samples exposed to solutions without initial Fe²⁺ and with 370 mg L^{− 1} of Fe²⁺, the corrosion scales had similar characteristics and were composed of light grey phases integrated in a darker matrix (Fig. 6a and c). This indicated that there was a chemical difference between these two regions and the corrosion layer must have contained two different phases. The XRD measurements (see Fig. 7) revealed many cementite (Fe₃C) peaks for samples exposed to solutions without Fe²⁺ and with 370 mg L^{− 1} of Fe²⁺. A single FeCO₃ peak was also detected when testing with 0 mg L^{− 1} of Fe²⁺, which is in contrast to the sample exposed to the solution with 370 mg L^{− 1} Fe²⁺, where no FeCO₃ could be detected. Additionally, the top surface scale analysis did not show characteristic FeCO₃ crystals for samples exposed to 0 and 370 mg L^{− 1} Fe²⁺ (see Fig. 6b and d respectively). This might indicate that either the crystals were too small to see with the resolution used in this paper due to short exposure time or there was another corrosion product formed with a non-crystalline structure, which would not be detectible with XRD. Consequently, it can be concluded that the corrosion scale formed in a solution with an initial Fe²⁺ of 0 and 370 mg L^{− 1} consisted of Fe₃C originating from the steel embedded in a corrosion product, which could be FeCO₃ (for 0 mg L^{− 1} of Fe²⁺) or non-crystalline phase. Owing to the higher X-ray penetration depth (17 μm for high diffraction angles) compared to the thicknesses of these scales, it cannot be excluded that some Fe₃C was measured from the substrate of the material. Surface analysis also revealed some cracks formed in the corrosion scales, which are seen in Fig. 6b and d, which are most likely formed during preparation of the samples. When exposing the samples to highly supersaturated solutions, a dense double layered scale was formed (Fig. 6e). It consisted of an upper, homogeneous layer and a bottom, inhomogeneous layer, which was also observed for samples tested in solutions with no initial Fe²⁺ and with 370 mg L^{− 1} of Fe²⁺ (Fig. 6a and c respectively), with light grey regions integrated in the darker matrix. The SEM analysis revealed protrusions of corrosion products with more compact scale beneath (Fig. 6f). The sample exposed to highly supersaturated solution for a prolonged time (Fig. 6g) also showed the formation of a double layer with additional protrusions on the top (Fig. 6h). Owing to extended exposure time, the corrosion scale was approximately twice the thickness compared to the scale formed after exposing for 242 h.

Cross-sectional and top surface morphologies of corrosion scales formed in solutions with initial Fe²⁺ content of a and b 0 mg L^{− 1}, c and d 370 mg L^{− 1}, and e and f 1300 mg L^{− 1} for 242 h; g and h 1300 mg L^{− 1} for 1613 h; and i and j 514 mg L^{− 1} for 4485 h

XRD analysis of corrosion scales formed in solutions with initial Fe²⁺ content of a 0 mg L^{− 1}, b 370 mg L^{− 1}, and c 1300 mg L^{− 1} for 242 h; d 1300 mg L^{− 1} for 1613 h; and e 514 mg L^{− 1} for 4485 h

Formation of a double layered scale was also observed for the sample exposed to the solution initially containing 514 m L^{− 1} of Fe²⁺ (Fig. 6i). The bottom layer was seven times thicker in the solution with 514 mg L^{− 1} (Fig. 6i) than in 1300 mg L^{− 1} of Fe²⁺ (Fig. 6g). Many cracks and pores were also present in the bottom part of the scale, while a compact and thin layer covering the bottom layer was formed on the top (Fig. 6i and j).

XRD results showed that a number of FeCO₃ peaks with high intensities were detected for samples exposed to highly supersaturated solutions (Fig. 7c and d) and electrolytes with 514 mg L^{− 1} of Fe²⁺ (Fig. 7e). This indicated that the volume fraction of this phase was much larger than obtained for the two other scales formed in solutions without iron ions and 370 mg L^{− 1} of Fe²⁺ (Fig. 7a and b respectively). Further, there was a shift in the position of FeCO₃ peaks for samples exposed to highly supersaturated solution (Fig. 7d) and with 514 mg L^{− 1} of Fe²⁺ (Fig. 7e) compared to peaks recorded for samples exposed to solutions without Fe²⁺ (Fig. 7a) and with 370 mg L^{− 1} Fe²⁺ (Fig. 7e shifted towards lower diffraction angles, indicating changes of the state of internal stresses, compared to those detected in Fig. 7a and c). The XRD measurements also revealed the presence of Fe₃C. However, only a single Fe₃C peak with low intensity was detected for samples tested in highly supersaturated solutions (Fig. 7c and d) and 514 mg L^{− 1} (Fig. 7e). Owing to the large thickness of the corrosion scale formed in a solution with 514 mg L^{− 1} of Fe²⁺, compared to the maximum penetration depth of the X-ray beam (17 μm), it can be concluded that the Fe₃C phase was present in the corrosion scale. This, together with the observed morphology of the corrosion scale, indicated that the bottom layer consisted of Fe₃C integrated in a FeCO₃ matrix, while the top layer contained only FeCO₃ phase, as shown in Fig. 6i. This is also applicable for double scales formed in highly supersaturated solutions (Fig. 6e and g) and confirms the results discussed above, i.e. that the Fe₃C measured from solutions with no Fe²⁺ and 370 mg L^{− 1}, indeed, originated from the corrosion scale and not the substrate.

Discussion

Overall analysis of the results showed that the initial concentration of Fe²⁺ and HCO₃ ⁻, together with resulting pH, in the solution had a significant impact on the electrochemical behaviour of steel, corrosion rate and scale formation. It is important to emphasise that it is not the concentration of Fe²⁺ that directly affects the corrosion rate, but the kinetics of scale formation and its morphology are influenced by the initial concentration of Fe²⁺, as shown by the results.

It was shown that the general corrosion rate decreased when increasing the initial concentration of Fe²⁺ in the solution, which is in agreement with formerly published results.^3,13,14,17 Despite the formation of a relatively thick corrosion scale (12 μm), which was formed in solutions with no initial Fe²⁺ and 370 mg L^{− 1}, the scale was not protective, which was reflected in the increasing corrosion rates with the exposure time, even though the saturation level for FeCO₃ was exceeded in both tests. Current results, from the tests performed with initial Fe²⁺ concentration of 0 and 370 mg L^{− 1}, confirm that protectiveness of the formed corrosion scale depends more on its structure and morphology rather than on thickness and the scales formed at room temperature can be even 100 μm thick and still non-protective. ^4,15,26 At room temperature, the process of FeCO₃ precipitation is much slower compared to high temperatures (>60°C), and when the corrosion rate is faster than the scale formation, a porous and non-protective corrosion layer is developed.²⁶ However, based on the SEM analysis, no porosity was seen for scales formed in solutions with an initial Fe²⁺ of 0 and 370 mg L^{− 1}. Han et al. ²² have reported that due to non-uniform formation of a corrosion scale, a galvanic effect between film covered surface (cathode) and uncovered parts of metal surface (anode) enhanced the corrosion rate of steel; this might have also taken place for samples tested in solutions with an initial Fe²⁺ of 0 and 370 mg L^{− 1}, which would explain their increasing corrosion rate with exposure time.

SEM analysis showed that for samples exposed to solutions with an initial Fe²⁺ of 0 and 370 mg L^{− 1}, a single corrosion layer was formed, which consisted of Fe₃C embedded in a corrosion product. XRD analysis, performed for samples exposed to solutions with 0 mg L^{− 1} Fe²⁺, suggested that the scale consisted of Fe₃C embedded in FeCO₃ matrix, which was also reported by Berntsen et al. ²⁷ However, corrosion scale did not have typical FeCO₃ crystals, which might indicate either that the crystals were too small to be seen with the used resolution or that another corrosion product was formed as well. It has been reported^28,29 that one of the corrosion products formed in aqueous CO₂ environment was ferrous hydroxide [Fe(OH)₂], which might not have a crystalline structure and thus could not be detected with XRD. Since both scales have a similar morphology, it is expected that both scales consisted of Fe₃C originating from the steel embedded in a corrosion product of Fe(OH)₂ and/or FeCO₃. Cementite (Fe₃C) is a second predominant phase in the corrosion scale formed in aqueous CO₂ environment. In a ferritic–pearlitic microstructure, cementite is considered as non-corroding part of steel and may become a site for cathodic reactions,^30,31 which lead to microgalvanic coupling between the Fe₃C and ferrite (α-Fe) and consequently increasing corrosion rates. This was confirmed by Mora-Mendoza and Turgoose,³² where it has been shown that the corrosion rate of mild steel in CO₂ solution with pH 5.5 increased with exposure time as a result of unoxidised Fe₃C and its galvanic effect on the corrosion rate. Berntsen et al. ²⁷ have reported that due to preferential dissolution of ferrite (α-Fe), Fe₃C accumulates on the surface and becomes embedded in the FeCO₃ film. Crolet et al. ²⁴ have shown that the main difference between protective and non-protective morphologies of the corrosion scales that formed in aqueous CO₂ environment is the presence of empty cementite (Fe₃C) layer in contact with steel surface in case of non-protective scales and fill of Fe₃C network with FeCO₃ in case of protective corrosion layers. However, despite a clear increase in corrosion rate, which indicated formation of non-protective scales for samples tested in solutions with initial Fe²⁺ concentration of 0 and 370 mg L^{− 1}, no empty Fe₃C layer was observed close to the metal surface with the SEM resolution employed in this paper.

With the generation of the test solution with a high supersaturation of Fe²⁺ by corrosion of steel wool under CO₂ sparging, a large concentration of Fe²⁺ and the equivalent amount of HCO₃ ⁻ were created. After transfer of the test solution to the test vessel, only the corrosion of the test specimens could provide additional Fe²⁺ to the solution. It was shown that such conditions lead to a formation of a protective corrosion scale and very low corrosion rates of 1.5 μm year^{− 1}, in accordance with the practical experience with flexible pipe. Decrease in corrosion rate was also seen in a solution containing initially 514 mg L^{− 1} of Fe²⁺, but the minimum corrosion rate of 25 μm year^{− 1} was first seen after 2324 h.

Results from tests performed with solutions highly supersaturated with Fe²⁺ showed that a double layer scale was formed. It was clear that high initial Fe²⁺ concentration facilitated formation of a dense and protective scale, which was reflected in a decreased corrosion rate and increased (more noble) corrosion potentials. These results are in a good agreement with already published data,²³ where it was shown that by exposing samples to solution initially highly supersaturated with Fe²⁺, thus by providing high SR, a rapid reduction of the corrosion rate was seen due to the formation of protective corrosion layer. It is known that higher supersaturation of Fe²⁺ increases the rate of a formation of FeCO₃ scale,²⁰ which might have also taken place when exposing samples to solutions with Fe²⁺ of 1300 mg L^{− 1}, which would support the statement that high supersaturation of Fe²⁺ is necessary for the formation of protective films.^13,14 The thermodynamic calculations presented in the present paper support this observation where the SR of FeCO₃, which is a driving force of scale formation,²⁰ was much higher in solutions with 1300 mg L^{− 1} (SR = 417) than in solutions with 370 mg L^{− 1} (SR = 6) and 514 mg L^{− 1} of Fe²⁺ (SR = 16).

Based on the detected phases by XRD measurements, it is likely that a bottom layer of corrosion product consisted of Fe₃C integrated in a FeCO₃ matrix and the top layer contains only the FeCO₃ phase. It has been reported^31,33,34 that the formation of FeCO₃ is associated with the steel microstructure. The steel used for this investigation has a microstructure with a very fine dispersion of Fe₃C in ferrite with < 10% by volume of carbide free ferrite. Ferrite grains are relatively small (2-5 μm) and widely distributed. According to Al-Hassan et al. ³⁰ and Sun et al.,³¹ due to more positive potential of Fe₃C, cementite does not corrode and provides effective cathodes; thus, in the present steel, cathodes are abundant even on the most local scales. Since the cementite (Fe₃C) does not dissolve, it will provide a good geometrical marker for the location of the original surface. With the observed two-layer scale, the outer (top) layer consisting only of FeCO₃ with no cementite may have grown at least partly by precipitation from the solution, which could explain decreasing concentration of Fe²⁺ in the electrolyte with time, as it was also reported by Berntsen et al. ²⁷ It cannot, however, be excluded that Fe²⁺ may also be provided by diffusion from the steel side. The present study does not allow a distinction between the two possible sources of Fe²⁺. It is believed that the FeCO₃ present in the bottom layer was primarily formed from the Fe²⁺ due to dissolution of the steel substrate, which explains the presence of Fe₃C originating from the steel in the scale, but certainly Fe²⁺ present in the bulk solution also contributed to the process of scale formation. It is suggested that the interface between the bottom and the top layer presents the original surface of the material. The formation of a double layer has been also reported by Dugstad et al. ³⁵ when exposing steel armours used in flexible pipes to artificial sea water, with a high initial Fe²⁺ concentration of 2000 ppm, at 1 bar CO₂. The corrosion layer consisted of an inner layer, which consisted of FeCO₃ and Fe₃C, together with alloying elements originating from the steel, and an outer layer of FeCO₃ precipitated from the Fe²⁺ in the solution. A similar morphology of the corrosion scale was also reported by Gao et al.,³⁶ where experimental and SEM results showed that by performing tests at 10 bar CO₂ at 75°C and pH 6.5, a formation of double layer corrosion film took place. Similar to current results, the two layers were divided by a straight line, which represented the initial surface of the material. Gao et al. ³⁶ have reported that the layer formed closer to the metal surface had a two-phase structure and a higher resistance compared to the outer homogenous layer. Results from extended exposure to highly supersaturated solutions with Fe²⁺ showed that the scale became even more protective with time. It is in a good agreement with formerly published results,^12,37 where it has been shown that the surface coverage percentage by FeCO₃ increased with exposure time, and at the same time, the corrosion scale became denser and more protective.

Formation of a double layer, with a similar morphology compared to the scale formed in highly supersaturated conditions, was also seen for the sample exposed to the solution with 514 mg L^{− 1} of Fe²⁺. However, the thickness of the bottom part of the scale was seven times larger than the one formed in highly supersaturated solution, which was also reflected in the extended time needed for the corrosion rate to decrease. The XRD measurements revealed a shift in the position of FeCO₃ peaks, similar to the pattern obtained for samples exposed to solutions initially highly supersaturated with Fe²⁺ for a prolonged time. The shift in the FeCO₃ peak position is most likely related to the presence of compressive stresses in the formed scale. It might be that the level of stress, which was high enough to be visible in a peak shift in XRD, could be first seen when the thickness of the corrosion scale was larger due to longer exposure times. It has been suggested by Nešić et al. ²³ that in order to form protective corrosion scales in solutions containing Fe²⁺ below supersaturated conditions, extended exposure times are necessary. Even though the current experiment was running with Fe²⁺ concentration that was much higher than the solubility limit of FeCO₃ (202 mg L^{− 1}), as predicted by thermodynamic calculations, long exposure times of 2324 h were needed to see formation of a protective corrosion layer and corresponding low corrosion rates of 25 μm year^{− 1}.

The results from prolonged exposure times in solutions with an initial Fe²⁺ of 514 mg L^{− 1} of Fe²⁺1300 mg L^{− 1} showed a clear correlation between decreasing concentration of Fe²⁺ and decreasing corrosion rates of tested samples. It was shown that for samples tested in a solution with an initial Fe²⁺ of 514 mg L^{− 1}, a decrease of Fe²⁺ concentration occurred approximately at the same time when the corrosion rate started to decrease, while for samples tested in a solution initially highly supersaturated with Fe²⁺(1300 mg L^{− 1}), a decrease of Fe²⁺ was seen when a corrosion rate of 1.5 μm year^{− 1} was observed. It is important to emphasise that when the solution is highly supersaturated with respect to FeCO₃, the precipitation of FeCO₃ will occur on all available surfaces, including the test vessel and its ancillaries, at the same time contributing to a long-term decrease in Fe²⁺ concentration and, due to removed excess of HCO₃ ⁻ with the general CO₂ gas bubbling, to a long-term decrease of pH towards that predicted by thermodynamics for a solution saturated with FeCO₃. Consequently, observed decreasing concentration of Fe²⁺ in the solution cannot be directly used to estimate the rate of scale formation, but the time when the Fe²⁺ concentration started to decrease, as shown by current results, can be used as an indication when the protective scale starts to be formed, which is reflected in decreasing corrosion rate.

Results from close packed experiments with free water volume to steel surface ratio (V/S) of 0.2 mL cm^{− 2} (Ref. 3) showed that it is not the thickness of the electrolyte above the steel surface but the Fe²⁺ concentration in the solution that influences steel corrosion rates. Despite the high V/S ratio (40 mL cm^{− 2}), materials exposed to highly supersaturated conditions obtained similar corrosion rates as those tested in the close packed conditions with a V/S of 0.2 mL cm^{− 2} (as in the annulus of a flexible pipe). However, the V/S ratio has a significant influence on the time, which is needed to reach supersaturation of Fe²⁺ in the test solution when starting with a non-Fe²⁺ containing electrolyte. With lower V/S ratio, supersaturation with Fe²⁺, together with HCO₃ ⁻, and the corresponding maximum pH are reached faster compared to tests executed with high V/S ratio. It has been reported by Rubin et al. ³ that the maximum pH for V/S = 0.17 mL cm^{− 2} was reached after 48 h, whereas at V/S = 40 mL cm^{− 2}, peak values were first noticed after ∼528 h. Traditional electrochemical laboratory experiments are designed with a large V/S ratio (small working electrode and large volume of electrolyte) and with relatively short test times. This means that the chemistry (Fe²⁺ concentration) of the electrolyte changes slowly during the testing period, and consequently, the behaviour found in this work might never be experienced. It should also be emphasised that in operating conditions, findings presented in this paper will only apply for conditions with no or low replenishment of corrosive media such as in confined conditions, which are relevant for the present case.

Corrosion rates presented in this paper for highly supersaturated conditions with Fe²⁺ correlated very well with former results from close packed cell tests reported in the literature,³ which show that when performing closely packed cell tests, a high pH of 6.2 was obtained within 48 h of exposure. This is negligible compared to the lifetime of a flexible pipe, which is typically 20-30 years. It can therefore be assumed that from the very beginning of the operation of the pipe, steel wires are exposed to highly supersaturated environment with Fe²⁺ during service. Consequently, it is expected that the low corrosion rate found in this investigation for highly supersaturated solutions is applicable to the annulus region for the entire lifetime of the flexible pipe.

It has been reported^38–42 that when exposing materials to corrosive media, their fatigue life is decreased compared to the air conditions. Consequently, the corrosive environment and the resulting corrosion behaviour of steel armours have a significant impact on the fatigue life of steel armours situated in the annulus region and consequently the fatigue performance of the entire flexible pipe. The results presented in this paper indicate that steel armours exposed to simulated annulus conditions (high initial supersaturation) only experience high corrosion rates initially, while after a short time, the corrosion rate decreases to 1.5 μm year^{− 1}, which has been also supported by other authors.^3,43–47 It indicates that the corrosion process of steel armours stops relatively quickly after the water and corrosive gases (CO₂) reach the annulus region. Consequently, when the protective layer is formed and the corrosion rate decreases to almost zero, the corrosion process is expected to have lower impact on the fatigue life of steel armours, which is beneficial for the lifetime of a flexible pipe.

In order to predict fatigue damage of a flexible pipe, corrosion fatigue tests of steel armours are performed on a laboratory scale. A correct simulation of the annulus conditions when performing corrosion fatigue experiments plays an important factor not only in predicting fatigue life of a flexible pipe but also in optimising its design. In the corrosion fatigue tests, the annulus conditions are typically simulated by exposing steel armours to solution with high initial concentration of Fe²⁺. However, the results presented in this paper show that even small changes in the initial Fe²⁺ concentration have a significant impact on the kinetics of scale formation and accordingly how fast a decrease of corrosion rate of steel armours can be observed. Consequently, in order to ensure a close simulation of the annulus conditions when performing corrosion fatigue testing, it is important to expose steel armours to solution with high initial supersaturation of Fe²⁺.

Conclusions

Analysis of Fe²⁺ evolution showed that a decrease in Fe²⁺ and corresponding pH was directly related to a decrease in a corrosion rate of analysed samples. A twice as fast decrease in Fe²⁺ concentration and pH was observed in solutions initially highly supersaturated with Fe²⁺ compared to solutions with Fe²⁺ concentration marginally above the solubility limit of FeCO₃.

Analysis of the Fe²⁺ together with obtained corrosion rates for samples exposed to different initial Fe²⁺ revealed that before protective FeCO₃ can be formed, high supersaturation of Fe²⁺(∼1300 mg L^{− 1}) needs to be built up in the solution.

Electrochemical measurements confirmed that the initial amount of Fe²⁺ in the solution had a significant impact on the corrosion rate of steel materials. Results showed that high initial supersaturation of Fe²⁺ is required for fast formation of protective layers and decrease of corrosion rates close to 1.5 μm year^{− 1}.

Metallographic and diffraction analysis showed that the morphology of the scale depended on the initial exposure environment and exposure time. Samples exposed to the solution without Fe²⁺ or in supersaturated solution for a short exposure time formed scales consisting of Fe₃C embedded in FeCO₃, while under highly supersaturated conditions, the scale showed a double layer with an inner layer of Fe₃C embedded in FeCO₃ and an outer FeCO₃ layer.

The results from experiments performed with simulated annulus conditions (high supersaturation of Fe²⁺) showed fast decrease of corrosion rate to 1.5 μm year^{− 1}, indicating that the corrosion process stops relatively quickly after the water and CO₂ gases reach the annulus region. The low corrosion rate of steel armours is believed to have a beneficial impact on the fatigue life of a flexible pipe.

The results showed that small changes in initial Fe²⁺ have a large impact on the corrosion rates of steel armours. Consequently, in order to correctly simulate annulus conditions, when performing corrosion fatigue testing, high supersaturation of Fe²⁺ must be provided from the beginning of the test.

Footnotes

Acknowledgement

The authors would like to acknowledge InnovationsFonden in Denmark for providing financial support in the project.

References

Rubin

and Gudme

: ‘Qualification of steel wire for flexible pipes’ NACE International, Houston, TX, March 2006, Paper 06149.

Rubin

and Gudme

: ‘Test method for corrosion fatigue testing of cold rolled steel wire in sour and sweet environment based on deflection controlled four point bending’ NACE International, Houston, TX, March 2009, Paper 09103.

Rubin

Overby

Nielsen

T. S.

Haahr

and Gudme

: ‘Corrosion rates of carbon steel in confined environments’ NACE International, Houston, TX, 2012, Paper 01495.

Nešić

: ‘Key issues related to modelling of internal corrosion of oil and gas pipelines—a review’ Corros. Sci., 2007 49 4308–4338. doi: 10.1016/j.corsci.2007.06.006

Nordsveen

Nešić

Nyborg

and Stangeland

: ‘A mechanistic model for carbon dioxide corrosion of mild steel in the presence of protective iron carbonate films—part 1: theory and verification’ Corrosion 2003 59 443–456. doi: 10.5006/1.3277576

Nešić

Nyborg

Stangeland

and Nordsveen

: ‘Mechanistic model for CO₂ corrosion with protective iron carbonate films’ NACE International, Houston, TX, March 2001, Paper 01040.

Nešić

: ‘An electrochemical model for prediction of corrosion of mild steel in aqueous carbon dioxide solutions’ Corrosion 1996 52 280–294. doi: 10.5006/1.3293640

Nešić

and Lee

K. -L. J.

: ‘A mechanistic model for carbon dioxide corrosion of mild steel in the presence of protective iron carbonate films—part 3: film growth model’ Corrosion 2003 59 616–628. doi: 10.5006/1.3277592

Nešić

Nordsveen

and Nyborg

: ‘A mechanistic model for carbon dioxide corrosion of mild steel in the presence of protective iron carbonate films—part 2: a numerical experiment’ Corrosion 2003 59 489–497. doi: 10.5006/1.3277579

10.

Sun

Nešić

and Woollam

R. C.

: ‘The effect of temperature and ionic strength on iron carbonate (FeCO₃) solubility limit’ Corros. Sci., 2009 51 1273–1276. doi: 10.1016/j.corsci.2009.03.009

11.

Mullin

J. W.

: ‘Crystallization’3rd edn242; 1993

Oxford

Oxford Univ. Press.

12.

Sun

and Nešić

: ‘Kinetics of corrosion layer formation: part 1—iron carbonate layers in carbon dioxide corrosion’ Corrosion 2008 64 334–346. doi: 10.5006/1.3278477

13.

Videm

and Dugstad

: ‘Corrosion of carbon steel in an aqueous carbon dioxide environment. II: film formation’ Mater. Perform., 1989 28 46–50.

14.

Videm

and Dugstad

: ‘Corrosion of carbon steel in an aqueous carbon dioxide environment. I: solution effects’ Mater. Perform., 1989 28 63–67.

15.

Dugstad

: ‘The importance of FeCO₃ supersaturation on the CO₂ corrosion of carbon steels’ NACE International, Houston, TX, March 1992, Paper 14.

16.

Nyborg

and Dugstad

: ‘Mesa corrosion attack in carbon steel and 0.5% chromium steel’ NACE International, Houston, TX, March 1998, Paper 29.

17.

Kinsella

Tan

Y. J.

and Bailey

: ‘Electrochemical impedance spectroscopy and surface characterization techniques to study carbon dioxide corrosion product scales’ Corrosion 1998 54 835–842. doi: 10.5006/1.3284803

18.

Nešić

Solvi

G. T.

and Enerhaug

: ‘Comparison of the rotating cylinder and pipe flow tests for flow-sensitive carbon dioxide corrosion’ Corrosion 1995 51 773–787. doi: 10.5006/1.3293555

19.

Ruzic

Veidt

and Nešić

: ‘Protective iron carbonate films—part 3: simultaneous chemo-mechanical removal in single-phase aqueous flow’ Corrosion 2007 63 758–769. doi: 10.5006/1.3278425

20.

Dugstad

: ‘Fundamental aspects of CO₂ metal loss corrosion—part 1: mechanism’ NACE International, Houston, TX’; 2006.

21.

Chokshi

Sun

and Nešić

: ‘Iron carbonate scale growth and the effect of inhibition in CO₂ corrosion of mild steel’ NACE International, Houston, TX, 2005, Paper 05285.

22.

Han

Brown

B. N.

and Nešić

: ‘Investigation of the galvanic mechanism for localized carbon dioxide corrosion propagation using the artificial pit technique’ NACE International, Houston, TX, 2007, Paper 07323.

23.

Nešić

Lee

and Ruzic

: ‘A mechanistic model of iron carbonate film growth and the effect on CO₂ corrosion of mild steel’. NACE International, Houston, TX; 2002.

24.

Crolet

J. L.

Thevenot

and Nešić

: ‘Role of conductive corrosion products in the protectiveness of corrosion layers’ Corrosion 1998 54 194–203. doi: 10.5006/1.3284844

25.

Stephen Tait

: ‘An introduction to electrochemical corrosion testing for practicing engineers & scientists’; 1994

Racine, WI

PairODocs Professionals.

26.

Dugstad

: ‘Mechanism of protective film formation during CO₂ corrosion of carbon steel’ NACE International, Houston, TX, 1998, Paper 31.

27.

Berntsen

Seiersten

and Hemmingsen

: ‘Effect of FeCO₃ supersaturation and carbide exposure on the CO₂ corrosion rate of carbon steel’ Corrosion 2013 69 601–613. doi: 10.5006/0553

28.

Han

Yang

Nesic

and Brown

B. N.

: ‘Roles of passivation and galvanic effects in localized CO₂ corrosion of mild steel’. NACE International, Houston, TX, 2008, Paper 08332.

29.

Pfennig

Wiegand

Wolf

and Bork

C. -P.

: ‘Corrosion and corrosion fatigue of AISI 420C (X46Cr13) at 60 °C in CO₂-saturated artificial geothermal brine’ Corros. Sci., 2013 68 134–143. doi: 10.1016/j.corsci.2012.11.005

30.

Al-Hassan

Mishra

Olson

D. L.

and Salama

M. M.

: ‘Effect of microstructure on corrosion of steels in aqueous solutions containing carbon dioxide’ Corrosion 1998 54 480–491. doi: 10.5006/1.3284876

31.

Sun

J. B.

Zhang

G. A.

Liu

and Lu

M. X.

: ‘The formation mechanism of corrosion scale and electrochemical characteristic of low alloy steel in carbon dioxide-saturated solution’ Corros. Sci., 2012 57 131–138. doi: 10.1016/j.corsci.2011.12.025

32.

Mora-Mendoza

J. L.

and Turgoose

: ‘Fe₃C influence on the corrosion rate of mild steel in aqueous CO₂ systems under turbulent flow conditions’ Corros. Sci., 2002 44 1223–1246. doi: 10.1016/S0010-938X(01)00141-X

33.

Ueda

and Ikeda

: ‘Effect of microstructure and Cr content in steel on CO₂ corrosion’ NACE International, Houston, TX, March 1996, Paper 13.

34.

López

D. A.

Pérez

and Simison

S. N.

: ‘The influence of microstructure and chemical composition of carbon and low alloy steels in CO₂ corrosion. A state-of-the-art appraisal’ Mater. Des., 2003 24 561–575. doi: 10.1016/S0261-3069(03)00158-4

35.

Dugstad

Børvik

Palencsar

and Eikrem

P. A.

: ‘Corrosion testing of steel armour wires in flexible pipes—a parametric study’ NACE International, Houston, TX, March 2015, Paper 5829.

36.

Gao

Pang

and Gao

: ‘The growth mechanism of CO₂ corrosion product films’ Corros. Sci., 2011 53 557–568. doi: 10.1016/j.corsci.2010.09.060

37.

Zhang

Pang

and Gao

: ‘Discussion of the CO₂ corrosion mechanism between low partial pressure and supercritical condition’ Corros. Sci., 2012 59 186–197. doi: 10.1016/j.corsci.2012.03.006

38.

Félix-Henry

: ‘Prevention and monitoring of fatigue-corrosion of flexible risers’ steel reinforcements’ ‘Proceedings of the 26^th International Conference on Offshore Mechanics and Arctic Engineering, San Diego, CA, June 2007, Paper 29186’

39.

Akid

: ‘Corrosion fatigue’ Shreir's Corros., 2010 928–953. doi: 10.1016/B978-044452787-5.00038-X

40.

Wang

Y. Z.

: ‘Corrosion fatigue’, in ’Uhlig's corrosion handbook’3rd edn195–202; 2011

New York

John Wiley & Sons.

41.

Wang

R. A. Y.

and Akid

: ‘Role of nonmetallic inclusions in fatigue, pitting, and corrosion fatigue’ Corrosion 1996 52 92–102. doi: 10.5006/1.3292108

42.

Congleton

Craig

I. H.

Olieh

R. A.

and Parkins

R. N.

: ‘Some electrochemical and microstructural aspects of corrosion fatigue’367–389; 1983

Philadelphia, PA

American Society for Testing and Materials.

43.

Ropital

Condat-Taravel

Saas

J. N.

and Duret

: ‘Methodology to study the general corrosion of steel armours in simulated conditions of flexible pipe annulus influence of confinement and evaluation of the pH’ The European Corrosion Congress, Moscow, Russia; September 2000.

44.

Taravel-Condat

and Desamais

: ‘Qualification of high strength carbon steel wires for use in specific annulus environment of flexible pipes containing CO₂ and H₂S’. Proc. 25^th International Conference on Offshore Mechanics and Arctic Engineering, Hamburg, Germany, June 2006, Paper 92394.

45.

Remita

Ropital

and Kittel

: ‘Experimental and theoretical investigation of the uniform corrosion in the annulus of offshore flexible pipelines’ NACE International, Houston, TX, March 2008, Paper 08538.

46.

Remita

Tribollet

Sutter

Ropital

Longaygue

Kittel

Taravel-Condat

and Desamais

: ‘A kinetic model of CO₂ corrosion in the confined environment of flexible pipe annulus’, ‘ The European Corrosion Congress, Freiburg, Germany’; September 2007.

47.

Clements

: ‘Corrosion assessment prediction for a confined flexible pipe annulus’ The European Corrosion Congress, Edinburgh, Scotland; September 2008.

Effect of Fe ion concentration on corrosion of carbon steel in CO 2 environment

Abstract

Keywords

Introduction

Experimental

Experimental set-up for corrosion testing

Material and specimen preparation

Electrolyte solution

Electrochemical measurements

Scale analysis

Thermodynamic calculations

Results

Thermodynamic calculations

Electrochemical testing and electrolyte analysis

Evolution of Fe2+ and pH

Evolution of corrosion potential

Corrosion rates

SEM and XRD analysis

Discussion

Conclusions

Footnotes

Acknowledgement

References

Evolution of Fe²⁺ and pH