Corrosion Science 88 (2014) 5665

Contents lists available at ScienceDirect

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

Carbon steel corrosion in clay-rich environment

Y. El Mendili a,, A. Abdelouas a, A. Ait Chaou a, J.-F. Bardeau b, M.L. Schlegel c
a

SUBATECH, CNRS-IN2P3, Ecole des Mines de Nantes, Universit de Nantes, 4 rue Alfred Kastler, BP 20722, 44307 Nantes cedex 03, France
LUNAM Universit, Institut des Molcules et Matriaux du Mans, UMR CNRS 6283, Universit du Maine, Avenue Olivier Messiaen, 72085 Le Mans Cedex 9, France
c
CEA, DEN, DANS/DPC/SEARS/LISL, F-91191 Gif-sur-Yvette, France
b

a r t i c l e

i n f o

Article history:
Received 22 April 2014
Accepted 10 July 2014
Available online 18 July 2014
Keywords:
A. Carbon steel
B. Raman spectroscopy
B. SEM
C. Microbiological corrosion

a b s t r a c t
We investigated the carbon steel corrosion in carbon dioxide clay-rich environment to understand its
behavior under geological conditions. The results show the formation of magnetite as the main corrosion
product in the rst step of the corrosion process, followed by the formation of different corrosion products with complex mixtures of iron-oxide, hydroxycarbonate, hydroxychloride and sulde phases. These
results strongly contrast with similar experiments conducted under H2 atmosphere where the major corrosion products consisted of iron suldes. It appears then important to consider all the geochemical
parameters including gas composition to better study corrosion of steel buried in geological formations.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
Deep geological disposal remains the preferred option for management of long-lived radioactive waste in several countries
including Belgium, Finland, Japan, Republic of Korea, Sweden, Switzerland, USA and France [18].
The French radioactive waste management agency (ANDRA) is
designing a deep geological repository in Callovo Oxfordian claystone (COx) formation in Eastern France. Current investigations
are conducted to optimize and nalize this repository concept with
the aim to ensure its long-term safety and its reversibility. The
high-level waste (HLW) shall be conned in a glass matrix and
then encapsulated into cylindrical carbon steel overpacks. These
containers shall be placed into micro-tunnels in the highly impermeable COx layer at a depth of 500 m [6,9,10]. Carbon steel containers must remain leakproof for a few thousand years as a
contribution to the multi-barrier containment system for the
waste [9,11,12]. The long-term safety assessment of the geological
repository has to take into account the degradation of the carbon
steel used for the waste overpacks, which is mainly caused by corrosion processes.
One of the most important features of any deep geological
repository is that the nature of the environment evolves over time.
Generally, the environment evolves in such a way that corrosion
becomes less severe and more predictable with time. However,
the presence of microorganisms can also affect the environment
[6,13]. In a recent study [6], we have shown the presence of
Corresponding author. Tel.: +33 2 51 85 86 35; fax: +33 2 51 85 84 52.
E-mail address: elmendil@subatech.in2p3.fr (Y. El Mendili).
http://dx.doi.org/10.1016/j.corsci.2014.07.020
0010-938X/ 2014 Elsevier Ltd. All rights reserved.

sulfate-reducing bacteria (SRB) in the COx claystone and sufcient

nutrients and trace elements for bacterial growth. SRB were isolated from argillite under anaerobic conditions and successfully
characterized and identied by 16sRNA sequencing. In addition,
H2 is expected to be produced in the disposal site and could be a
tremendous source of energy for the growth of anaerobic bacteria,
in particular the SRB [6].
Recently, El Mendili et al. [14] investigated the effects of SRB on
carbon steel corrosion at 30 C for 1, 3 and 6 months under anaerobic conditions in the presence of COx claystone and groundwater,
with an atmosphere of 5% H2/N2, simulating hydrogen production
in the site. We showed that the rst corrosion products formed on
steel surface in these conditions were poorly crystallized iron
sulde and mackinawite, followed then by a fast transformation
process into pyrrhotite phase. Finally, our investigations have
attempted to demonstrate that in the presence of hydrogen, the
corrosion of carbon steel containers under anoxic and suldogenic
environment sustained by SRB may not be a problem notably due
to the formation of a passive layer on the steel surface.
Another possible scenario of deep geological disposal conditions
was investigated [15], i.e., the carbon steel corrosion under
sequential aerobic and anaerobic microbiologically induced corrosion in the presence of COx claystone and groundwater with an
atmosphere of 5% H2/N2. The results showed the formation of mixture of lepidocrocite, maghemite and magnetite under aerobic conditions. Upon oxygen consumption and establishment of
suldogenic conditions, by SRB activity, all these oxides disappeared via transformation into iron sulde. It was also shown that
corrosion rate of steel in anaerobic cultures was higher than that of
steel initially corroded in aerobic condition, suggesting a protective

57

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

role of the corrosion product layer formed under sequential aerobicanaerobic conditions.
Although great care was taken to investigate the inuence of
major factors on carbon steel corrosion under geological conditions, the mechanism responsible is probably more complex.
Indeed, the corrosion of iron and steel surfaces in natural environments often lead to the formation of corrosion layers made of
extremely complex mixtures of iron-sulde, (hydro)oxide, carbonate, and chloride solids [16,17]. According to Anderko and Shuler
[18], magnetite and siderite are precursors to a less stable form
of iron sulde such as mackinawite. Mackinawite is a particular
form of iron sulde that is commonly observed in immersion corrosion studies and also easily produced from iron and iron oxides
by SRB activity [19]. Mackinawite can then transform into pyrrhotite which is considered to be more thermodynamically favored in
anoxic environments, hence pyrrhotite is expected to predominate
over long term conditions [18,2022].
In the present work, we investigate the carbon steel corrosion
under carbon dioxide-rich atmosphere in the presence of claystone. The morphology and the nature of corrosion products evolution were analyzed step by step using both scanning electron
microscopy (SEM) and confocal micro-Raman spectroscopy and
the role of SRB in the mechanism of steel corrosion was then
discussed.
2. Experimental section
2.1. Materials
The Callovo-Oxfordian (COx) claystone from the Bure site (east
of France) contains three main mineralogical phases: a dominant
argillaceous phases 4050% clay (5070% interstratied illite/
smectite), a carbonate phase (2230%, essentially calcite with a
few percent of dolomite) and a quartz phase (1832%) [23]. These
three main phases are accompanied by other minor phases including pyrite, feldspar, and organic matter, in mass quantities of less
than a few percent. The clay used for our experiments was sampled
under oxygen-free conditions, sealed in plastic/aluminium bags,
and transferred to the laboratory. Powder samples were prepared
for batch experiments in a glove box lled with nitrogen gas. A
simplied groundwater composition was calculated by BRGM
(French Geological Survey) and beaucaire et al. [24,25], from analyses of samples coming from the site of the Meuse/Haute-Marne
laboratory (France). The simulated groundwater was prepared in
the laboratory at 25 C with the following reagents: NaHCO3,
Na2SO4, Na2SiO39H2O, NaCl, KCl, CaCl22H2O, MgCl26H2O,
SrCl6H2O. All the reactants except metasilicate (Na2SiO39H2O)
were weighed out and grinded, then placed in a 2-liter ask containing 1.5 L of ultrapure water. The ask was closed and the solution is then brought into equilibrium (pH stabilization) under
magnetic stirring for 24 h until the reagents were completely dissolved. The solution was bubbled with a mixture of 89.5% Ar + 5%
H2 + 5.5% CO2 until the pH reached about 6.5. Metasilicate was
then added and the solution volume in the ask was adjusted to
2 L. After clarication, the solution was again degassed to pH 7.2.
The nal groundwater composition is detailed in Table 1. Lactate
(1 ppm) was added at the beginning of the experiments to simulate
the SRB activity as indicated by El Hajj et al. [11].
2.2. Steel preparation
The P235GH carbon steel was used because it has a number of
obvious benets as a canister material, including good combination of strength and ductility, good corrosion properties in the
expected repository environment, extensive experience with

Table 1
Synthetic water composition in equilibrium with the COx claystone (pH = 7.2).
Elements

Measured concentration
(mmol/L)

Theoretical concentration
(mmol/L) [25]

Na+
K+
Mg2+
Ca2+
Cl
SO2
4

44.9
1
6.4
7.3
42.1
15.5

45.6
1.03
6.67
7.36
41.0
15.6

fabricating and sealing large cylindrical objects, and the relative

abundance and cost of the material. The steel coupons were cut
into 10  10  0.5 mm and both surfaces were polished with a
Buehler polisher using a PSA discs (Pressure Sensitive Adhesive
discs) to a surface roughness of 3 lm. According to the ASTM standard method for corrosion measurement (ASTM, 1966), the steel
coupons were then cleaned with HCl (15%) and sodium carbonate
(5%) before experiments. For each experiment four steel coupons
were used, two coupons for weight loss measurements and two
for structural characterizations.
2.3. Experimental procedure
Each batch experiment was conducted with 100 ml of groundwater mixed with 20 g of claystone and 4 steel coupons in a serum
bottle. Before addition of the steel coupons to the groundwater and
the clay contained in the serum bottle, anaerobic conditions were
achieved rst by degassing the medium using a vacuum pump,
then placing the bottles in a glove box lled with nitrogen gas,
and nally by continuously bubbling a 3% Ar/CO2 gas at 0.5 bar
in the solution for 30 min. At this stage the groundwater/clay system is considered to be air-free but small traces of oxygen cannot
be excluded. Then, the steel coupons were added horizontally
inside the clay in the bottom of each bottle, and bottles were
sealed. Finally, the whole system (bottle containing clay, groundwater and steel) was bubbled with 1 bar CO2 at 30 C for 10 min
and the pressure was kept at 1 bar during the time exposure (1,
3 and 6 months). The nal addition of CO2 gas is meant to better
simulate the geochemical conditions in the COx claystone
formation.
2.4. Weight loss and corrosion rate
For all experiments, the steel corrosion rate was determined
according to the ASTM standard method for corrosion measurement. The corrosion products are stripped using the chemical
HCl (15%) and isopropanol with 5 mg/l of hexamethylenetetramine
to determine the mass loss.
The corrosion rate could be calculated using the following
formula:

Corrosion rate lm=y :

3650  weight lossmg

Densityg=cm3  Areacm2  Timedays

With a density of 7:85 g=cm3 for steel

2.5. Solution analysis
At the end of reaction time, an aliquot of the solution was syringe-ltered using a 0.2 lm polypropylene lter (Whatman). The
solution was then analyzed by Inductively Coupled Plasma-Mass
Spectrometry (ICP-MS, Perkin-Elmer) to quantitatively determine
the iron concentration. The microbial activity was monitored by
studying sulde production by sulfate reduction into sulde. Sulfate was measured by ion chromatography (DIONEX ICS 2500)

58

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

and sulde was measured by UVVIS spectroscopy (UV-2401 PC

Shimadzu) at 665 nm.

3. Results and discussion

3.1. Surface analysis

2.6. Surface analysis

The morphology and chemical composition of the corrosion
products were investigated by an Energy Dispersive Spectroscopy
(EDS) coupled to a SEM. The analyses were done under an accelerating voltage of 15 kV. The Si(Li) detector for EDS was equipped
with a beryllium window that allows to detect and quantify
oxygen.
The corrosion products have been characterized by Raman analysis. The confocal micro-Raman spectrometry is well suited for the
non-destructive analysis of very small and heterogeneous samples.
The Raman experiments were performed at room temperature
using a T64000 JobinYvon/Horiba spectrometer equipped with a
600 lines/mm diffraction grating and a nitrogen cooled CCD detector. The Raman spectra were recorded under a microscope (Olympus Bx41) in the backscattering geometry with a 100x objective
focusing the 514 nm line from an ArgonKrypton ion laser (coherent, Innova). The spot size of the laser was estimated at 0.8 lm and
the spectral resolution at 2 cm1. Single spectra were recorded
twice in the wavenumber 804000 cm1 region with an integration time of 600 s. Acquisition and basic treatments of spectra have
been made with the LabSpec V5.25 (Jobin Yvon-Horiba) and Origin
8.6 software. Raman measurements were carried out at very low
laser power to minimize possible sample deterioration or phase
transition with operating time. Indeed, in previous studies [26
28], we clearly evidenced that phase transitions can be induced
by local heating due to laser irradiation, especially with nanoscale
iron oxide nanoparticles. So, in this study, we chose to work with
an output laser power of 0.6 mW in order to prevent phase
transformations.
It should be noted that the coupons were transported in a container under Ar atmosphere before structural characterization by
micro-Raman spectroscopy to protect them from atmospheric contamination. After Raman analysis the same coupons were then
stored under Ar atmosphere waiting for SEM/EDS analysis.

3.1.1. The 1 month experiment

The corrosion products was localized by SEM micrographs and
identied by EDS microanalysis.
The SEM micrograph of the black deposit shows a surface completely covered by deposits of micrometer dimensions on a compact sub-layer (Fig. 1a and b). The EDS spectrum recorded on the
entire image clearly shows the presence of strong peaks of Fe
and O due to the presence of iron oxide (Fig. 1d). It is important
to note that the steel coupons turned all black, the black colour
suggests that this phase is likely magnetite. The presence of a small
peak of sulfur was also observed in the spectrum indicating a partial formation of iron sulde, which can be due to a SRB activity
from the clay autochthonous micro-organisms [6].
A representative Raman spectrum recorded on the corrosion
products formed on the steel surface after 1 month of exposure is
presented in Fig. 1c. All spectra obtained on the steel surface present the same vibrational bands observed at 195, 306, 450, 538 and
668 cm1 characteristics of magnetite [29]. The peaks at 195, 450
and 538 cm1 are assigned to T2g modes and the peaks at 306
and 668 cm1 to Eg and Ag modes, respectively. The lack of peak
at higher wavenumber (1378 cm1), conrms the presence of
Fe3O4 rather than c-Fe2O3 [30]. The absence of carbonate vibrational modes indicates that CO2 does not contribute to the total
corrosion products formation at this point. An important aspect
of the magnetite phase precipitation is that a protective layer of
corrosion products is then formed on the steel surface [14,31].
Indeed, the magnetite layer formed on steel surface increases resistance to corrosion, which is explained by a high thermodynamic
stability of magnetite and its compactness, adherence to the metal
surface and its small molar volume.
3.1.2. The 3 months experiment
Fig. 2 shows the SEM images, EDS and Raman spectra of the corrosion products formed on the steel surface after 3 months of time

Fig. 1. (a) SEM image of the steel surface after 1 month, (b) zoom of the SEM photograph with the corresponding chemical microanalysis spectrum and (d and c) Raman
spectrum of the corrosion products formed on the steel surface after 1 month of time exposure. The elemental composition is Fe: 41, S: 3, O: 53, C: 2 and Al: 1 (atomic %).

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

59

Fig. 2. (a) SEM photograph of the steel surface after 3 months, (b) zoom of the SEM photograph (a) with the corresponding chemical microanalysis SEM micrograph, 1, 2 and
3: spectra associated with carbonate, iron sulde and magnetite. The elemental composition is Fe: 38, Ca: 3, S: 7, O: 36 and C: 16 (atomic %).

exposure. Several solid phases with distinct backscattered electron

(BE) intensities can be observed on the SEM images, implying the
coexistence of several distinct phases. The combination of microRaman spectroscopy and SEM allows identication of the nature
of corrosion products more accurately. Analysis by EDS and
micro-Raman spectroscopy was performed in the area delimited
by the rectangle drawn on the SEM photograph (Fig. 2a).
The Raman spectrum collected on the point 1 is typical of a
rhomboedral carbonate phase. The Raman lines correspond to
the lattice modes at 169 and 294 cm1 corresponding to external
TO (Transversal Optical) and LO (Longitudinal Optical) phonon
modes, and to the internal modes of carbonate ions vibration at
725 (m4), 1095 (m1) and at 1450 (m3) cm1 [32]. The EDS microanalyses (Fig. 2c) revealed the presence of some elements originating
from COx groundwater, such as calcium at 15 mass% and Mg in
mass quantities of less than a few percent. The strong m1(CO3)
Raman lines in various carbonates are close to each other, and
the types of carbonates can be identied from the positions of their
respective Raman active lattice modes: the lattice modes of
dolomite CaMg(CO3)2 (174, 297 cm1), siderite FeCO3 (169,
294 cm1), and magnesium carbonate MgCO3 (212 and 329 cm1)
are widely separated [32]. The lattice vibration modes in our
spectrum correspond perfectly to those of siderite.
The Raman spectrum obtained for the black spot (zone 2) shows
the presence of Raman bands at 208, 228, 253, 286, 355 and
378 cm1. The peaks around 208 and 286 cm1 were attributed

to poorly crystallized iron sulde (PC) and are assigned to the

asymmetric and symmetric stretching mode of FeS, respectively
[33] and the other peaks around 228, 253, 355 and 378 cm1 are
attributed to mackinawite phase (Fe1+xS) [21]. The presence of iron
sulde on the surface implies minor chemical conversion of the
corrosion lm, suggesting that the corrosion rates could eventually
increase [34]. The formation of iron sulde probably results from
the SRB activity. The energy metabolism of these bacteria is due
to their ability to use lactate added to COx water and hydrogen
formed by steel dissolution as an energy source for sulfate
reduction.
The spectrum obtained in area 3 is typical of magnetite. From
the EDS analysis of the SEM images (analysis of the images contrasts), it is estimated that magnetite cover 70% of the steel surface. Most corrosion products contained only traces of sulde
compounds, which amount to less than 3%.

3.1.3. The 6 months experiment

An analysis by SEM and micro-Raman spectroscopy of a crosssection of the sample reacted during 6 months is shown in Fig. 3.
The steel was imbedded in an epoxy-resin (Epox of Struers) in
the glove box lled with nitrogen gas before polishing.
The SEM micrograph showing the chemical contrast of the analyzed zone (Fig. 3a), reveals the presence of different kinds of corrosion products with contrasted BE intensities (shades of grey on

60

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

the backscattered electron image, corresponding to different

chemical contrasts).
Raman spectroscopy has been used here because it allows to
discriminate easily the siderite FeCO3 and the chukanovite Fe2(OH)2CO3 phases. The band observed at 1072 cm1 is the signicant
m1-CO2
symmetric stretching mode for Fe2(OH)2CO3, which is
3
slightly shifted from the m1-CO2
of siderite at 1091 cm1 [35].
3

The band observed at 696 cm1 is assigned to m4 mode of CO2

3
and the 1425 cm1 band to the m3-CO2
3 anti-symmetric stretching
even if this mode seems to be affected by the hydration of the compound (redshift of about 20 cm1). The two lower bands at 211 and
256 cm1 can be assigned to the lattice mode [36]. In addition,
unlike with siderite, chukanovite has two strong bands around
3565 and 3559 cm1 due to the OH ions in the structure (Fig. 3).

Fig. 3. (a) SEM photograph of the steel surface after 6 months; (b) zoom of the SEM photograph; (c) the Raman spectra correspond to the corrosion products formed onto
surface: mackinawite, magnetite, carbonate (Ca, Mg)FeCO3, chukanovite and hydroxychloride b-Fe2(OH)3Cl.

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

61

We can notice that chukanovite is not present after 3 months.

This can be explained by the solubility of theses carbonates phases.
Indeed, siderite is thermodynamically more stable than chukanovite [37]. Fe2(OH)2CO3 is considered to be an intermediate, metastable species, leading to formation of FeCO3 via the following
multistep mechanism [3841]:

Fes H2 Ol ! FeOHaq Haq 2e

CO2g ! CO2aq

2FeOHaq

CO2
3aq

! Fe2 OH2 CO3s


Fe2 OH2 CO3s CO2
3aq ! 2FeCO3s 2OHaq

3
4

Thus, a slightly basic pH, resulting from e.g. water reduction at the
steel surface, would favor the chukanovite over siderite.
In addition to the described corrosion products, ferrous
hydroxychloride b-Fe2(OH)3Cl was also clearly identied by Raman
spectroscopy (Fig. 3). Our attribution is based on a similar Raman
spectrum reported by Rmazeilles et al., Neff et al. and Rguer
et al. on terrestrial ferrous archaeological objects [4244]. Their
spectra shows peaks around 127, 161, 322, 429, 615, 789 and
985 cm1, but no Raman band assignment was given by theses
authors. Accordingly, to reveal clearly the formation of ferrous
hydroxychloride phase, we have tested the transformation of ferrous hydroxychloride into Akaganeite (b-FeOOH containing chlorine) [42] in oxidizing conditions by exposing the steel coupons
to open air for 1 month. Indeed, Akaganeite the b phase of ferric
oxihydroxide b-FeO12x(OH)1+xClx, is often observed as a corrosion
product of iron in chloride containing environments, such as marine environments or in corroded archaeological iron artefacts
[45,46]. Akaganeite has a monoclinic crystal structure and always
contains some chloride ions. As shown by Keller [47], it can contain
up to 6% chloride. It is also preferentially formed compared to goethite and lepidocrocite in environments with high chloride and
Fe2+ concentration, by the oxidation of ferrous hydroxychloride
b-Fe2(OH)3Cl [48,49].
Fig. 4 shows the Raman spectrum obtained after oxidation of
ferrous hydroxychloride for 1 month. The Raman spectrum with
the bands around 308, 390, 535 and 720 cm1 is characteristic
for akaganeite [43,44,50] and our Raman spectrum is similar to
that reported by Neff et al. [43,44]. This corroborates the identication of the ferrous hydroxychloride as a corrosion product in our
experiments.
Finally, the Raman spectrum of black zone (Fig. 3) clearly identies the presence of mackinawite vibrational modes at 225 cm1
(B1g), 255 cm1 (Eg), 367 cm1 (A1g) and 398 cm1 (Eg) [51]. In
addition, the observation of band at 175 demonstrates the presence of elemental sulfur onto the corroded surface. Accordingly,
our Raman investigations reveal clearly the formation of poorly
crystallised FeS after 1 month of time exposure followed by the
crystallization of mackinawite after 6 months.
The corrosion of steel surface under CO2 pressure leads to the
formation of different corrosion products with extremely complex
mixtures of iron-oxide (magnetite), carbonates (siderite and chukanovite), sulde (mackinawite), and chloride (hydroxychloride).

Fig. 4. Raman spectrum of akaganeite formed on the steel surface after 1 month of
ferrous hydroxychloride oxidation.

the surface will give minimum iron transport and thus maintaining
low Fe concentration in solution.
The evolution of dissolved concentration of sulfate ([SO2
4 ]aq)
and sulde ([S2]aq) as a function of experiment time is presented
in Fig. 6. It is important to note that the concentration of sulfate at
the beginning of the runs (1325 mg/L) is higher than that expected
in COx water (1023 mg/l), which can be explained by the possible
slight oxidation of claystone (oxidation of pyrite) during transport
and storage before being used in the experiments.
Between 0 and 6 months, [SO2
4 ]aq decreased by about 45 mg/L
(0.47 mM) while sulde was produced. This indicates the SRB bioactivity in the rst step of the corrosion processes, which explains
the presence of sulde in the EDS spectrum (Fig. 1d). The high
2
decrease in [SO2
]aq, are
4 ]aq paralleled by the low increase of [S
the signature of the formation of iron sulde on the steel surface.
In fact, after 6 months the average measured mass loss for each
coupon is 75.5 mg and the expected total mass loss for the 4 coupons used in the experiment is 302 mg in 100 mL of solution.
Hence the maximum iron concentration provided by the steel corrosion is 3020 mg/L (54 mM). This indicates that only a minor fraction of oxidized Fe (less than 1%) is needed to co-precipitate the
entire amount of produced sulde. This is consistent with SEM
observations of the corroded steel showing that mackinawite is
only a minor corrosion product.
We did not measure oxygen concentrations in the systems but a
measurement of redox potential in the beginning of the experiments give values (+100 mV) indicating the presence of some
oxygen traces. However, at the end of the experiment, Eh measurements (200 mV) show reducing conditions, which is consistent
with the activity of sulfate-reducing bacteria. Hence at the

3.2. Solution analysis

The variations in concentrations of dissolved Fe ([Fe]aq) in synthetic groundwater as a function of time exposure is plotted in
Fig. 5. In the rst month, [Fe]aq is high (about 620 lg/l), probably
a consequence of a rapid steel corrosion and formation of iron
oxide on the steel surface. From 30 to 180 days, the continuous
decrease of [Fe]aq can be explained by the precipitation of phases
with a greater stability, i.e. a lower solubility and with protective
properties. This protective corrosion layer of iron compounds onto

Fig. 5. Evolution of iron concentration in synthetic groundwater as a function of

time exposure.

62

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

Fig. 6. Concentrations of sulfate and sulde as a function of incubation time.

Fig. 7. SEM photograph of the steel surface after 6 months of corrosion.

beginning of the experiment where oxygen traces persist, magnetite (Fe3O4) is expected to form as follow:

3FeOH2 1=2O2 ! Fe3 O4 3H2 O

With increasing time exposure and oxygen consumption the

reduction of water governs the corrosion process and produces
hydrogen and iron metal oxidizes, forming magnetite and resulting
in a net reaction of iron oxidation and water reduction:

3Fe0 4H2 O ! Fe3 O4 4H2

Hydrogen production leads then to establish reducing conditions propitious for SRB growth. In fact, SRB are anaerobic bacteria
that gain energy by oxidizing organic compounds such as lactate or
H2 formed by the steel corrosion. The SRB bacteria promote reduction of dissolved sulfate to H2S. SRB metabolize as follows:

4H2 SO2
4 2H ! 4H2 O H2 S

Ferrous sulde, formed by the reaction of biogenic sulde ions

with the metal surface, produces an adhesive layer. At the anodic
site the dissolved Fe2+ reacts stoichiometrically with H2S to form
iron sulde FeS [51]. The proposed reactions are as follows:
Formation of bisulde:

H2 Saq ! Haq HSaq

Bisulde dissociation:

HSaq ! Haq S2

Iron sulde formation:

Fe2 S2 ! FeS

3.3. pH variation
The variation results of solution pH as a function of time exposure is given in Fig. 8. The pH increases from 7 to 9 within
3 months. This increase of pH can due to the overall corrosion process in our system, according to the reaction:


4Fe SO2
4 3HCO3 5H ! FeS 3FeCO3 4H2 O

This reaction is proton consuming and leads to an increase of pH

even in a buffered solution particularly near the steel surface.
Between 90 and 180 days, the pH slightly decreases, possibly
due to the dissolution of clay solid phases, mostly silicate, which
can buffer the system. This hypothesis is suggested by the concentrations of the major components of the COx claystone (Na, Si, Ca)
increase in solution (Table 2). This increase is more pronounced
between the third and sixth month, reecting probably the dissolution of some mineral phases originating from claystone such as
illite and carbonate.
The alkaline pH indicates that all type of iron sulde corrosion
products could precipitate including mackinawite [14,21]. In addition, the formation of magnetite depends on the total concentration of dissolved Fe and was favoured at higher pH (magnetite
only forms at pH > 7.3).
Depending on environmental conditions, specically the CO2
pressure and dissolved anion concentrations in the soil or water,
carbonate phases, chukanovite Fe2(OH)2CO3, and siderite (FeCO3)
could also be intermediate corrosion products in addition to magnetite under neutral to slightly alkaline conditions [40,59].

10

The chemical reactions proposed above are the most widely

accepted reactions describing sulde chemistry in aqueous solution. Other authors [5256] have proposed different pathways for
iron sulde formation. An example is the direct reaction:

Fe2 H2 S ! FeS H2

11

When magnetite is formed in the rst step of steel corrosion, it

can transform into iron sulde at low temperature (25 C) in the
presence of excess of H2S provided by SRB [57,58]:

Fe3 O4 4H2 S ! 3FeS S 4H2 O

12

The presence of sulde can be seen mostly after 6 months of

time exposure in Fig. 7. Additional SEM observations on cross section (see Fig. 3) did not show localized corrosion (e.g. pitting) but
only a general corrosion trend.

13

Fig. 8. Evolution of solution pH as a function of time exposure.

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

63

Table 2
Concentrations of Cl, Ca2+, Na+ and Si4+ in the COx groundwater as function of time
experiment.
Elements (mg/l)

First day

30 days

60 days

90 days

180 days

[Cl]
[Ca2+]
[Na+]
[Si4+]

1353
318
887
2.1

1335
324
891
3.2

1310
321
894
5.0

1299
326
900
6.9

1270
364
922
12.0

The reactions of the iron carbonate formation on carbon steel

can be described as anodic (oxidation) and cathodic (reduction)
reactions.
CO2 react in aqueous solution to give carbonic acid, H2CO3, as
follows [60]:

CO2 H2 O ! H2 CO3

14

This weak and partly dissociated acid is responsible for high

corrosion rates of steel in CO2-containing aqueous solutions. The
electrochemistry of CO2 corrosion is still not certain although a
number of detailed studies have been reported [6166]. One of
the simplest assumptions is that the dominant cathodic reaction
is the reduction of hydrogen ions, where the hydrogen ions are
supplied by dissociation of carbonic acid:

H2 CO3 ! H HCO3

15

HCO3 ! H CO2
3

16

The other possibility is the direct reduction of carbonic acid:

H e ! 1=2H2 e provided by iron oxidative dissolution

H2 CO3 e ! 1=2H2 CO2
3

17

FeCO3 forms on the surface of carbon steels according to [67]:

Fe

CO2
3 ! FeCO3

18

or

Fe2 HCO3 ! FeCO3 H

19

Carbon dioxide, bicarbonate, and carbonate are all just different

species of inorganic carbon in water and which species predominates is determined primarily by pH. Dissolved CO2 predominates
2
at low pH, HCO
3 at near neutral pH (6.4 < pH < 10.3) [59], and CO3
at high pH.
The formation of the iron (II) hydroxychloride b-Fe2(OH)3Cl can
be described by the following reaction:


2FeCO3 H 3H2 O Cl ! b  Fe2 OH3 Cl 2H2 CO3

20

The phase b-Fe2(OH)3Cl may also directly precipitate from the

solution on the surface according to:


2Fe2 3H2 O Cl ! b  Fe2 OH3 Cls 3H

21

The formation of b-Fe2(OH)3Cl can explain the slight decrease in

the Cl concentration (Table 2).
3.4. Corrosion rate
Corrosion rate determined from weight loss are plotted in Fig. 9.
We reported data from similar experiments conducted under H2
atmosphere where iron suldes were the main corrosion products
[14]. The evolution of the corrosion rate is directly attributed to the
nature of corrosion products. In both cases, a clear increase of the
corrosion rate can be observed between 0 and 3 months. This can
be explained by the replacement of the protective Fe-oxide (magnetite) lm by non-protective porous lms of carbonate or iron sulde, which are then responsible for increasing the corrosion rate.

Fig. 9. Evolution of the steel corrosion rate as a function of time exposure and
under different atmospheres. Under CO2: this study and under H2: El Mendili et al.
[14].

Schlegel et al. [68] studied the corrosion of steel under anaerobic conditions in the presence of COx claystone reacted at 90 C
and 50 bar (pH = 6.7 and PCO2 = 0.43 atm) for 8 months. Their
results show the presence of a corrosion layer and a clay transformation layer. The corrosion layer is made of a magnetite-containing internal sub-layer and a Fe-phyllosilicate external sub-layer
enriched in Na, with traces of goethite. The clay transformation
layer is made of predominantly Ca-rich siderite (FeCO3). Silica
released upon clay dissolution diffused into the corrosion layer
and co-precipitated with oxidized Fe to form Fe-phyllosilicate.
The absence of iron sulde is probably due to the high temperature
conditions which prevent bacterial activity. It is important to also
notice that in our study conducted at 30 C, Fe-phyllosilicates are
not detected within the corrosion products after 8 months of experience. The absence of silicates is related to a problem of dissolution kinetics of silicate, very low compared to that of the
carbonate at 30 C [69].
Under only H2 atmosphere, the main corrosion product is a
continuous iron sulde layer [14] which led to a drastic drop in
corrosion rate with time due to the formation of well crystallized
iron sulde phases (pyrrhotite). Under disposal conditions sulde
production may be promoted by SRB activity due to hydrogen
production at low temperature (3050 C) and perhaps by COx
claystone pyrite reduction by hydrogen at higher temperature
(90 C).
While this experiment was too short to determine what rates
are achievable after long-term experiment of corroded steel under
anaerobic conditions and CO2 atmosphere, it demonstrates the formation of corrosion layers made of extremely complex mixtures of
iron-sulde, oxide, carbonate, and chloride solids. Formation of
iron sulde is not expected to increase the corrosion but could play
a protective role via the formation of highly stable iron sulde minerals including pyrrhotite and pyrite as indicated by El Mendili
et al. [14].

64

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

4. Conclusion
Carbon steel corrosion in simulated geological conditions (clay
and groundwater) results in a variety of corrosion products of
varying proportions with exposure time. Hence, at short time
exposure magnetite was identied as the main corrosion product
while at longer times carbonate (siderite and chukanovite), mackinawite and ultimately iron hydroxychloride are formed. These
structural modications are due to changes of localized geochemical conditions with time exposure including hydrogen and sulde
production due to steel corrosion and SRB activity, respectively.
The present work shows that it is important to take into account
all the components of a disposal site including the atmosphere
composition to intent to explain the corrosion mechanism. A lack
of hydrogen would lead to a mineralogical paragenesis dominated
by iron carbonate while under sufcient hydrogen supply, sulde
production via SRB stimulation by hydrogen and possibly via pyrite
reduction by hydrogen would lead to iron sulde formation with
ultimately high passivation properties.

Acknowledgment
Special thanks to the Institut des Matriaux de Nantes (CNRS/
universit de Nantes) for the SEM-EDX measurements.

References
[1] S.S. Kim, K.S. Chun, C.H. Kang, J.W. Choi, K.W. Han, Estimated Corrosion
Thickness of Candidate Materials for High-Level Waste Disposal Container in a
Disposal Condition, KAERI/TR-2172/02, Korea Atomic Energy Research
Institute, KAERI, 2002.
[2] C.H. Kang, J.W. Kim, K.S. Chun, J.H. Park, W.J. Cho, J.W. Choi, J.W. Lee, Y.M. Lee,
High-Level Radwaste Disposal Technology Development: Geological Disposal
System Development, KAERI/RR-2336/02, Korea Atomic Energy Research
Institute, KAERI, 2002.
[3] Nuclear Energy Agency, Organization for Economic Cooperation and
Development, Partnering for Long-term Management of Radioactive Wastes:
Evolution and Current Practice in Thirteen Countries, NEA No. 6823, 2010.
[4] O. Bildstein, L. Trotignon, M. Perronnet, M. Jullien, Modelling iron-clay
interactions in deep geological disposal conditions, Phys. Chem. Earth 31
(2006) 618625.
[5] D. Feron, D.D. Macdonald, Prediction of long term corrosion behaviour in
nuclear waste systems, Mat. Res. Soc. Symp. Proc. 932 (2006) 785795.
[6] H. El Hajj, A. Abdelouas, B. Grambow, C. Martin, M. Dion, Microbial corrosion of
P235GH steel under geological conditions, Phys. Chem. Earth 35 (2010) 248
253.
[7] Nuclear Energy Agency of the Organisation for Economic Co-operation and
Development, Safety Cases for Deep Geological Disposal of Radioactive Waste:
Where Do We Stand?, Symposium Proceedings Paris, France, 2007.
[8] B. Kursten, F. Druyts, P. Van Iseghem, Geological disposal of conditioned highlevel and long lived radioactive waste, Studies on container materials
laboratory corrosion tests on stainless steel, Belgian Nuclear Research Centre
Report, SCKCEN-R-3774, 2003.
[9] C. Tournassat, C. Lerouge, P. Blanc, J. Brendl, J.-M. Grenche, S. Touzelet, E.C.
Gaucher, Cation exchanged Fe(II) and Sr compared to other divalent cations
(Ca, Mg) in the Bure Callovian-Oxfordian formation. Implications for porewater
composition modelling, Appl. Geochem. 23 (2008) 641654.
[10] J. Morel, F. Bumbieler, N. Conil, G. Armand, Feasibility and behavior of a full
scale disposal cell in a deep clay layer, Rock Mechanics for Resources, Energy
and Environment, in: Kwasniewski & Lydzba (eds), 2013, Taylor & Francis
Group, London.
[11] H. El Hajj, A. Abdelouas, B. Grambow, Role of microbial activity on corrosion of
steel used for radioactive waste disposal, Clays in Natural & Engineered
Barriers for Radioactive Waste Connement, 4th International Meeting,
Nantes, France, 2010.
[12] P. Van Iseghem, B. Kursten, E. Valcke, H. Serra, J. Fays, A. Sneyers. In situ testing
of waste forms and container materials: contribution to the identication of
their long term behaviour, in: D. Fron, D.D. Macdonald (Eds.), Proceedings of
the International Workshop on Prediction of Long Term Corrosion Behaviour
in Nuclear Systems, Cadarache, France, November 2629, 2001, EFC
Publications n 36, 2002.
[13] R.M. Maier, I.L. Pepper, C.P. Gerba, Environmental microbiology, second ed.,
Elsevier, 2009, pp. 387439, doi: B978-0-12-370519-8.00020-1.
[14] Y. El Mendili, A. Abdelouas, J.-F. Bardeau, Insight into the mechanism of carbon
steel corrosion under aerobic and anaerobic conditions, Phys. Chem. Chem.
Phys. 15 (2013) 91979204.

[15] H. el Hajj, A. Abdelouas, Y. El Mendili, G. Karakurt, B. Grambow, C. Martin,

Corrosion of carbon steel under sequential aerobicanaerobic environmental
conditions, Corros. Sci. 76 (2013) 432440.
[16] R. Jeffery, R.E. Melchers, Bacteriological inuence in the development of iron
sulde species in marine immersion environments, Corros. Sci. 45 (2003) 693
714.
[17] D.H. Phillips, D.B. Watson, Y. Roh, B. Gu, Mineralogical characteristics and
transformations during long-term operation of a zerovalent iron reactive
barrier, J. Environ. Qual. 32 (2003) 20332045.
[18] A. Anderko, P.J. Shuler, A computational approach to predicting the formation
of iron sulde species using stability diagrams, Comput. Geosci. 23 (1997)
647658.
[19] W.A. Hamilton, Biocorrosion: the action of sulfate-reducing bacteria, in: C.
Ratledge (Ed.), Biochemistry of Microbial Degradation, Kluwer Academic,
Dordrecht, 1994, pp. 555570.
[20] M.B. McNeil, B.J. Little, Mackinawite formation during microbial corrosion,
Corrosion 46 (1990) 599600.
[21] Y. El Mendili, A. Abdelouas, J.-F. Bardeau, Impact of sulphidogenic environment
on the corrosion behavior of carbon steel at 90 C, RSC. Adv. 3 (2013) 15148
15156.
[22] Y. El Mendili, A. Abdelouas, J.-F. Bardeau, Corrosion of P235GH carbon steel in
simulated Bure soil solution, J. Mater. Environ. Sci. 4 (2013) 786791.
[23] H.-R. Wenk, M. Voltolini, M. Mazurek, L.R. Van Loon, A. Vinsot, Preferred
orientations and anisotropy in shales: Callovo-Oxfordian shale (France) and
Opalinus Clay (Switzerland), Clays Clay Miner. 56 (2008) 285306.
[24] P. Vieillard, S. Ramirez, A. Bouchet, A. Cassagnabre, A. Meunier, E. Jacquot,
Alteration of Callovo-Oxfordian clay from Meuse-Haute Marne underground
laboratory (France) by alkaline solutions: II. Modelling of mineral reactions at
120 C and at hyperalkaline solutions, Appl. Geochem. 19 (2004) 16991709.
[25] C. Beaucaire, E. Tertre, E. Ferrage, B. Grenut, S. Pronier, B. Made, A
thermodynamic model for the prediction of pore water composition of
clayey rock at 25 and 80C - Comparison with results from hydrothermal
alteration experiments, Chem. Geol. 334 (2012) 6276.
[26] Y. El Mendili, J.-F. Bardeau, N. Randrianantoandro, A. Gourbil, J.-M. Grenche,
A.-M. Mercier, F. Grasset, New evidences of in situ laser irradiation effects on
c-Fe2O3 nanoparticles: a Raman spectroscopic study, J. Raman. Spectrosc. 42
(2011) 239242.
[27] Y. El Mendili, J.-F. Bardeau, N. Randrianantoandro, F. Grasset, J.-M. Grenche,
Insights into the mechanism related to the phase transition from c-Fe2O3 to aFe2O3 nanoparticles induced by thermal treatment and laser irradiation, J.
Phys. Chem. C 116 (2012) 2378523792.
[28] T. Beuvier, J.-F. Bardeau, B. Calvignac, G. Corbel, F. Hindre, J.-M. Greneche, F.
Boury, A. Gibaud, Phase transformation in CaCO3/iron oxide composite
induced by thermal treatment and laser irradiation, J. Raman Spectrosc. 44
(2013) 489495.
[29] O.N. Shebanova, P. Lazor, Raman study of magnetite (Fe3O4): laser induced
thermal effects and oxidation, J. Solid. Stat. Chem. 174 (2003) 424430.
[30] D.L.A. De Faria, S.V. Silva, M.T. de Oliveira, Raman micro spectroscopy of some
iron oxides and oxyhydroxides, J. Raman Spectrosc. 28 (1997) 873878.
[31] Ph. Dillmann, F. Mazaudier, S. Hoerle, Advances in understanding atmospheric
corrosion of iron. I. Rust characterisation of ancient ferrous artefacts exposed
to indoor atmospheric corrosion, Corros. Sci. 46 (2004) 14011429.
[32] S.K. Sharma, A.K. Misra, S.M. Clegg, J.E. Bareeld, R.C. Wiens, T.E. Acosta, D.E.
Bates, Remote-Raman spectroscopic study of minerals under supercritical CO2
relevant to Venus exploration, Spectrochim. Acta. Part. A 80 (2011) 75
81.
[33] E.B. Hansson, M.S. Odziemkowski, R.W. Gillham, Formation of poorly
crystalline iron monosuldes: surface redox reactions on high purity iron,
spectroelectrochemical studies, Corros. Sci. 48 (2006) 37673783.
[34] B.W.A. Sherar, P.G. Keech, J.J. Nol, R.G. Worthingham, D.W. Shoesmith, Effect
of sulde on carbon steel corrosion in anaerobic near-neutral pH saline
solutions, Corrosion 69 (2013) 6776.
[35] E. Chan, D. John, S. Bailey, B. Kinsella, S. Nesic, Formation of magnetite scale
under anaerobic conditions at low temperatures, Paper presented at the
Australasian Corrosion Association, Coffs Harbour, 2009.
[36] R.L. Frost, A Raman spectroscopic study of selected minerals of the rosasite
group, J. Raman Spectrosc. 37 (2006) 910921.
[37] I. Azoulay, C. Remazeilles, Ph. Refait, Determination of standard Gibbs free
energy of formation of chukanovite and Pourbaix diagrams of iron in
carbonated media, Corros. Sci. 58 (2012) 229236.
[38] T.R. Lee, R.T. Wilkin, Iron hydroxy carbonate formation in zerovalent iron
permeable reactive barriers: characterization and evaluation of phase stability,
J. Contam. Hydrol. 116 (2010) 4757.
[39] R. De Marco, Z.-T. Jiang, B. Pejcic, E. Poinen, An in situ synchrotron radiation
grazing incidence X-ray diffraction study of carbon dioxide corrosion, J.
Electrochem. Soc. 152 (2005) B389B392.
[40] C. Rmazeilles, Ph. Refait, Fe(II) hydroxycarbonate Fe2(OH)2CO3 (chukanovite)
as iron corrosion product: synthesis and study by Fourier Transform Infrared
Spectroscopy, Polyhedron 28 (2009) 749756.
[41] M.L. Schlegel, C. Bataillon, C. Blanc, D. Prt, E. Foy, Anodic activation of iron
corrosion in clay media under water-saturated conditions at 90 C:
characterization of the corrosion interface, Environ. Sci. Tech. 44 (2010)
15031508.
[42] C. Rmazeilles, D. Neff, F. Kergourlay, E. Foy, E. Conforto, E. Guilminot, S.
Reguer, P. Refait, P. Dillmann, Mechanism of long-term anaerobic corrosion of
iron archaeological artefacts in seawater, Corros. Sci. 51 (2009) 29322941.

Y. El Mendili et al. / Corrosion Science 88 (2014) 5665

[43] D. Neff, P. Dillmann, L. Bellot-Gurlet, G. Beranger, Corrosion of iron
archaeological artefacts in soil: characterisation of the corrosion system,
Corros. Sci. 47 (2005) 515535.
[44] S. Rguer, P. Dillmann, F. Mirambet, Buried iron archaeological artefacts:
corrosion mechanisms related to the presence of Cl-containing phases, Corros.
Sci. 49 (2007) 27262744.
[45] F.R. Prez, C.A. Barrero, K.E. Garca, Factors affecting the amount of corroded
iron converted into adherent rust in steels submitted to immersion tests,
Corros. Sci. 52 (2010) 25822591.
[46] J. Monnier, D. Neff, S. Rguer, P. Dillmann, L. Bellot-Gurlet, E. Leroy, E. Foy, L.
Legrand, I. Guillot, A corrosion study of the ferrous medieval reinforcement of
the Amiens cathedral. Phase characterisation and localisation by various
microprobes techniques, Corros. Sci. 52 (2010) 695710.
[47] P. Keller, Vorkommen, Entstehung und Phasenumwandlung von b-FeOOH in
Rost, Werkst. Korros. 20 (1969) 102108.
[48] P. Refait, J.M.R. Genin, The mechanisms of oxidation of ferrous
hydroxychloride b-Fe2(OH)3Cl in aqueous solution: the formation of
akaganeite vs goethite, Corros. Sci. 39 (1997) 539553.
[49] C. Remazeilles, P. Refait, On the formation of b-FeOOH (akaganite) in
chloride-containing environments, Corros. Sci. 49 (2007) 844857.
[50] R. Autengruber, G. Luckeneder, A.W. Hassel, Corrosion of press-hardened
galvanized steel, Corros. Sci. 63 (2012) 1219.
[51] Y. El Mendili, B. Minisini, A. Abdelouas, J.-F. Bardeau, Assignment of Ramanactive vibrational modes of tetragonal mackinawite: Raman investigations and
ab-initio calculations, RSC. Adv. 4 (2014) 2582725834.
[52] W. Sun, Kinetics of iron carbonate and iron sulde scale formation in CO2/H2 S
corrosion, Ph.D. dlss Ohio University, Chemical and Biomolecular Engineering
Department, 2006.
[53] W. Sun, S. Nesic, D. Young, R.C. Woollam, Equilibrium expressions related to
the solubility of the sour corrosion product mackinawite, Ind. Eng. Chem. Res.
47 (2008) 17381742.
[54] D. Rickard, G.W. Luther, Chemistry of iron sulphides, Chem. Rev. 107 (2007)
514562.
[55] J. Vera, S. Kapusta, N. Hackerman, Localized corrosion of iron in alkaline sulde
solutions iron sulde formation and the breakdown of passivity, J.
Electrochem. Soc. 133 (1986) 461467.
[56] A.R. Lennie, D.J. Vaughan, in: M.D. Dyar, C. McCammon, M.W. Schaefer (Eds.),
Mineral Spectroscopy: A Tribute to Roger G. Burns, The Geochemical Society,
Houston, 1996, p. 117.

65

[57] M. Robert, J. Berthelin, Role of biological and biochemical factors in soil

mineral weathering, in: P.M. Huaing, M. Schnitzer (Eds.), Interaction of Soil
Minerals with Natural Organics and Microbes, Soil Sci. Soc. Am. Special
Publication 17, Soil Science Society of America, Madison, 1986, pp. 453496.
[58] J. Ning, Y. Zheng, D. Young, B. Brown, S. Nesic, A thermodynamic study of
hydrogen sulde corrosion of mild stee, NACE International Corrosion
Conference & Expo, Paper No. 2462, 2013.
[59] P. Refait, M. Abdelmoula, J.-M.R. Genin, R. Sabot, Green rusts in
electrochemical and microbially inuenced corrosion of steel, C.R. Geosci.
338 (2006) 476487.
[60] S. Nesic, J. Postlehwaite, S. Olsen, An electrochemical model for prediction of
corrosion of mild steel in aqueous carbon dioxide solutions, Corrosion 52
(1996) 280294.
[61] C. De Waard, D.E. Milliams, Carbonic acid corrosion of steel, Corrosion 31
(1975) 177181.
[62] G. Schmitt, B. Rothman, Studies of the corrosion mechanism of unalloyed
steels in oxygen-free carbon dioxide solutions. Part I. Kinetics of the liberation
of hydrogen, Werkst. Korros. 28 (1997) 816822.
[63] T. Hurlen, S. Gunvaldsen, R. Tunold, F. Blaker, P.G. Lunde, Effects of carbon
dioxide on reactions at iron electrodes in aqueous salt solutions, J. Electroanal.
Chem. 180 (1984) 511526.
[64] L.G.S. Gray, B.G. Anderson, M.J. Danysh, P.G. Tremaine, Mechanism of carbon
steel corrosion in brines containing dissolved carbon dioxide at pH 4,
corrosion/89, paper no. 464, Houston, TX: NACE International, 1989.
[65] M.R. Bonis, J.L. Crolet, Basics of the prediction of the risks of CO2 corrosion in
oil and gas wells, CORROSION/89, paper no. 466, Houston, TX: NACE
International, 1989.
[66] J.L. Mora-Mendoza, S. Turgoose, Fe3C inuence on the corrosion rate of mild
steel in aqueous CO2 systems under turbulent ow, Corros. Sci. 44 (2002)
12231246.
[67] D.A. Lopez, W.H. Schreiner, S.R. De Sanchez, S.N. Simison, The inuence of
carbon steel microstructure on corrosion layers: an XPS and SEM
characterization, Appl. Surf. Sci. 207 (2003) 6985.
[68] M.L. Schlegel, C. Bataillon, K. Benhamida, C. Blanc, D. Menut, J.L. Lacour, Metal
corrosion and argillite transformation at the water-saturated hightemperature iron-clay interface. A microscopic-scale study, Appl. Geochem.
23 (2008) 26192633.
[69] J.L. Palandri, Y.K. Kharaka, A compilation of rate parameters of water-mineral
interaction kinetics for application to geochemical modeling, U.S. Geological
survey, Open le report, 20041068.