Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Corrosion Science 85 (2014) 380–393

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

On the theory of CO2 corrosion reactions – Investigating their


interrelation with the corrosion products and API-X100 steel
microstructure
Faysal Fayez Eliyan ⇑, Akram Alfantazi
Corrosion Group, Department of Materials Engineering, The University of British Columbia, Vancouver, BC V6T 1Z4, Canada

a r t i c l e i n f o a b s t r a c t

Article history: This electrochemical research investigates the interrelation between the corrosion reactions, the growth
Received 28 November 2013 and characteristics of the corrosion products, and the microstructures of simulated Heat-Affected Zones
Accepted 29 April 2014 (HAZs) from API-X100 steel. In saline CO2-saturated solutions, the corrosion products suppress more the
Available online 9 May 2014
anodic dissolution than the cathodic reactions. The corrosion products of ferritic microstructures are
thicker and of higher cathodic currents than those of acicular-ferritic and martensitic microstructures,
Keywords: which are intrinsically less reactive. The significances of chloride, iron oxides, and Fe3C with the growth
A. Carbon steel
and reactivity of the corrosion products were explored. Schematics depicting the open-circuit reactions
B. Polarization
B. EIS
were made from electrochemical impedance spectroscopy measurements.
C. Passive films Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction investigations on its corrosion resistance improves the predictabil-


ity of the corrosion behavior of the new steels in the field and opti-
The internal corrosion of welded pipeline steel segments is one of mizes the welding methods so that they result in minimal
the common problems that could critically undermine the opera- metallurgical corrosion causes.
tional safety and integrity of oil pipelines [1–3]. At a welding joint, This paper correlates the stability, growth, and characteristics of
it is the finite metallurgical variety that exacerbates and localizes the corrosion products to the CO2 corrosion reactions and HAZ
the corrosion attacks. It is of significance that could be as high as that microstructures. Depending on the microstructure and the envi-
of the physicochemical and hydrodynamic effects acting on the rest ronmental conditions – which relate to the high temperature and
of the pipeline [4–8]. The ‘‘metallurgical variety’’ refers to a multiple chloride concentration in this paper – the corrosion products
of microstructures – across a narrow region over the pipe circumfer- become natural corrosion-inhibiting barriers of changing charac-
ence – of types, distributions, and proportions that depend on the teristics with time. The corrosion behavior of the HAZ is different
thermal fluxes associated with a pipeline welding method [9–11]. from the adjacent base steel, and this raises two issues. First, the
Corrosion-wise, the vulnerability of such a metallurgical system interrelation between the corrosion behavior, species transport,
arises from the micro-galvanic conductance between many micro- and the characteristics of the corrosion products need to be deter-
structures of different corrosion reactivities [12,13]. The emphasis mined. Second, the characteristics of the corrosion products need
in this study is on the heat-affected zone (HAZ), a region that lies to be linked to the HAZ microstructures. With time, the corrosion
between the weldment fusion and the parent base steel, which has products on the HAZ grow different from those of the base steel.
distinct microstructures that depend mainly on the cooling rate – Considering the thickness or compactness, for instance; their vari-
which is a function of the welding method and pipe thickness ation over the two regions affects the underlying reactions differ-
[14,15]. The subject base steel is the high-strength API-X100 of HAZs ently and generates hydrodynamic vortices that add to the
simulated by GleebleÓ thermal simulation cycles. API-X100 is a inhomogeneity of the reactions [16–19]. This exacerbates the seg-
new-generation steel of special alloying composition that makes it regation of the galvanic microcells that are already established
of high-pressure reliability for a number of future oil transmission between the microstructures of the HAZ and base steel.
projects in North America. Carrying out electrochemical In literature, the significance of the HAZ microstructure with CO2
corrosion reactions received little interest. Mostly, the test HAZs
⇑ Corresponding author. Tel.: +1 778 997 4878. were not properly simulated, compromising their microstructural
E-mail address: faysal09@interchange.ubc.ca (F.F. Eliyan).
specificity and electrochemical significance. The open-circuit

http://dx.doi.org/10.1016/j.corsci.2014.04.055
0010-938X/Ó 2014 Elsevier Ltd. All rights reserved.
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 381

potential (OCP) tests [20,21] were seldom used alone to investigate pearlite and bainite. The microstructure of 30 °C/s HAZ, as shown
CO2 corrosion reactions, and the results did not usually lead to clear in Fig. 1c, is mainly acicular-ferritic, and the microstructure of
outcomes. In some studies, the OCPs were monitored during long- 60 °C/s HAZ, as shown in Fig. 1d, consists mainly of martensite–
term immersion tests or in combination with other measurements, retained austenite (M/A), martensite, and grain-boundary ferrite.
such as the electrochemical impedance spectroscopy (EIS) [22–25].
In this study, we carried out potentiodynamic scans across three 2.3. Test solutions
stages of OCP monitoring, during which the growth, stability, mor-
phology, and composition of the corrosion products are linked to the The test solutions were synthesized from double-distilled, deion-
corrosion reactions. The significance of chloride [26–28] is re- ized water, continuously purged with a 1-bar carbon dioxide (CO2)
examined. Independent EIS measurements were carried out during flow with a high purity (99.99%). They had chloride concentrations
free immersion to analyze the growth of the corrosion products of 0.17, 0.85, and 1.38 M. The chloride was introduced in the form
which, along with the interfacial interactions, are depicted in of the analytical grade, Fisher-procured sodium chloride (NaCl).
schematics. The test temperatures were 90 ± 1 °C. The pH, which was around
4.5, was measured by an ALpHAÒ series epoxy electrode, equipped
2. Experimental details with annular ceramic liquid junction and manufactured by OMEGA.
The test solutions simulate the upstream flows that have high water
2.1. Corrosion test setup cuts and changeful salinity [29] at high temperatures.

The tests were carried out in a one-liter, three-electrode, multi- 2.4. Electrochemical and surface-analysis tests
port jacketed cell. The working electrodes were machined from an
API-X100 pipeline shell into flat disks. They were 5 mm thick and 1 Before starting the experiments, a cathodic conditioning was
cm2 in surface area, fitted into a sample holder made of Tefzel, and applied at 2 V vs. SCE, for 800 s right after immersing a test sam-
sealed by a Kalrez washer to maintain a standard exposure to the ple, by which the air-formed oxides destabilize and dissolve under
test solutions. The potentials were measured against the Saturated hydrogen generation. The experiments were repeated three times
Calomel Electrode (SCE), of +0.240 V vs. SHE, which was in a salt to ensure reproducibility. First, consecutive OCP and potentiody-
bridge isolated at room temperatures. The counter electrode was namic polarization tests were carried out. They started with mon-
made of a graphite rod. The jacketed chamber of the test cell was itoring the OCPs for 1 ks, denoted as OCP1, followed by a
connected with tubes to a circulator, pumping a 95 °C water flow potentiodynamic polarization, shifted from the OCPs to be scanned
to the chamber to heat the solutions. The experiments were carried from 1 to 0.4 V vs. SCE, at 0.5 mV/s. The OCPs were monitored
out by a Versastat 4 potentiostat, synchronized to a VersaStudio again for 6 ks, denoted as OCP2, and then the same polarization
software program to set and control the experiments and analyze was scanned before the OCPs were finally monitored for 8 ks,
the results. denoted as OCP3. The morphology and composition of the corro-
sion products of these experiments were investigated by scanning
2.2. Test material electron microscopy (SEM), energy dispersive X-ray spectrometry
(EDS), and X-ray photoelectron spectroscopy (XPS). This was for
The specimens, prior to immersion in the test solutions, were samples rinsed with distilled water and ethanol and dried and
wet ground in sequence of 120, 320, and 600 grit emery papers, stored in a dessicator until surface analysis.
ultrasonically degreased in ethyl alcohol, rinsed in distilled water, Independent electrochemical impedance spectroscopy (EIS)
and dried in a hot air stream. The chemical composition of API- measurements were carried out after 5, 10, and 15 ks of free immer-
X100, presented in Table 1, was analyzed by the inductive coupled sion, with a range of frequency of 10,000–0.05 Hz (and in reverse, to
plasma (ICP). A Nikon EPIPHOT 300 optical microscope was used to ensure the electrochemical stability during some measurements),
produce the optical micrographs of the as-received and HAZ micro- and a sampling rate of 10 points per decade.
structures. Prior to the microstructural analysis, selected samples
were wet-ground up to 1200 grit finish and polished with 6 and
1 lm diamond suspensions. They then were immersed in a nital 3. Results and discussion
etchant (2 mL of 70% nitric acid and 98 mL of anhydrous, dena-
tured ethyl alcohol), treated with alcohol swapping, and dried in 3.1. Suppression of anodic dissolution by the corrosion products
an air stream. Fig. 1a shows the as-received microstructure. It con-
sists of a mixture of ferrite, pearlite, and bainite. A GleebleÓ 3500 Consecutive OCP and potentiodynamic polarization tests were
thermal simulation machine simulated three types of near-fusion carried out to study the interrelation between the corrosion reac-
HAZs. The simulations were made by heating the samples at tions, the growth and properties of the corrosion products, and
100 °C/s to a peak temperature of 950 °C, held for 0.5 s and fol- the microstructure. The open-circuit potentials were first moni-
lowed by controlled cooling at 10, 30, and 60 °C/s, which are repre- tored for 1 ks (OCP1) and then swept to 1 V vs. SCE, from which
sented in Fig. 2. The cooling rates simulate some rates of the heat the potentials were scanned at 0.5 mV/s to 0.4 V vs. SCE. After-
dissipation from a welded pipeline segment – depending on its wards, the OCPs were monitored for 6 ks (OCP2) before a second
thickness and the welding method – to have HAZs of different same polarization was applied. Finally, the OCPs were monitored
microstructures. The microstructure of 10 °C/s HAZ, as shown in for 8 ks (OCP3). In 0.85 M chloride solutions, of profiles shown in
Fig. 1b, consists of large, equiaxed ferrite and some dispersed Fig. 3, the first open-circuit potentials (OCP1) of 10 °C/s HAZ were
the lowest with 0.77 V vs. SCE, and they were 0.76, 0.73, and
0.70 V vs. SCE for the as-received, 30 and 60 °C/s HAZs, respec-
Table 1 tively. The low OCP1 and OCP2 of 10 °C/s HAZ and base steel can
Chemical composition of the test API-X100 steel.
be attributed to two different phenomena. They either can be
Composition (wt%) C.E. attributed to the high dissolution rates of their ferritic constituents
C Mn Mo Ni Al Cu Ti Nb Cr V or to the suppression of the cathodic reactions by thick corrosion
products. The corrosion products form with a tendency propor-
0.1 1.66 0.19 0.13 0.02 0.25 0.02 0.043 0.016 0.003 0.45
tional with the concentrations of the active adsorbents that
382 F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393

Fig. 1. Optical micrographs of (a) the as-received API-X100 steel, (b) 10 °C/s HAZ, (c) 30 °C/s HAZ and (d) 60 °C/s HAZ microstructures. The same samples of these
microstructures were tested in high-temperature bicarbonate solutions [31].

Fig. 2. GleebleÓ thermal simulation cycles set to simulate the HAZs by heating at Fig. 3. Profiles of the consecutive tests of open-circuit potentials monitoring and
100 °C/s to a peak temperature of 950 °C, held for 0.5 s, followed by controlled potentiodynamic polarization of the as-received steel and the HAZs in 0.85 M
cooling at 10, 30, and 60 °C/s. The same thermal cycles were considered for chloride solutions.
simulated HAZs tested in high-temperature bicarbonate solutions [31].

generate from dissolution. And microstructure-wise, after the vig- of the corrosion products – different from what was indicated from
orous dissolution of the large ferrite grains, the corrosion products longer OCP tests carried out by Moiseeva and Kuksina [20].
grow thicker and become anchored by the stable, conductive Fe3C After OCP1, when polarized for the first time, the initially pre-
[30–38]. The corrosion products afterwards inhibit the corrosion cipitated scales dissolve, possibly substantially, to become thinner
reactions which are reported in many studies as being under the and/or porous during the cathodic part of the polarization. It is
cathodic control [6]. Therefore, this might imply that the second from 1 to nearly 0.7 V vs. SCE during which the cathodic cur-
phenomenon relates to the inhibition of the cathodic reactions to rents account partially for the reduction of H+ and H2CO3, as will
result in low OCPs. At the end of the experiments, all samples were be explained in Section 3.2. During the anodic part, up to 0.4 V
covered by intact, gray–black scales of the type previously reported vs. SCE, the surfaces get covered again, possibly by net thicker
in a number of studies [26,39,40]. The OCP3 regimes did not scales. The OCPs during OCP2 were higher than those of OCP1 for
indicate critical transformations in the morphology or composition all samples, as represented in Fig. 4, and they showed a slight
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 383

increase over time. The increase relates to the degree of dominance


of the anodic reactions (sustained through the porous products) vs.
that of the cathodic reactions (sustained more likely onto the sur-
face of the products) – in relation to interrelated factors that result
in the two phenomena explained above.
The scenario of the second phenomenon – the inhibition of the
cathodic reactions by the corrosion products – does not fully
encompass the consequence of water adsorption, dissolution, the
formation and thickening of the corrosion products, the inhibition
of dissolution and, finally, the inhibition of the controlling cathodic
reactions, which can be sustained through the porous corrosion
products [41], or onto them [42,43] in mechanisms different from
those of the film-free surfaces [44]. Therefore, the second polariza-
tion and OCP3 monitoring were carried out to determine the dom-
inant reactions during OCP2, and whether it was the anodic or
cathodic reactions that the corrosion products mainly suppress.
For all samples, the values of OCP3 were higher than those of
OCP1 and OCP2, and the OCP3 of 10 °C/s HAZ was the highest, as
presented in Fig. 4. This indicates that the corrosion products dur-
ing OCP3 were thicker than those of OCP2, as shown in Fig. 5, and
that 10 °C/s HAZ developed the thickest corrosion products during
OCP3. The thickness of the corrosion products of 10 °C/s HAZ
increased from nearly 10, 30 to 70 lm at the end of OCP1, OCP2,
and OCP3, respectively in 0.85 M solutions. Therefore, the consec-
utive experiments indicate that the corrosion products suppress
mainly the anodic dissolution, rather than the cathodic reactions,
and during OCP2 it was the high dissolution rates of 10 °C/s HAZ,
not its suppressed cathodic reactions that made it of lowest OCPs
during OCP2.

3.2. The relationship between the corrosion products and cathodic


reduction

Fig. 6 presents the first and second polarizations in 0.85 M solu-


tions. During the second polarizations, the cathodic currents were
higher and so were the corrosion potentials. The factors, physico-
chemical and/or electrochemical, which these cathodic trends
can be attributed to, are many and interlinked. The corrosion

Fig. 5. SEM micrographs of the cross sections of the corrosion products of 10 °C/s
HAZ by the end of (a) OCP1, (b) OCP2, and (c) OCP3 in 0.85 M chloride solutions.

products, right before the second polarization, are thicker than


those polarized at the first time, i.e. those during OCP1. They
became thicker by further dissolution that made Fe3C, a conductive
phase, have higher proportions in their growing morphology. The
corrosion products are typically comprised of FeCO3 [45–47] as
well as the iron oxides Fe2O3 and Fe3O4[48]. Cathodic conduc-
tion-wise, FeCO3 is nonconductive [49,50], and the extents of
Fig. 4. The values of OCP1, OCP2, and OCP3, with error bars corresponding to the
average of values of three tests (for reproducibility), for the as-received steel and Fe2O3 and Fe3O4 conduction are difficult to estimate, in part
the HAZs in 0.85 M chloride solutions. because their mechanisms of formation are not yet fully
384 F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393

reactions were higher is therefore a possibility – due to the same


reasons mentioned above – to be partially responsible besides
the suppressed dissolution in making the OCPs higher.

3.3. Significance of microstructure and chloride concentration

Similar to previous findings [31,32,60], for the same microstruc-


tures tested in this study, 30 and 60 °C/s HAZs had lower corrosion
rates and cathodic currents than those of the base steel and 10 °C/s
HAZ, as revealed from the polarization profiles in Fig. 6. Moreover,
they showed evidence of thin corrosion products, as they had the
highest OCPs, as presented in Figs. 3 and 4. The acicular ferrite in
30 °C/s HAZ and martensite–retained austenite in 60 °C/s HAZ
were not as reactive as the equiaxed ferrite in the base steel and
10 °C/s HAZ to be of high dissolution rates. Moreover, 60 °C/s
HAZ was particularly of low cathodic activity to enhance dissolu-
tion under galvanic coupling. The microstructures of the 30 and
60 °C/s HAZs therefore seemed to have a more homogenous distri-
bution of the anodic and cathodic regions of which the net currents
were low, as were the ionic fluxes of the electroactive species. This
might not necessarily have changed the mechanisms of the corro-
sion reactions of these rapidly-cooled microstructures of dense
lattice defects – to which the formation of thin, unstable passive
Fig. 6. Profiles of the first and second potentiodynamic polarization for the as- films was attributed [61–63]. The effect of the applied potentials
received steel and the HAZs in 0.85 M chloride solutions.
on the OCP trends was independent of the chloride concentration
and microstructure. The average values of OCP1 for each of the
microstructures were generally higher with the chloride concen-
established in literature. The reported mechanisms describe their tration, as presented in Fig. 7. The dissolution process might have
formation as electrochemical and chemical oxidations of ferrite been catalyzed in proportion with the chloride concentration to
and FeCO3, in many paths. In addition, FeCO3 is a non-stable phase develop thicker corrosion products, during which the chloride
(to quantify) when the sample is extracted from the test solution simultaneously causes pitting. This, as well as the physicochemical
and exposed to air (conventionally on which XPS and XRD tests interrelation between the chloride concentration, CO2 hydration,
are carried out [51–53]– except for a few recent in situ studies pH, and the concentrations of H2CO3 and HCO 3 will be subject
[17,54]). Therefore, it would be unattainable to infer accurate pro- for investigation in the future.
portions of Fe2O3 and Fe3O4, whose cathodic significance are
sought for a sample immersed in a test solution in our case. But
as for Fe3C, the proportions of these oxides increase by the applied 3.4. Morphology and composition of the corrosion products
potentials during the first polarization and OCP2.
The semi-conductive significance of Fe2O3 with the corrosion The top surface and composition of the corrosion products of
reactions received no interest, except in a few peripheral studies the consecutive OCP and polarization tests were analyzed by
[55,56]. Correlating its electronic properties to the mechanisms
or rates of H+ reduction is limited by the kinetic interference of
Fe2O3 dissolution when polarized at low potentials. However,
Fe3C is widely considered an active phase onto which the cathodic
reactions, semi-galvanically, sustain in low-pH, CO2-saturated
solutions [57,58]. In a valuable study by Staicopolus, he synthe-
sized Fe3C by a special procedure and, with early constant-current
and constant-potential techniques, he showed that the rate of the
cathodic reduction on Fe3C was higher than that on iron in chloride
and acidified sulfate solutions [59]. At low currents, he showed
that Fe3C exhibited a high affinity to H+ reduction, but at higher
currents it decomposed to ferrite and hydrocarbons – while on
which hydrogen generation simultaneously took place.
As mentioned in the beginning of this section, the higher catho-
dic currents during the second polarization can be ascribed to
interlinked, comparably-significant factors. First, they can be
ascribed to the dissolution of the corrosion products, either by
the direct reduction or under the disruptive influence of hydrogen
evolution. Second, the higher amounts of the conductive iron
oxides and Fe3C could play surface/volume roles in increasing the
conduction for the cathodic reactions to become in turn of higher
currents. Thirdly, the thicker corrosion products could make the
cathodic reduction increasingly under mass-transfer control,
during which the hydrogen generation might occur more onto Fig. 7. Values of OCP1 with error bars corresponding to the average of values taken
the covered substrate, through the porous corrosion products, to from three tests (for reproducibility), for the as-received steel and the HAZs in 0.17,
be of higher rates. During OCP3, that the rates of the cathodic 0.85, and 1.38 M chloride solutions.
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 385

Fig. 8. SEM micrographs of the corrosion products that formed at the end of the consecutive OCP and polarization tests, for the as-received steel in (a) 0.17 M and (b) 0.85 M,
and for 10 °C/s HAZ in (c) 0.85 M and (d) 1.38 M, and for 30 °C/s HAZ in (e) 0.85 M, and for 60 °C/s HAZ in (f) 0.85 M chloride solutions.

Fig. 9. SEM micrograph and EDS elemental mapping of the cross section of the corrosion products of 10 °C/s HAZ in 0.17 M solution.
386 F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393

SEM, EDS, and XPS. In literature, to be noted, the significance of


chloride with the corrosion products did not receive as much inter-
est as temperature [53,64,65], CO2 pressure [66], and pH [67].
Chloride exacerbates the underlying dissolution and acidifies the
pores of the corrosion products – in a process that tends to result
in thicker corrosion products of greater proportions of carbonate-
and oxide-based products.
In Fig. 8, selected top-surface SEM micrographs of the corrosion
products are presented. For the base steel in 0.17 M chloride solu-
tions, the surface consisted, as shown in Fig. 8a, of round, dispersed
grains of 5 lm wide. They were not as closely compact or sharp-
edged as those of corrosion products formed during free immer-
sion in similar solutions [68,69]. The surface is similar to another
formed on API-X100 after a 0.5 mV/s polarization in higher-pH,
CO2-saturated solutions at 50 °C [70]. In 0.85 M solutions, as
shown in Fig. 8b, another thin, cracked layer seemed to have
formed onto an inner layer that is similar to that in 0.17 M solu-
tions. The chloride ions in higher concentrations enhanced the
underlying dissolution and the activity of the porous corrosion
products to seemingly result in a second layer of different mor-
phology, and possibly of different composition. In Fig. 8c, the mor-
phology reflects, as do the cross sections in Fig. 5, the
electrochemical behavior of 10 °C/s HAZ as dissolving at high rates
in the beginning and then covered with thick corrosion products
after the second polarization. Whether the partial disbondment
of the upper thick layer over a highly compact one is caused by
the second polarization or upon removing the sample from the
solution is unclear. In the more concentrated 1.38 M solution, the
corrosion products seemed more imbricated, as shown in Fig. 8d,
under a synergistic influence of chloride and polarization – which
is difficult to link to their morphology. The corrosion products of
30 °C/s HAZ in 0.85 M solution, shown in Fig. 8e, showed few
cracks across a seemingly outer, thin adherent layer of the type
previously reported for a quenched and tempered steel polarized
in 3% NaCl solutions at 25 °C [71]. For 60 °C/s HAZ, the corrosion
products were of compact grains similar to those of the as-received
sample. They were, as shown in Fig. 8f, relatively larger and, inter-
estingly, there appears to be an incompletely-formed outer layer.
An elemental analysis was performed by EDS on the top-surface
of the corrosion products. Those of 10 °C/s HAZ had the highest
iron concentration and had appreciable amounts of Mn, Mo, Cu,
and Ni of up to 1.5, 2, 3, and 2 wt%, respectively. For 30 °C/s HAZ,
relatively high 4 wt% Ni and 5 wt% Cu were detected.
A cross-sectional mapping of the corrosion products of 10 °C/s
HAZ in 0.17 M solutions is presented in Fig. 9. The thickness was
nearly 10 lm and at the upper part the corrosion products
appeared as uniform, compact, large platelets. Below there was a
degree of porosity and the dark areas correspond to high carbon
concentrations adjacent to the substrate. They represent undis-
solved alloy-carbides – phases previously reported in CO2 corrosion Fig. 10. SEM micrograph and EDS elemental mapping of the cross section of the
corrosion products of 30 °C/s HAZ in 0.17 M solution.
studies [72,73]. The concentration of iron was higher in the upper
areas so that it was associated with oxygen (which was homoge-
nously distributed) in the form of iron oxides. The corrosion prod-
high-resolution curve by XPSPEAK41Ó. Peak I appeared at 710.67
ucts of 30 °C/s HAZ, as shown in Fig. 10, were thinner. They
and Peak II appeared at 723.89, of shifts similar to those previously
appeared as small, compact grains over the substrate that seemed
reported [77,78]. They correspond, as do those by Heuer, Stubbins
it did not corrode evenly. The relative concentrations of iron and
and Wang et al., to a-Fe2O3[45,79], but whether it formed during
manganese, compared with those of 10 °C/s HAZ, were less.
the experiments and/or part of it resulted from Fe3O4 oxidation
The XPS spectra of the top-surface corrosion products of the as-
during sample preparation for XPS testing is unclear.
received sample, in 0.17 and 0.85 M solutions, and 30 °C/s HAZ in
0.17 M solution are shown in Fig. 11a. They similarly reflect the
compositions of the corrosion products with slight relevance to 3.5. Electrochemical impedance spectroscopy (EIS) tests
the microstructure and chloride concentration. The high-resolution
spectra of Fe 2p1/2, along with the binding energy ranges of Fe2O3 The EIS tests were carried out to analyze the growth of the corro-
and Fe3O4, cited from [74–76] for 2p3/2 and 2p1/2, are presented sion products after 5, 10, and 15 ks of immersion in OCP conditions.
in Fig. 11b. The two peaks of the as-received sample in 0.85 M solu- The corrosion products thickness and morphology with time
tion, as presented in Fig. 11c, were decomposed out of the fitting depend on the microstructure and the interfacial concentrations
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 387

Fig. 11. (a) XPS spectra of the corrosion products of the as-received steel in 0.17 M and 0.85 M chloride solutions, and of 30 °C/s HAZ in 0.17 M solution. (b) High-resolution Fe
2p spectra and (c) Curve fitting and peak decomposition of the high-resolution Fe 2p spectrum for the as-received steel in 0.85 M solution.

2
of water, H2CO3, HCO 3 , and CO3 . These two factors affect the sur- corrosion reactions and ionic transport. Fig. 12 presents a schematic
face distributions of the anodic vs. cathodic reactions, the rates representation of the ferritic–pearlitic microstructure of the as-
and mechanisms of dissolution, the species that result, the cathodic received steel along with the main interfacial interactions in a
reduction, hydrogen generation, and finally the corrosion products CO2-saturated corrosion environment. CO2 combines with water
2
formation and properties. As explained earlier, the steel microstruc- to form H2CO3, which dissociates into HCO þ
3 ; H and CO3 as follows
ture contains the active ferrite phase whose type, proportion, and [80,81]:
distribution control the anodic reactions, and it contains the stable
cementite (Fe3C) phase, which anchors the corrosion products and CO2ðgÞ $ CO2ðaqÞ 0:26 cm3 CO2 =g water; 90  C ð1Þ
is of a cathodic significance. In addition, there are the ‘‘foreign’’
phases; the corrosion products which form with kinetics interde- CO2ðaqÞ þ H2 O $ H2 CO3 ð2Þ
pendent with those of the surface – which, depending on their thick-
ness, composition and morphology, affect the course of the H2 CO3 $ Hþ þ HCO3 pK1 ¼ 6:4; 90  C ð3Þ
388 F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393

HCO3 $ Hþ þ CO2
3 pK2 ¼ 10:2; 90  C ð4Þ
In our case of a pH of 4.5, HCO
and carbonate are considered
3
insignificant to affect the course of the reactions. Schmitt reported
that H2CO3 dissociates in the close vicinity of the surface to result
in H+ ions that get reduced [82] – or that H2CO3 gets directly
reduced to result in adsorbed H atoms in a rate-determining step,
as reported by de Waard and Milliams as follows [83]:
H2 CO3 þ e ! Hads þ HCO3 ð5Þ
HCO
3 resulting from the reduction above, not the dissociation of
H2CO3, was reported in some mechanistic studies to be reduced
to CO23 , which engages with the local FeCO3 supersaturation.
Ogundele and White proposed that the reduction of bicarbonate
results in adsorbed H atoms in a rate-determining step, and it
combines with them to result in H2 molecules as follows [7]:

Fig. 12. Schematic representation of the interfacial anodic and cathodic interac- HCO3 þ e ! Hads þ CO2
3 ð6Þ
tions for the ferritic–pearlitic microstructure of the as-received steel in a
CO2-saturated solution.
HCO3 þ Hads þ e ! H2 þ CO2
3 ð7Þ

Fig. 13. Nyquist EIS plots of the as-received steel, and 10 and 60 °C/s HAZs in 0.17 M solutions, and of 10 °C/s HAZ in 0.85 M solution.
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 389

In Fig. 12, the cathodic reduction is figuratively represented


with a curved arrow from H2CO3 and H+ being reduced on the
pearlite grains to H2 and CO2
3 .
Regarding the multi-step dissolution of ferrite, it proceeds, as
indicated in the schematic, after H2O adsorption as follows [84]:
Fe þ H2 O ! FeOHads þ Hþ þ e ð8Þ

FeOHads ! FeOHþ þ e ð9Þ

FeOHþ þ Hþ ! Fe2þ þ H2 O ð10Þ


In Fig. 13, the Nyquist profiles of the as-received, 10, and 60 °C/s
HAZs in 0.17 and 0.85 M solutions are shown. The governing inter-
facial interactions seemed to be the same, but the rates of growth Fig. 15. The equivalent circuit fitting the EIS data of the ferritic–pearlitic, as-
of the corrosion products (and their protectiveness, accordingly) received steel sample during (a) the first stages of dissolution and (b) corrosion
were apparently affected by the microstructure. They were, there- products growth.

fore, of different admittance (Ycp) and porosity resistance (Rp)


values (cp in the subscript stands for ‘‘corrosion products’’ and p Ogundele and White, and Linter and Burstein reported that
stands for ‘‘porosity’’) which in turn affected the charge-transfer Fe(OH)2 could form first, but it reacts with CO2 and H2CO3, or
resistance (Rch) and double layer capacitance (Qdl) differently. HCO 3 (only likely if the pH in our case was at least 6.8) to result

The Nyquist profiles consisted of overlapped capacitive arcs repre- in a series of reactions in FeCO3 [7,88]. The corrosion products, as
senting the simultaneously-occurring processes of adsorption and represented in Fig. 14c, continue growing by the underlying disso-
filming during the first stages of CO2 corrosion [85,86]. The catho- lution, ionic transport through them, and the chemical and electro-
dic reduction as well occurs, but its surface proportion vs. those of chemical transformations that convert the carbonate-based to
dissolution and filming changes with time, which is difficult to oxide-based products.
estimate in relation to the growth of the corrosion products. The The equivalent {R(Q(R(QR)))} circuit, which is shown in Fig. 15
growth of the corrosion products, of behavior and properties influ- and previously considered in many CO2 corrosion studies, fits our
enced mainly by the microstructure, affects the most the low-fre- EIS data, as selectively shown in Fig. 16, and interprets the interfa-
quency impedance, from which the governing reactions can be cial transformations from the first stages of dissolution until com-
inferred. For the as-received sample in 0.17 M solution, the capac- plete filming. In Fig. 15, the constant phase element (CPE) is
itive arcs increased in size with time, indicating the increasing denoted in the beginning as Qp to represent the heterogeneous
charge-transfer resistance under thickening corrosion products. capacitance across the pearlite network, and later denoted as Qcp
With time, the surface area available for anodic adsorption to represent the capacitance across the corrosion products. Its
decreases as a result of the increasing ferrite dissolution, repre- admittance (Y) expression is as follows [89]:
sented in Fig. 14a, which becomes faster and less reaction-control- np np
ling. The cathodic reduction, which becomes the slower process, Y ¼ Y Q xn cos þ jY Q xn sin ð12Þ
2 2
mainly sustains onto the pearlite grains. The relatively high pro-
portion of pearlite makes the reactions increasingly under cathodic where x is the angular frequency and n is the CPE exponent.
control and facilitates the precipitation of the corrosion products The circuitry elements and values are presented selectively in
by their ‘‘anchoring’’ effect [37], and by stagnating Fe2+ and CO2 3 Tables 2 and 3 for all microstructures. The charge transfer of the
within the Fe3C bands so that they react and precipitate as FeCO3 as-received sample increased over time from 200 to 600 X cm2,
[38,87], as represented in Fig. 14b, as follows: and its double layer was pseudo-capacitive, as the nch was less than
0.7, especially in the beginning, indicating that the precipitation
Fe2þ þ CO2
3 ! FeCO3 ð11Þ began well before 5 ks of immersion.

Fig. 14. Schematic representation of the interfacial anodic and cathodic reactions, and the development of the corrosion products for a ferritic–pearlitic microstructure of the
as-received sample during (a) 5, (b) 10, and (c) 15 ks of free immersion at the OCPs.
390 F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393

Fig. 16. Selected Nyquist profiles showing fitting with the calculated EIS data from the equivalent {R(Q(R(QR)))} circuit.

Table 2
Equivalent circuit fitting of EIS data of the base steel and 10 and 60 °C/s HAZs in 0.17 M Cl solutions.

Circuitry Base steel 10 °C/s HAZ 60 °C/s HAZ


5 ks 10 ks 15 ks 5 ks 10 ks 15 ks 5 ks 10 ks 15 ks
n 2
Ycp (S s /cm ) 9.2E04 3.4E04 3.8E04 2.2E04 2.4E04 2.5E04 6.6E04 8.4E04 1.1E03
nfilm 0.80 0.79 0.78 0.80 0.82 0.82 0.77 0.80 0.81
Ydl (S sn/cm2) 1.7E03 2.9E03 2.7E03 2.0E02 2.1E02 5.5E02 3.2E02 8.5E02 7.0E02
nch 0.67 0.51 0.68 0.97 0.94 0.95 0.60 0.66 0.64
Rch (X cm2) 197.90 225.50 559.20 275.5 249.2 248.3 102.1 66.4 35.3
Rp (X cm2) 144.50 131.70 41.13 50.8 52.5 57.5 61.1 66.4 54.7
Chi-square 6.15  105 1.21  105 1.74  105 3.35  105 1.73  105 1.85  105 1.17  105 2.35  105 4.27  105

Table 3
a longer period of time, as explained in Section 3.1., during which
Equivalent circuit fitting of the EIS data of 10 °C/s HAZs in 0.85 M Cl solutions. the vigorous dissolution of ferrite was not matched by the physical
and electrochemical effects which would counteract it by greater
Circuitry 10 °C/s HAZ
pearlite proportions – in precipitating more effective and thicker
5 ks 10 ks 15 ks corrosion layers. The proportion of pearlite in 10 °C/s HAZ is low,
Ycp (S sn/cm2) 4.7E04 3.7E04 3.6E04 as represented in Fig. 17, leading to a predominant dissolution pro-
nfilm 0.73 0.76 0.77 cess of a capacitive surface, during which most Fe2+ ions diffuse to
Ydl (S sn/cm2) 3.4E03 3.4E03 4.0E03
the bulk solution without engaging in an effective precipitation
nch 0.83 0.88 0.75
Rch (X cm2) 317.20 333.20 314.10 process, which remained of unchanging porosity resistance. The
Rp (X cm2) 218.80 159.20 181.10 EIS response of 10 °C/s HAZ is interestingly different from what
Chi-square 1.40  105 7.30  105 1.49  105 we previously measured in bicarbonate solutions at 90 °C [31], in
which it exhibited a higher tendency to passivate with thick corro-
sion layers, which were simulated with multi-time-constant cir-
The admittance of the corrosion products (Ycp) decreased, as did cuitry. This reflects the significance of bicarbonate in higher-pH
the porosity resistance (Rp), indicating that the channels of trans- solutions in promoting passivation of the ferritic microstructures,
port were increasingly occupied by the precipitating products. unlike H2CO3 in lower-pH ones, which seems more of cathodic sig-
10 °C/s HAZ did not show an increase in charge-transfer resistance, nificance. The charge-transfer resistance of 60 °C/s HAZ, which had
which generally was lower than that of the as-received, and its an EIS response similar to that of 30 °C/s HAZ, was lower than that
double layer was noticeably more capacitive. This corresponds to of 10 °C/s HAZ and decreased over time. Given that its microstruc-
a stronger surface activity that counteracted the precipitation for tures intrinsically has low anodic and cathodic activity, the passive
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 391

Fig. 17. Schematic representation of the interfacial anodic and cathodic reactions, and the development of the corrosion products for the ferritic microstructure of 10 °C/s
HAZ during (a) 5, (b) 10, and (c) 15 ks of free immersion at the OCPs.

films could become considerably thinner and of high porosity 7. The governing reactions of all microstructures during free
increasing by the chloride pitting attack. Increasing the chloride immersion were generally independent of the microstruc-
concentration made the charge-transfer resistance higher under ture and salinity. The charge-transfer resistance of the fer-
thicker corrosion products that formed from exacerbated ritic–pearlitic microstructure of the base steel showed a
underlying dissolution, which generated more carbon-carrying, continuous increase with time as the corrosion products
film-forming adsorbents, as explained in Section 3.1. The corrosion were growing thicker and faster. The ferritic microstructure
products were, however, of higher admittance, as presented in of 10 °C/s HAZ remained active for a longer period of time
Table 3. and was more capacitive.
8. In the field, cooling down the welded pipeline HAZs at dif-
4. Conclusion ferent rates – depending on the welding method – affects
the corrosion behavior but not necessarily the integrity in
This detailed study investigated the interrelation between the the long run. Cooling the HAZs at low rates make them of
CO2 corrosion reactions, the corrosion products growth and charac- higher reactivity but later of thick corrosion layers that pro-
teristics, HAZ microstructures, and salinity. Consecutive OCP and tect them. Cooling them at high rates makes the dissolution
potentiodynamic polarization tests and independent EIS scans rates intrinsically low but the corrosion products become
were carried out in 1-bar CO2-saturated solutions of 0.17, 0.85, thin and slow-growing.
and 1.38 M chloride at 90 °C. The morphology and composition
of the corrosion products were analyzed by SEM, EDS, and XPS. Acknowledgment
Based on the EIS analysis, schematic diagrams depicting the inter-
facial interactions and the corrosion products growth were made. This publication was made possible by NPRP Grant # 09-211-2-
The following in brief, is concluded: 089 from the Qatar National Research Fund (a member of Qatar
Foundation).
1. The corrosion products generally suppress the anodic disso-
lution, rather than the cathodic reactions.
References
2. The ferritic microstructures of the API-X100 base steel and
10 °C/s HAZ dissolve at higher rates than the acicular-ferritic [1] F. Schremp, G. Roberson, Effect of supercritical carbon dioxide (CO2) on
and martensitic microstructures of 30 and 60 °C/s HAZs. construction materials, SPE J. 15 (1975) 227–233.
3. The base steel and 10 °C/s HAZ have a higher tendency to [2] T. Yoshida, S. Matsui, T. Atsuta, K. Itoga, Corrosion problems of pipeline and a
solution, Offshore Technology Conference, Paper no. 3891-MS, 1980.
form thick corrosion products, promoted more by the pearl- [3] R. Tuttle, Corrosion in oil and gas production, SPE J. 39 (1987) 756–762.
itic content of the base steel. Their low OCPs correspond to [4] M. Joosten, J. Kolts, J. Kiefer, P. Humble, J. Marlow, M. Marlow, Aspects of
high anodic activity while their high OCPs, after an applied Selective Weld and HAZ Attack in CO2 Containing Production Environments,
CORROSION/1996, Paper no. 79, NACE, 1996.
potential or a long time of free immersion, correspond to [5] B. Lu, J. Luo, Relationship between yield strength and near neutral pH stress
thick corrosion products suppressing the dissolution. corrosion cracking resistance of pipeline steels – an effect of microstructure,
4. 30 and 60 °C/s HAZs are of intrinsic low anodic and cathodic Corrosion 62 (2006) 129–140.
[6] S. Nesic, Key issues related to modelling of internal corrosion of oil and gas
activity and their corrosion products are thinner and more pipelines – a review, Corros. Sci. 49 (2007) 4308–4338.
vulnerable to chloride pitting. [7] G. Ogundele, W. White, Observations on the influences of dissolved
5. Increasing the chloride concentration exacerbates dissolu- hydrocarbon gases and variable water chemistries on corrosion of an API-
L80 steel, Corrosion 43 (1987) 665–673.
tion and acidifies more the corrosion products in a fashion [8] R. Jasinski, Corrosion of N80-type steel by CO2/water mixtures, Corrosion 43
that they become of higher tendency to grow thicker. This (1987) 214–218.
is associated to physicochemical and electrochemical effects [9] C. Li, Y. Wang, Y. Chen, Influence of peak temperature during in-service
welding of API X70 pipeline steels on microstructure and fracture energy of the
that will be investigated in future studies.
reheated coarse grain heat-affected zones, J. Mater. Sci. 46 (2011) 6424–6431.
6. The cathodic polarization currents are higher for thicker [10] G. Goodall, J. Gianetto, J. Bowker, M. Brochu, Thermal simulation of HAZ
corrosion products. This can be ascribed to the dissolution regions in modern high strength steel, Can. J. Metall. Mater. Sci. 51 (2012) 58–
of the products, the greater amounts of the conductive iron 66.
[11] W. Xu, Q. Wang, T. Pan, H. Su, C. Yang, Effect of welding heat input on
oxides and Fe3C, and the possibility for the cathodic reac- simulated HAZ microstructure and toughness of a V–N microalloyed steel, J.
tions to be more under mass-transfer control. Iron Steel Res. Int. 14 (2007) 234–239.
392 F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393

[12] G. Zhang, Y. Cheng, Micro-electrochemical characterization and Mott– [42] E. Dayalan, F. de Moraes, J. Shadley, E. Rybicki, S. Shirazi, CO2 Corrosion
Schottky analysis of corrosion of welded X70 pipeline steel in carbonate/ Prediction in Pipe Flow Under FeCO3 Scale-Forming Conditions, CORROSION/
bicarbonate solution, Electrochim. Acta 55 (2009) 316–324. 1998, Paper no. 51, NACE, 1998.
[13] H. Mitsui, R. Takahashi, H. Asano, N. Taniguchi, M. Yui, Susceptibility to stress [43] S. Papavinasam, R. Winston Revie, A. Doiron, Predicting Internal Pitting
corrosion cracking for low-carbon steel welds in carbonate–bicarbonate Corrosion of Oil and Gas Pipelines: Review of Corrosion Science Models,
solution, Corrosion 64 (2008) 939–948. CORROSION/2005, Paper no. 643, NACE, 2005.
[14] W. Zhao, W. Wang, S. Chen, J. Qu, Effect of simulated welding thermal cycle on [44] S. Nesic, J. Postlethwaite, S. Olsen, An electrochemical model for prediction of
microstructure and mechanical properties of X90 pipeline steel, Mater. Sci. corrosion of mild steel in aqueous carbon dioxide solutions, Corrosion 52
Eng. A 528 (2011) 7417–7422. (1996) 280–294.
[15] F. Fang, Q. Yong, C. Yang, H. Su, Microstructure and precipitation behavior in [45] J. Heuer, J. Stubbins, Microstructure analysis of coupons exposed to carbon
HAZ of V and Ti microalloyed steel, J. Iron Steel Res. Int. 16 (2009) 68–72. dioxide corrosion in multiphase flow, Corrosion 54 (1998) 566–575.
[16] W. Sun, S. Nesic, Kinetics of corrosion layer formation: Part 1 – Iron carbonate [46] A. Muñoz, J. Genesca, R. Duran, J. Mendoza, Mechanism of FeCO3 Formation on
layers in carbon dioxide corrosion, Corrosion 64 (2008) 334–346. API X70 Pipeline Steel in Brine Solutions Containing CO2, CORROSION/2005,
[17] J. Zhang, Z. Wang, Z. Wang, X. Han, Chemical analysis of the initial corrosion Paper no. 297, NACE, 2005.
layer on pipeline steels in simulated CO2-enhanced oil recovery brines, Corros. [47] L. Moiseeva, N. Rashevskaya, Effect of pH value on corrosion behavior of steel
Sci. 65 (2012) 397–404. in CO2-containing aqueous media, Russ. J. Appl. Chem. 75 (2002) 1625–1633.
[18] G. Schmitt, C. Bosch, M. Mueller, A Probabilistic Model for Flow Induced [48] L. Moiseeva, Carbon dioxide corrosion of oil and gas field equipment, Prot. Met.
Localized Corrosion, CORROSION/2000, Paper no. 49, NACE, 2000. 41 (2005) 76–83.
[19] R. Barker, X. Hu, A. Neville, The influence of high shear and sand impingement [49] J. Han, S. Nešić, Y. Yang, B. Brown, Spontaneous passivation observations
on preferential weld corrosion of carbon steel pipework in CO2-saturated during scale formation on mild steel in CO2 brines, Electrochim. Acta 56 (2011)
environments, Tribol. Int. (2013). <http://dx.doi.org/10.1016/ 5396–5404.
j.triboint.2012.11.015>. [50] H. Wang, H. Wang, H. Shi, C. Kang, P. Jepson, Why Corrosion Inhibitors Do not
[20] L. Moiseeva, O. Kuksina, On the dependence of steel corrosion in oxygen-free Perform Well in Some Multiphase Conditions: A Mechanistic Study,
aqueous media on pH and the pressure of CO2, Prot. Met. Phys. Chem. Surf. 39 CORROSION/2002, Paper no. 276, NACE, 2002.
(2003) 490–498. [51] J. Heuer, J. Stubbins, An XPS characterization of FeCO3 films from CO2
[21] J. Han, Y. Yang, B. Brown, S. Nesic, Electrochemical Investigation of Localized corrosion, Corros. Sci. 41 (1999) 1231–1243.
CO2 Corrosion on Mild Steel, CORROSION/2007, Paper no. 323, NACE, 2007. [52] D. Li, Y. Feng, Z. Bai, M. Zheng, Characteristics of CO2 corrosion scale formed on
[22] G. Zhang, M. Lu, Y. Qiu, X. Guo, Z. Chen, The relationship between the N80 steel in stratum water with saturated CO2, Appl. Surf. Sci. 253 (2007)
formation process of corrosion scales and the electrochemical mechanism of 8371–8376.
carbon steel in high pressure CO2-containing formation water corrosion [53] Z. Yin, Y. Feng, W. Zhao, Z. Bai, G. Lin, Effect of temperature on CO2 corrosion of
science and technology, J. Electrochem. Soc. 159 (2012) C393–C402. carbon steel, Surf. Interf. Anal. 41 (2009) 517–523.
[23] J. Sun, G. Zhang, W. Liu, M. Lu, The formation mechanism of corrosion scale and [54] B. Ingham, M. Ko, N. Laycock, J. Burnell, P. Kappen, J. Kimpton, D. Williams, In
electrochemical characteristic of low alloy steel in carbon dioxide-saturated situ synchrotron X-ray diffraction study of scale formation during CO2
solution, Corros. Sci. 57 (2012) 131–138. corrosion of carbon steel in sodium and magnesium chloride solutions,
[24] O. Yevtushenko, R. Bäßler, A. Pfennig, Corrosion behaviour of Cr13 steel in CO2 Corros. Sci. 56 (2012) 96–104.
saturated brine with high chloride concentration, Mater. Corros. 63 (2012) [55] R. Chang, J. Wagner Jr., Direct-current conductivity and iron tracer diffusion in
517–521. hematite at high temperature, J. Am. Ceram. Soc. 55 (1972) 211–213.
[25] V. Pandarinathan, K. Lepková, S. Bailey, R. Gubner, Evaluation of corrosion [56] A. Atkinson, R. Taylor, Diffusion of 55Fe in Fe2O3 single crystals, J. Phys. Chem.
inhibition at sand-deposited carbon steel in CO2-saturated brine, Corros. Sci. Solids 46 (1985) 469–475.
72 (2012) 108–117. [57] J. Crolet, N. Thevenot, S. Nesic, Role of conductive corrosion products in the
[26] F. Eliyan, F. Mohammadi, A. Alfantazi, An electrochemical investigation on the protectiveness of corrosion layers, Corrosion 54 (1998) 194–203.
effect of the chloride content on CO2 corrosion of API-X100 steel, Corros. Sci. [58] S. Nešić, L. Lunde, Carbon dioxide corrosion of carbon steel in two-phase flow,
64 (2012) 37–43. Corrosion 50 (1994) 717–727.
[27] F. Eliyan, A. Alfantazi, Electrochemical investigations on the corrosion behavior [59] D. Staicopolus, The role of cementite in the acidic corrosion of steel, J.
and corrosion natural inhibition of API-X100 pipeline steel in acetic acid and Electrochem. Soc. 110 (1963) 1121–1124.
chloride-containing CO2-saturated media, J. Appl. Electrochem. 42 (2012) [60] F. Eliyan, A. Alfantazi, Sensitivity of the passive films on API-X100 steel heat-
233–248. affected zones (HAZs) towards trace chloride concentrations in bicarbonate
[28] F. Eliyan, A. Alfantazi, Adsorption of oil onto API-X100 pipeline steel in CO2- solutions at high temperature, Mater. Corros. (2013), http://dx.doi.org/
saturated solutions, Metall. Mater. Trans. B 44 (2013) 1598–1604. 10.1002/maco.201206984.
[29] F. Manning, R. Thompson, Oilfield Processing Crude Oil, first ed., Pennwell, [61] G. Zhang, Y. Cheng, Micro-electrochemical characterization of corrosion of
Oklahoma, USA, 1995. welded X70 pipeline steel in near-neutral pH solution, Corros. Sci. 51 (2009)
[30] D. López, S. Simison, S. de Sánchez, The influence of steel microstructure on 1714–1724.
CO2 corrosion. EIS studies on the inhibition efficiency of benzimidazole, [62] D. Clover, B. Kinsella, B. Pejcic, R. De Marco, The influence of microstructure on
Electrochim. Acta 48 (2003) 845–854. the corrosion rate of various carbon steels, J. Appl. Electrochem. 35 (2005)
[31] F. Eliyan, A. Alfantazi, Corrosion of the heat-affected zones (HAZs) of API-X100 139–149.
pipeline steel in dilute bicarbonate solutions at 90 °C – an electrochemical [63] D. López, T. Pérez, S. Simison, The influence of microstructure and chemical
evaluation, Corros. Sci. 74 (2013) 297–307. composition of carbon and low alloy steels in CO2 corrosion. A state-of-the-art
[32] F. Eliyan, F. Icre, A. Alfantazi, Passivation of HAZs of API-X100 pipeline steel in appraisal, Mater. Des. 24 (2003) 561–575.
bicarbonate–carbonate solutions at 298 K, Mater. Corros. (2013), http:// [64] G. Zhang, Y. Cheng, Localized corrosion of carbon steel in a CO2-saturated
dx.doi.org/10.1002/maco.201206985. oilfield formation water, Electrochim. Acta 56 (2011) 1676–1685.
[33] S. Al-Hassan, B. Mishra, D. Olson, M. Salama, Effect of microstructure on [65] W. Sun, S. Nesic, Basics Revisited: Kinetics of Iron Carbonate Scale
corrosion of steels in aqueous solutions containing carbon dioxide, Corrosion Precipitation in CO2 Corrosion, CORROSION/2006, Paper no. 365, NACE, 2006.
54 (1998) 480–491. [66] G. Lin, M. Zheng, Z. Bai, X. Zhao, Effect of temperature and pressure on the
[34] M. Gao, X. Pang, K. Gao, The growth mechanism of CO2 corrosion product films, morphology of carbon dioxide corrosion scales, Corrosion 62 (2006) 501–507.
Corros. Sci. 53 (2011) 557–568. [67] T. Tanupabrungsun, B. Brown, S. Nesic, Effect of pH on CO2 Corrosion of Mild
[35] W. Sun, S. Nešić, R. Woollam, The effect of temperature and ionic strength on Steel at Elevated Temperatures, CORROSION/2013, Paper no. 48, NACE, 2013.
iron carbonate (FeCO3) solubility limit, Corros. Sci. 51 (2009) 1273–1276. [68] G. Schmitt, M. Mueller, M. Papenfuss, Understanding Localized CO2 Corrosion
[36] Y. Zhang, X. Pang, S. Qu, X. Li, K. Gao, Discussion of the CO2 corrosion of Carbon Steel from Physical Properties of Iron Carbonate Scales, CORROSION/
mechanism between low partial pressure and supercritical condition, Corros. 1999, Paper no. 38, NACE, 1999.
Sci. 59 (2012) 186–197. [69] Z. Bai, C. Chen, M. Lu, J. Li, Analysis of EIS characteristics of CO2 corrosion of
[37] K. Videm, J. Kvarekvaal, T. Perez, G. Fitzsimons, Surface Effects of the well tube steels with corrosion scales, Appl. Surf. Sci. 256 (2006) 7578–7584.
Electrochemistry of Iron and Carbon Steel Electrodes in Aqueous CO2 [70] F. Eliyan, A. Alfantazi, Influence of temperature on the corrosion behavior of
Solutions, CORROSION/1996, Paper no. 1, NACE, 1996. API-X100 pipeline steel in 1-bar CO2–HCO 3 solutions: an electrochemical
[38] C. Palacios, J. Shadley, Characteristics of corrosion scales on steels in a CO2- study, Mater. Chem. Phys. 140 (2013) 508–515.
saturated NaCl brine, Corrosion 47 (1991) 122–127. [71] E. Gulbrandsen, R. Nyborg, Effect of Steel Microstructure and Composition on
[39] S. Seal, W. Jepson, M. Gopal, K. Sapre, V. Desai, Surface Chemical and Inhibition of CO2 Corrosion, CORROSION/2000, Paper no. 23, NACE, 2000.
Morphological Changes in Corrosion Product Layers and Inhibitors in CO2 [72] A. Dugstad, H. Hemmer, M. Seiersten, Effect of steel microstructure on
Corrosion in Multiphase Flowlines, CORROSION/2000, Paper no. 46, NACE, corrosion rate and protective iron carbonate film formation, Corrosion 57
2000. (2001) 369–378.
[40] L. Xu, S. Guo, W. Chang, T. Chen, L. Hu, M. Lu, Corrosion of Cr bearing low alloy [73] C. Chen, M. Lu, D. Sun, Z. Zhang, W. Chang, Effect of chromium on the pitting
pipeline steel in CO2 environment at static and flowing conditions, Appl. Surf. resistance of oil tube steel in a carbon dioxide corrosion system, Corrosion 61
Sci. 270 (2013) 395–404. (2005) 594–601.
[41] X. Guo, Y. Tomoe, The effect of corrosion product layers on the anodic and [74] H. Mathieu, D. Landolt, An investigation of thin oxide films thermally grown
cathodic reactions of carbon steel in CO2-saturated MDEA solutions at 100 °C, in situ on Fe–24Cr and Fe–24Cr–llMo by auger electron spectroscopy and X-
Corros. Sci. 41 (1999) 1391–1402. ray photoelectron spectroscopy, Corros. Sci. 26 (1986) 547–559.
F.F. Eliyan, A. Alfantazi / Corrosion Science 85 (2014) 380–393 393

[75] B. Tan, K. Klabunde, P. Sherwood, X-ray photoelectron spectroscopy studies of [82] G. Schmitt, Fundamental Aspects of CO2 Corrosion, Corrosion 83, Paper no. 43,
solvated metal atom dispersed catalysts. Monometallic iron and bimetallic NACE International, Houston, TX, 1983.
iron–cobalt particles on alumina, Chem. Mater. 2 (1990) 186–191. [83] C. de Waard, D. Milliams, Carbonic acid corrosion of steel, Corrosion 31 (1975)
[76] K. Kishi, Adsorption of ethylenediamine on clean and oxygen covered Fe/ 177–181.
Ni(1 0 0) surfaces studied by XPS, J. Electron Spectrosc. Relat. Phenom. 46 [84] J. Bockris, D. Drazic, The kinetics of deposition and dissolution of iron: effect of
(1988) 237–247. alloying impurities, Electrochim. Acta 7 (1962) 293–313.
[77] D. López, W. Schreiner, S. de Sánchez, S. Simison, The influence of inhibitors [85] F. Farelas, M. Galicia, B. Brown, S. Nesic, H. Castaneda, Evolution of dissolution
molecular structure and steel microstructure on corrosion layers in CO2 processes at the interface of carbon steel corroding in a CO2 environment
corrosion: an XPS and SEM characterization, Appl. Surf. Sci. 236 (2004) 77–97. studied by EIS, Corros. Sci. 52 (2010) 509–517.
[78] Z. Cui, S. Wu, S. Zhu, X. Yang, Study on corrosion properties of pipelines in [86] Y. Chen, W. Jepson, EIS measurement for corrosion monitoring under
simulated produced water saturated with supercritical CO2, Appl. Surf. Sci. 252 multiphase flow conditions, Electrochim. Acta 44 (1999) 4453–4464.
(2006) 2368–2374. [87] T. Berntsen, M. Seiersten, T. Hemmingsen, Effect of FeCO3 supersaturation and
[79] B. Wang, M. Du, J. Zhang, C. Gao, Electrochemical and surface analysis studies carbide exposure on the CO2 corrosion rate of carbon steel, Corrosion 69
on corrosion inhibition of Q235 steel by imidazoline derivative against CO2 (2013) 601–613.
corrosion, Corros. Sci. 53 (2011) 353–361. [88] B. Linter, G. Burstein, Reactions of pipeline steels in carbon dioxide solutions,
[80] D. Li, Z. Duan, The speciation equilibrium coupling with phase equilibrium in Corros. Sci. 41 (1999) 117–139.
the H2O–CO2–NaCl system from 0 to 250 °C, from 0 to 1000 bar, and from 0 to [89] B. Boukamp, Equivalent Circuit User’s Manual, second ed., University of
5 molality of NaCl, Chem. Geol. 244 (2007) 730–751. Twente, Enschede, 1989.
[81] Physical and Engineering Data, January 1978 ed., Shell Internationale
Petroleum Maatschappij BV, The Hague, 1978.

You might also like