Engineering Failure Analysis 89 (2018) 46–56

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Failure analysis of a wet gas pipeline due to localised CO2 corrosion

T
⁎
Mehdi Javidi , Shima Bekhrad
Department of Materials Science and Engineering, School of Engineering, Shiraz University, Shiraz, Iran

A R T IC LE I N F O ABS TRA CT

Keywords: CO2 corrosion is one of the main causes of failures of pipelines transporting wet gas. It can
Carbonate scale happen in general and localised form. In this study, the impact of the presence of iron carbonate
CO2 corrosion scale which formed during process on CO2 corrosion behaviour of API 5L X65 carbon steel has
Galvanic corrosion been investigated by electrochemical measurements. The chemical composition and morphology
Pipeline failures
of the scale has been studied via X-ray diffraction and scanning electron microscope. The scales
were found to be iron carbonate scale with trace amount of iron sulphide. The result of elec-
trochemical analysis revealed that the uniform corrosion rate of carbon steel decreases in the
presence of iron carbonate scale. Furthermore, the bare and continuously covered sample with
iron carbonated did not show susceptibility to localised CO2 corrosion. However, the presence of
uneven iron carbonate scale or formation of thin, porous or defected scale can result in a po-
tential difference between this area and adjacent area covered with compact scale followed by
localised corrosion.

1. Introduction

Carbon steel is the preferred material choice for oil and gas transmission and distribution pipeline systems. API 5L is a material
specification for the application of carbon steel in line pipes carrying gas, oil and water which provided in variety of heat treatment
and mechanical strength. This class of material is susceptible to CO2 corrosion in wet gas services. Pipeline failures are serious and
expensive problems that might even end with catastrophic accidents [1–3].
The term “wet gas” means unprocessed natural gas containing liquid hydrocarbon (gas condensate) and water. When CO2 which
accompanying natural gas dissolves in formation water, carbonic acid is formed which is a weak, but corrosive acid to the carbon
steel. The influencing parameters on CO2 corrosion are CO2 partial pressure, pH, temperature and flow rate and the presence of scales
which have thoroughly been discussed in the literature elsewhere [4–6]. It is reported that in the absence of protective scales,
corrosion rate increases with increasing CO2 partial pressure and in high values of pH, increasing the CO2 partial pressure enhances
the protective scale precipitation which decreases corrosion rate [7]. Furthermore, increasing temperature accelerates the CO2
corrosion rate in low pH, but in higher pH, as iron carbonate solubility decreases, formation of protective scales decreases the
corrosion rate [6]. In the absence of inhibitor and non-scaling conditions (pH < 6, T < 60 °C), the increase in flow rate increases
corrosion rate. On the other hand, in case of scaling condition (pH ≥ 6, T ≥ 60 °C) or presence of inhibitor film, the effect of flow rate
is insignificant [4]. The mechanism of CO2 corrosion for carbon steel is widely discussed and the iron carbonate (FeCO3) has been
proposed as the main corrosion product. Iron carbonate scale can be protective or non-protective depending on the water chemistry
and temperature [4,6]. It is well-known that the media under the scale has different physical and chemical properties compared to the
bulk electrolyte which can result in under deposit corrosion as a localised accelerated corrosion [8–10]. Thus, removing of scales is
important for preventing corrosion. The use of scale inhibitors and mechanical removing via pigging are common removal methods of

⁎
Corresponding author.
E-mail address: mjavidi@shirazu.ac.ir (M. Javidi).

https://doi.org/10.1016/j.engfailanal.2018.03.006
Received 18 July 2017; Received in revised form 24 December 2017; Accepted 3 March 2018
Available online 04 March 2018
1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

hard scales and other deposits such as sand from the internal of the pipelines [11,12].
A number of environmental factors have been associated with the onset of localised corrosion of carbon steel including poor
corrosion inhibition, local water separation in oil-water flow, differential condensation in wet gas flow, and flow disturbances such as
weld beads, flanges, the presence of bacteria, solids, organic acids (such as acetic acid, formic acid and propionic acid) [13], hy-
drogen sulphide. On the other hand, temperature, pH of the solution, CO2 partial pressure, salt concentration, flow rate and flow
regime are the key environmental affecting parameters on the propagation of the localised CO2 corrosion [14–16].
In this study, failure of an API 5L X65 carbon steel pipeline which failed because of internal localised CO2 corrosion has been
investigated. Also, the susceptibility of X65 carbon steel to localised CO2 corrosion in the presence of iron carbonate scale has been
studied.

2. Materials and methods

2.1. History of the pipeline

The investigated pipeline is a buried flow line which carrying wet gas from a gas condensate well to a gas manifold. The material
of the pipe is API 5L X65 with outside diameter of 273 mm (10″ nominal pipe size, NPS) and 8.7 mm thickness which 4034 m long and
with 3-layer polyethylene external coating. Also, the operating temperature of the pipeline is 55–60 °C and the pH was reported to be
4.8. Furthermore, the well-head pressure is 16.9 MPa with 72 ppm H2S concentration in the gas phase. In other words, the partial
pressure of H2S is 3.3 kPa. The salt concentration in the water phase is reported to be around 50,000 ppm (30,000 ppm chloride ion).

2.2. Visual inspection

First of all, the location of the pipeline failure was inspected and the section of the pipe which failed due to corrosion was cut and
delivered to laboratory for further studies. The internal corrosion of the pipe and also the presence on internal scale were investigated
via scanning electron microscopy as mentioned in Section 2.5. Also, the primary data including the history of the pipeline con-
struction and maintenance and also process condition were collected from the site.

2.3. Material characterisation

The chemical composition of the pipe was investigated by optical emission spectroscopy (OES, Foundry-Master Pro).
Furthermore, in order to investigate the phase composition of the surface scale and view the cross sectional of the scale and per-
forming electrochemical measurements some samples were cut from the body of the pipe. Three samples with 10 mm × 10 mm
dimension were prepared, mounted in epoxy resin and wet ground up to 2000 grit SiC paper, then etched in Nital 2% and their
microstructure was investigated via Leica optical microscope.

2.4. Phase composition and surface investigation of scale

Phase composition analysis of the hard scale was investigated via X-ray diffraction (XRD) technique using Bruker D8 Advance
instrument operating at 40 kV and 30 mA with Cu-Kα radiation (λ = 0.15405 nm) with the scan rate of 3°/min. For this purpose, a
sample with 10 mm × 10 mm dimension of the pipe with the presence of surface scale was used. The resulted XRD pattern was
analysed by X'pert Highscore Plus Software Version 2.2.2 (PANalytical software). Furthermore, the surface morphology and cross
section of the samples were investigated via scanning electron microscope (SEM, Cambridge Stereo-scan S360).

2.5. Electrochemical measurements

The surface area for both bare and scale covered samples was set to be 0.25 cm2 with dimensions of 5 mm × 5 mm. A wire was
soldered to the rear surface of the samples then; bare and scale covered samples were embedded in epoxy resin. The surface of some
specimen was wet ground via 80, 180, 400 and 600 grit SiC papers then degreased by ethanol and rinsed with doubled distilled water
and air dried for the study of the bare sample.
In order to investigate the effect of the presence of iron carbonate scale on CO2 corrosion behaviour of X65 carbon steel, all the
experiments were performed on both scale covered and bare samples. Electrochemical measurements were performed in a three-
electrode-glass cell with API 5L X65 carbon steel as working electrode (WE) and platinum wire and saturated Ag/AgCl electrode as
reference and counter electrodes, respectively. The working electrode was placed between the reference and counter electrode close
to the reference electrode with fixed distance. The cell configuration is schematically shown in Fig. 1. The glass cell was filled with
400 ml of 3.5 wt% sodium chloride (NaCl; Merck) solution. Before each experiment N2 gas was purged for 30 min to deoxygenate the
solution. Then CO2 gas was purged for 1 h in order to saturate the solution. It is necessary to note that purging of CO2 was continued
over the solution during the experimental measurements to maintain the saturation condition.
Electrochemical measurements were performed using potentiostat/galvanostat μAutolab-Type III. First of all, the open circuit
potential (OCP) of the samples were monitored with time until the potential of the electrodes become stabilised. Then electro-
chemical impedance spectroscopy (EIS) was performed on the samples. The excitation amplitude of 10 mV over a frequency range of
100 kHz to 0.01 Hz at 10 points per decade was conducted using Frequency Response Analyser Software (FRA version 4.9.006). All

47
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 1. The schematic view of the electrochemical cell.

EIS spectra were fitted and analysed via ZView™ Software (3.4 Scribner Associates Inc.). Furthermore, cyclic polarisation test was
performed on new samples. The potential sweep was started from −300 mV with respect to OCP in the positive direction up to
+700 mV, and then the polarisation was reversed to the start potential at scan rate of 1 mV/s. The cyclic polarisation was performed
using General Purpose for Electrochemical Systems Software (GPES Version 4.9.006). All experiments were performed at 30, 50 and
70 °C and prior to each test, the pH of the solution was measured with a Metrohm pH/temperature (°C) meter with (accuracy
of ± 0.01) which was calibrated with buffer solution (pH 4 and pH 7) and found to be 3.9, 4.3 and 4.5 at 30, 50 and 70 °C, re-
spectively.

3. Results and discussion

3.1. Material characterisation

The chemical composition of the investigated material which studied via optical emission spectroscopy is reported in Table 1. As
can be seen, the material met the specified range of API 5L X65 carbon steels.
Fig. 2(a) shows the location of the failure of the investigated pipeline. Also, Fig. 2(b) and (c) shows the dimension of the localised

Table 1
The chemical composition of API 5L X65 carbon steel (wt%).

Element Fe C Mn Si P Cr Al V

wt% Bal. 0.053 1.29 0.21 0.014 0.012 0.039 0.029

48
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 2. (a). Location of the failure of the investigated pipeline. (b). A view of the corroded area. (c). Internal views of the pipe.

corrosion and the presence of internal scale through the internal surface of the pipe. This type of localised CO2 corrosion is usually
named as “mesa attack” which included large areas free of corrosion products that have corroded severely and sharply separated from
the surrounded protected areas which covered with a corrosion product (here iron carbonate) [14].
Furthermore Fig. 3(a) and (b) shows the surface morphology and the thickness of scale on the internal part of the pipeline,
respectively. As can be seen in Fig. 3(b), the scale shows a porous structure. It is believed that the iron carbonate scale which
precipitates at temperature below 70 °C is loose and porous [17]. Also Fig. 4 shows the depth of the pit in the corroded area. On the
other hand, Fig. 5 shows the microstructure of the investigated carbon steel sample. The microstructure consists of fine-grained ferrite
matrix with distributed pearlite, which is in agreement with the same microstructure investigated by other researchers [18,19].

3.2. Chemical composition of the scale

The XRD pattern of the surface scale is shown in Fig. 6 which revealed the presence of mainly iron carbonate (siderite) with the
chemical formula of FeCO3 and rhombohedral crystal structure and minor amount of FeS (Mackinawite). No peak associated with
iron was seen in the X-ray sweep which means that the iron carbonate scale was thick enough to prevent the passage of diffraction

Fig. 3. (a). Surface morphology of iron carbonate scale. (b). Cross sectional view of the base metal and iron carbonate scale.

49
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 4. The depth of the pit in the corroded area.

Fig. 5. The microstructure of the investigated API 5L X65 carbon steel after metallography.

from the steel substrate [20]. By considering Fig. 3(b), the thickness of the iron carbonate scale was found to be around 200 μm.
The iron carbonate is the corrosion product of carbon steel via CO2 corrosion. As the wet gas containing H2S gas, the FeS is also
formed as corrosion product via below reactions [21]:
First of all, CO2 is dissolved and then forms weak H2CO3 acid which is corrosive to carbon steel:
CO2 (g ) ⇌ CO2 (aq·) (1)

CO2(aq·) + H2 O(l) ⇌ H2 CO3 (2)

Corrosive species, H , +
HCO3− 2−
and CO3 might reduce on the steel surface. Reduction of the reducing species is a pH dependent
process. At pH < 4 cathodic reduction of H+ is the prominent reaction and when pH increases to the intermediate range of
4 < pH < 6 reduction of H2CO3 or HCO3− plays an important role. When pH > 6 direct reduction of water signifies an important
role in determining the corrosion rate.
H+ + e−⟶H (3)

2H2 CO3 + 2e−⟶2HCO3− + H2 (4)

2HCO3− + 2e−⟶2CO32 − + H2 (5)

2H2 O + 2e−⟶2OH− + H2 (6)

50
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 6. The XRD pattern of the scale which present in the internal surface of the pipe.

Iron dissolution which is a multistep anodic reaction balances the possible cathodic reactions by following reaction [22]:
Fe⟶Fe 2 + + 2e− (7)
Thus, the overall reaction in CO2-saturated solution which leads to the formation of iron carbonate can be shows as:
Fe(s) + CO2(g ) + H2 O(l) ⇌ FeCO3(s) + H2(g ) (8)
On the other hand, H2S gas dissolves in aqueous solution via below reaction and can result in formation of Fes as corrosion
product.
H2 S(aq·) ↔ H+ + HS− (9)

HS− ↔ H+ + S2 − (10)

Fe + H2 S + H2 O⟺FeSH−ads + H3 O+ (11)

FeSH− ads ⟺Fe(SH)ads + e− (12)

Fe(SH)ads ⟺FeSH+ + e− (13)

The FeSH+ can change to FeS on the electrode surface via below reaction:
FeSH+ → FeS(1 − x) + xSH− + (1 − x)H+ (14)
It is necessary to note that in the media containing H2S, different polymorphous phases of iron sulphide can form including
amorphous iron sulphide (FeS), mackinawite (FeS), cubic ferrous sulphide (FeS), troilite (FeS), pyrrhotites (FeS), smythite
(Fe(3+X)S4), greigite (Fe3S4), pyrite (FeS2) and marcasite (FeS2) [23].
As the concentration of the H2S in the investigated wet gas is low (72 ppm in the gas phase) the fraction of FeS in the corrosion
product was found to be low.

3.3. Electrochemical measurements

3.3.1. Open circuit potential measurements (OCP)

Fig. 7 shows the comparison of the OCP of the iron carbonate scale covered and bare X65 carbon steel in 3.5 wt% NaCl solution at
30, 50 and 70 °C. According to Fig. 7 it can be seen that the OCP of the scale covered sample is more positive than the bare surface
which means carbon steel in the presence of the carbonate scale is less active. Furthermore, comparison of the OCP of the bare and
iron carbonate scale covered sample at 50 °C after 2000 S shows about 55 mV potential difference. This observation indicates that a
galvanic cell can be established between the bare and protected surfaces.

3.3.2. Electrical impedance spectroscope (EIS)

Fig. 8 shows the Nyquist and Bode plots of bare API 5L X65 carbon steel in 3.5 wt% NaCl solution at 30, 50 and 70 °C. All the
Nyquist plots have a same shape of a depressed semicircle and the only difference is their diameter which corresponded to the
polarisation (corrosion) resistance. According to Fig. 8(a) the diameter of the Nyquist plots decreases by increasing temperature,
which means that increase in temperature, causes the corrosion process to proceed with higher rates. The depressed semicircle is
attributed to the double layer capacitance which means the corrosion process is under charge transfer control. The number of

51
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 7. The comparison of the open circuit potentials of the bare X65 carbon steel and iron carbonate scale covered samples in 3.5 wt% NaCl solution at 30, 50 and
70 °C.

Fig. 8. (a). Nyquist plots. (b). Bode plots of the bare API 5L X65 carbon steel in 3.5 wt% NaCl solution at 30, 50 and 70 °C.

52
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 9. (a). Nyquist plots. (b). Bode plots of the iron carbonate scale covered samples in 3.5 wt% NaCl solution at 30, 50 and 70 °C.

semicircle in Nyquist plots defines the number of dominant electrochemical reactions which control the corrosion process. Fig. 8(b)
shows the Bode-magnitude and Bode phase plots of the bare X65 carbon steel samples which depicts only one time constant.
Fig. 9 shows the Nyquist and Bode plots of iron carbonate scale covered samples in 3.5 wt% NaCl solution at 30, 50 and 70 °C.
According to Fig. 9(a), the Nyquist plots consist of two depressed semicircles which means that the presence of iron carbonate scale
alters the CO2 corrosion mechanism. Furthermore, the diameter of the Nyquist plots has increased in the presence of iron carbonate
scale, as the scale has acted as a physical barrier to the corrosive species to reach the surface. Furthermore, the surface active area of
the scale covered steel is less than the bare steel as the carbonate scale has uniformly covered the surface. The diameter of the Nyquist
plots decreased with increasing temperature, which indicates increasing temperature increases the corrosion rate. According to
Fig. 9(b), there are two time constants visible in Bode plots.
All impedance investigations were fitted by electrical equivalent circuits shown in Fig. 10. The parameters used for data fitting are
Rs (solution resistance), Rct (charge transfer resistance), CPEdl (constant phase element of the double layer), CPEscale (constant phase
element describing the porous scale), Rpore (pore resistance of the scale). The constant phase element (CPE, Z(CPE)=Y0−1(jw)n) was
used instead of pure capacitors to reduce errors and increase the precision of data fitting. The parameter n is a non-integer parameter
(−1 ≤ n ≤ 1). The n parameter defines the deviation from ideal behaviour; n = 1 means an ideal capacitor, n = 0 indicates the
behaviour of an ideal resistor, and n = −1 indicates an ideal inductor behaviour. The non-ideal behaviour which was seen in all EIS
measurements is usually attributed to the surface porosities and non-homogeneities of the surface scale.
The extracted electrochemical parameters are reported in Table 2. In this table decrease in charge transfer resistance (Rct) and
pore resistance (Rpore) and on the other hand increase in Y0 with increasing temperature correspond to the increase in corrosion rate.
Furthermore, the data also shows increase in CPEdl and decrease in Rct for the scale covered sample relative to the bare sample in all
investigated temperature. Thus, according to the EIS measurements, the rate of uniform CO2 corrosion behaviour of the carbon steel
decreases in the presence of iron carbonate scale. In order to investigate the localised corrosion behaviour of carbon steel with and
without presence of carbonate scale, cyclic polarisation investigations were conducted.

53
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 10. Electrical equivalent circuit used for fitting the resulted EIS data for API 5L X65 carbon steel in CO2-saturated 3.5 wt% NaCl solution. (a). For the bare sample.
(b). For the iron carbonate covered sample (Rs: solution resistance, Rct: charge transfer resistance, CPEdl: constant phase element of the double layer, CPEscale: constant
phase element describing the porous scale, Rpore: pore resistance of the scale).

Table 2
Electrochemical parameters extracted from impedance measurements of the bare and iron carbonate scale covered API 5L X65 carbon steel in CO2-saturated 3.5 wt%
NaCl in 30, 50 and 70 °C.

Corrosion rate CPEdl CPEscale

(mm/y)
2 3
Samples T (°C) Rs (Ω·cm ) Rct Y0 × 10 n Rpore Y0 × 103 n
(Ω·cm2) (sn/cm2·Ω) (sn/cm2·Ω)

Bare samples 30 2.19 70.64 1.13 0.80

50 2.76 43.70 1.3 0.80
70 0.97 28.21 2.5 0.81
Iron carbonate scale covered samples 30 3.19 146.7 82 0.99 59.6 7 0.69
50 3.65 87.88 104 0.81 71.5 8.1 0.65
70 1.56 48.88 46 0.47 5.83 4 0.68

3.3.3. Cyclic polarisation measurements

Fig. 11 shows the cyclic polarisation curves for bare and carbonate covered X65 carbon steel in CO2-saturated 3.5 wt% solution at
30, 50 and 70 °C. Cyclic hysteresis can provide information on susceptibility to pitting corrosion and how readily a passive film can
repair itself. In other words, positive hysteresis (when the current density of reverse scan is greater than the forward scan) corre-
sponds to susceptibility to pitting. In other words, in this case the damaged passive film is not repaired and pits initiate. On the other
hand in case of negative hysteresis, the damaged passive film repairs itself and pits do not initiate [24].
According to the Fig. 11(a) as it is expected, the bare carbon steel does not show susceptibility to localised corrosion as the current
density of the reverse scan is less than the current density of the forward scan which corresponds with negative hysteresis. Fur-
thermore, as it is seen in Fig. 11(a), the current densities have increased with increasing temperature, which means that as expected,
the corrosion rate has increased with increasing temperature. Fig. 11(b) represents the cyclic polarisation curves of iron carbonate
scale covered samples. Negative hysteresis with lower current densities in the reverse scans can be seen in all curves which corre-
spond to negative hysteresis. Thus, the X65 carbon steel in the presence of iron carbonate scale does not show susceptibility to
localised corrosion. Furthermore, the repassivation potential for the iron carbonate covered steel was found to happen in potentials
more noble than the corrosion potential. This means that if pit initiate in high potentials, they can't grow. The polarisation parameters
are parameters are reported in Table 3.
Thus, by considering the above discussion, the reason for happening localised CO2 corrosion and formation of localised pit in the
investigated pipe could be due to formation of micro galvanic cell between the scale precipitated and bare surface. In other words, as
reported in Fig. 7, the open circuit potential of the iron carbonate covered carbon steel is nobler than the bare samples. Thus, the
potential difference between the bare and iron carbonate covered surface can result in the formation of galvanic cell in which the iron
carbonate covered surface is cathode while the bare area is anode. Therefore, the difference between open circuit potential of the bare
and iron carbonate covered surface is a significant driving force for galvanic current followed by localised CO2 corrosion [14–16].
The galvanic mechanism and propagation kinetics of the localised CO2 corrosion have been studied by Han et al. [Corrosion Vol.66,
no.9]. They showed that the bare and iron carbonate layer covered carbon steel can form a galvanic cell in CO2 corrosion in which
bare steel surface acts as anode and the iron carbonated covered parts act as cathode resulting in localised CO2 corrosion. They also
investigated the effect of pH, CO2 partial pressure, salt concentration and flow rate on the localised CO2 corrosion propagation rate.

54
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

Fig. 11. Cyclic polarisation curves for API 5L X65 carbon steel in CO2-saturated 3.5 wt% NaCl solution at 30, 50 and 70 °C. (a). Bare sample. (b). Iron carbonate
covered samples.

Table 3
Electrochemical parameters extracted from cyclic polarisation curves of the bare and iron carbonate scale covered API 5L X65 carbon steel in CO2-saturated 3.5 wt%
NaCl in 30, 50 and 70 °C.

Sample T (°C) Ecorr (V vs. Ag/AgCl) icorr × 105 (A·cm−2) ba (V/dec.) bc (V/dec.) Corrosion rate (mm/year)

Bare samples 30 −0.684 4.57 0.051 0.503 0.30

50 −0.712 7.15 0.047 0.365 0.47
70 −0.670 10.70 0.076 0.441 0.70
Iron carbonate scale covered samples 30 −0.672 0.77 0.083 0.172 0.05
50 −0.714 0.91 0.077 0.115 0.06
70 −0.720 4.71 0.101 0.198 0.31

They also concluded that working temperature above 50 °C, pH range of 5.5–6.5, high salt concentration and turbulent flow condition
are the key environmental condition for localised CO2 corrosion propagation. Considering the process condition of the investigated
pipeline which reported in Section 2.1, it can be concluded that the process condition is close to the requirement condition for
localised CO2 corrosion propagation.

55
M. Javidi, S. Bekhrad Engineering Failure Analysis 89 (2018) 46–56

4. Conclusions

It was found here that the uniform CO2 corrosion rate of API 5L X65 carbon steel decreases in the presence of iron carbonate scale
as the scale acts as a barrier to the corrosive species by covering the surface. According to the results, bare and iron carbonate covered
X65 carbon steel are not susceptible to localised attack in CO2-saturated aqueous solution. However, formation of micro galvanic cell
between the stagnant environment under the scale and the bulk electrolyte and also the galvanic cell between scale covered and bare
area are the main causes of localised CO2 corrosion. The driving force for localised CO2 corrosion is the difference between open
circuit potential of the bare and iron carbonate covered area.

References

[1] C.-H. Peng, Z.-Y. Liu, X.-Z. Wei, Failure analysis of a steel tube joint perforated by corrosion in a well-drilling pipe, Eng. Fail. Anal. 25 (2012) 13–28.
[2] Y. Liu, B. Zhang, Y. Zhang, L. Ma, P. Yang, Electrochemical polarization study on crude oil pipeline corrosion by the produced water with high salinity, Eng. Fail.
Anal. 60 (2016) 307–315.
[3] S.S. Abedi, A. Abdolmaleki, N. Adibi, Failure analysis of SCC and SRB induced cracking of a transmission oil products pipeline, Eng. Fail. Anal. 14 (2007)
250–261.
[4] S. Nešić, Key issues related to modelling of internal corrosion of oil and gas pipelines–a review, Corros. Sci. 49 (2007) 4308–4338.
[5] M. Kermani, A. Morshed, Carbon dioxide corrosion in oil and gas production—a compendium, Corrosion 59 (2003) 659–683.
[6] G. Schmitt, M. Horstemeier, Fundamental aspects of CO2 metal loss corrosion-part II: influence of different parameters on CO2 corrosion mechanisms,
CORROSION 2006, NACE International, Houston/TX, 2006(Paper No. 06112).
[7] S. Nesic, J. Lee, V. Ruzic, A mechanistic model of iron carbonate film growth and the effect on CO2 corrosion of mild steel, CORROSION 2002, NACE
International, Houston/TX, 2002(Paper No.02237).
[8] J. Huang, B. Brown, X. Jiang, B. Kinsella, S. Nesic, Internal CO2 corrosion of mild steel pipelines under inert solid deposits, CORROSION 2010, NACE
International, Houston/TX, 2010(Paper No. 10379).
[9] J.R. Vera, D. Daniels, M.H. Achour, Under deposit corrosion (UDC) in the oil and gas industry: a review of mechanisms, testing and mitigation, CORROSION
2012, NACE International, Houston/TX, 2012(Paper No. C2012-0001379).
[10] V. Pandarinathan, K. Lepková, W. Van Bronswijk, Chukanovite (Fe2(OH)2CO3) identified as a corrosion product at sand-deposited carbon steel in CO2-saturated
brine, Corros. Sci. 85 (2014) 26–32.
[11] Y. Tan, Y. Fwu, K. Bhardwaj, Electrochemical evaluation of under-deposit corrosion and its inhibition using the wire beam electrode method, Corros. Sci. 53
(2011) 1254–1261.
[12] A. Pedersen, K. Bilkova, E. Gulbrandsen, J. Kvarekvål, CO2 corrosion inhibitor performance in the presence of solids: test method development, CORROSION
2008, NACE International, Houston/TX, 2008(Paper No. 08632).
[13] G. Zhang, Y. Cheng, On the fundamentals of electrochemical corrosion of X65 steel in CO2-containing formation water in the presence of acetic acid in petroleum
production, Corros. Sci. 51 (2009) 87–94.
[14] J. Han, B. Brown, S. Nešic, Investigation of the galvanic mechanism for localized carbon dioxide corrosion propagation using the artificial pit technique,
Corrosion 66 (2010) 095003–095012.
[15] Y. Sun, K. George, S. Nesic, The effect of Cl−and acetic acid on localized CO2 corrosion in wet gas flow, CORROSION 2003, 2003 (Paper No.03327).
[16] J. Han, Y. Yang, B. Brown, S. Nesic, Electrochemical investigation of localized CO2 corrosion on mild steel, Corrosion/2007, 2007 (paper).
[17] M. Gao, X. Pang, K. Gao, The growth mechanism of CO2 corrosion product films, Corros. Sci. 53 (2011) 557–568.
[18] D. Clover, B. Kinsella, B. Pejcic, R. De Marco, The influence of microstructure on the corrosion rate of various carbon steels, J. Appl. Electrochem. 35 (2005)
139–149.
[19] A. Cervantes-Tobón, J. Godínez-Salcedo, J.L. Gonzalez-Velazquez, M. Díaz-Cruz, Corrosion rates of API 5L X-52 and X-65 steels in synthetic brines and brines
with H2S as a function of rate in a rotating cylinder electrode, Int. J. Electrochem. Sci. 9 (2014) 2454–2469.
[20] C. Palacios, J. Shadley, Characteristics of corrosion scales on steels in a CO2-saturated NaCl brine, Corrosion 47 (1991) 122–127.
[21] S. Nesic, J. Postlethwaite, S. Olsen, An electrochemical model for prediction of corrosion of mild steel in aqueous carbon dioxide solutions, Corrosion 52 (1996)
280–294.
[22] J.M. Bockris, D. Drazic, A. Despic, The electrode kinetics of the deposition and dissolution of iron, Electrochim. Acta 4 (1961) 325–361.
[23] J. Ning, Y. Zheng, D. Young, B. Brown, S. Nešić, Thermodynamic study of hydrogen sulfide corrosion of mild steel, Corrosion 70 (2013) 375–389.
[24] W.S. Tait, An Introduction to Electrochemical Corrosion Testing for Practicing Engineers and Scientists, PairODocs Publications, 1994.

56