Journal of Natural Gas Science and Engineering 59 (2018) 287–296

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

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

Influence of calcium and magnesium ions on CO2 corrosion of carbon steel T

in oil and gas production systems - A review
H. Mansoori∗, D. Young, B. Brown, M. Singer
Institute for Corrosion & Multiphase Technology, Department of Chemical & Biomolecular Engineering, Ohio University, Athens, OH, 45701, USA

A R T I C LE I N FO A B S T R A C T

Keywords: Calcium and magnesium ions are major constituent species in produced brines associated with oil and gas
CO2 corrosion production. Existing literature is often contradictory and unclear as to whether the presence of such ions ac-
Scale celerates or retards general corrosion. Furthermore, their influence on initiation and propagation of localized
Corrosion product corrosion remains ambiguous. This review presents state-of-the-art knowledge concerning how Ca2+ and Mg2+
CaCO3
influence the CO2 corrosion mechanism. In addition, a best way forward is proposed in order to bridge the
Isostructurality
Mixed metal carbonates
existing literature gaps in studying the influence of these alkaline earth cations on CO2 corrosion.

1. Introduction the gas composition and related water chemistry. Different brines with
complex chemistries are encountered in oil and gas production en-
Oil and gas production involves the transmission of large quantities vironments. However, most laboratory corrosion studies have been
of fluid – hydrocarbons, water, CO2 (carbon dioxide), H2S (hydrogen performed only in various dilute NaCl (sodium chloride) electrolytes
sulfide) gases – in carbon steel pipelines. In this environment, corrosion while, in reality, many other ions are present in the produced brine.
and scaling happen simultaneously (Mansoori et al., 2017). The aqu- Divalent cations that form various carbonate layers have the potential
eous phase, also called brine, is a by-product of almost every hydro- to play a major role in the corrosion behavior of mild steel, especially in
carbon production system, containing significant concentrations of downhole conditions of oil and gas wells at high temperature and high
dissolved cations such as Ca2+ (calcium ion) and Mg2+ (magnesium pressure. Among these cations, Ca2+ and Mg2+ are usually present at
ion) as well as dominant Na+ (sodium ion) and Cl− (chloride ion) the highest concentration in reservoir fluids. Table 1 shows an example
(Oddo and Tomson, 1994), (Mansoori et al., 2013a). Depending on the of these high ion concentrations from wells in Western Pennsylvania
conditions, scales in the form of CaCO3 (calcium carbonate), MgCO3 (Dresel and Rose, 2010). In terms of corrosion products, it is widely
(magnesium carbonate) or a mixture of both can precipitate. These accepted that FeCO3 can form a protective layer and plays a significant
scales are different from other precipitates such as FeCO3 (iron carbo- role in CO2 corrosion of mild steel (Kermani and Morshed, 2003). The
nate) and FeS (iron sulfide) where the constituent cation, Fe2+ (ferrous presence of Ca2+ and Mg2+ can change the composition of the corro-
ion), comes from the corroding surface steel. In this manuscript, FeCO3 sion product layer since their carbonate crystal structures, calcite (one
and FeS are defined as a corrosion product since the constituent cation of the mineral names for CaCO3) and magnesite (the mineral name for
comes from the corrosion process, where CaCO3 and MgCO3) are de- MgCO3), are isomorphous with siderite (the mineral name for FeCO3)
fined as a scale since the constituent cation comes from the bulk solu- possessing a hexagonal unit cell (Chai and Navrotsky, 2015), (Zoltai
tion. Distinct bodies of research have been conducted that separately and Stout, 1984). However, the effect of Ca2+/Mg2+-containing solu-
address scaling and corrosion in the oil and gas industry. However, tions on the properties of the FeCO3 layer and subsequent effect on CO2
there is minimal information in the literature relating to situations corrosion mechanisms are insufficiently documented. Because of this
where corrosion and scaling simultaneously occur, despite there being lack of knowledge, current mechanistic models do not account for the
indications that heavy scaling may lead to localized attack and loss of presence of Ca2+ and Mg2+ in their corrosion predictions. In most
facility integrity (Mansoori et al., 2017), (Salman et al., 2007). Conse- corrosion prediction models, Ca2+ and Mg2+ are only considered re-
quently, there is a need to explore the potential relationships between garding the ionic strength of the aqueous solution (Nordsveen et al.,
scaling and encountered corrosion phenomena. 2003). Therefore, it is critical to extend the capability of predictive
CO2 corrosion and scaling processes are both highly dependent on models to bridge the gap in knowledge associated with the presence of

∗
Corresponding author.
E-mail address: hm419213@ohio.edu (H. Mansoori).

https://doi.org/10.1016/j.jngse.2018.08.025
Received 26 February 2018; Received in revised form 5 May 2018; Accepted 25 August 2018
Available online 05 September 2018
1875-5100/ © 2018 Elsevier B.V. All rights reserved.
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

Table 1
Example of ion concentrations from an oilfield brine.
TDS (g/L) Conductivity (μS/cm) Na+ (mg/L) K+ (mg/L) Mg2+(mg/L) Ca2+(mg/L) Sr2+(mg/L) Ba2+(mg/L) Cl− (mg/L)

213 190,000 56700 190 2520 18000 691 171 122000

Ca2+ and Mg2+. The main goals of this paper are to review the relevant Fe (s) → Fe(2aq+) + 2e− (6)
literature and knowledge about the effect of such ions on CO2 corrosion
mechanisms and to provide a pathway for conducting future studies on the possible cathodic reactions in CO2 environments in the absence of
this topic based on the shortcomings in the existing scientific literature. oxygen are listed below:
2H(+aq) + 2e− → H2(g ) (7)
2. CO2 corrosion and relevant corrosion products/scales
−
2H2 CO3(aq) + 2e− → H2(g ) + 2HCO3(aq) (8)
This section presents the fundamentals of CO2 corrosion, including a
− − 2−
brief review of the chemical and electrochemical reactions involved in 2HCO3(aq) + 2e → H2(g ) + 2CO3(aq) (9)
CO2 corrosion of mild steel. In addition, relevant corrosion products
and scales encountered in CO2 environments are discussed. 2H2 O (l) + 2e− → H2(g ) + 2OH(−aq) (10)
Past research mainly focused on the cathodic reactions to describe
2.1. CO2 corrosion of carbon steel in brief and model observed corrosion mechanisms (Nyborg, 2010). First, de
Waard and Milliams proposed that the cathodic reaction in the form of
Despite its vulnerability to CO2 corrosion, carbon steel is the most Equation (8), the direct reduction of carbonic acid, would be dominant
used material for pipelines in the oil and gas industry, considering its if the solution pH is greater than 4 (De Waard and Milliams, 1975).
cost-effectiveness. CO2 corrosion, also known as sweet corrosion, is one Ogundele and White also proposed that the direct reduction of bi-
of the major concerns in the oil and gas industry (Mansoori et al., carbonate ion, Equation (9), is important when the pH is higher than 5
2013b), (Graver et al., 1984). CO2 corrosion of carbon steel represents a (Ogundele and White, 1987). Nesic et al. later proposed that hydrogen
phenomenon occurring when CO2 dissolves in water and hydrates to ion reduction expressed by Equation (7) is the dominant reaction when
become carbonic acid (H2CO3), which influences solution pH and the solution pH is less than 4 (Nesic et al., 1996). However, recent work
provides cathodic species for reaction with the metal pipe surface. CO2 by Tran et al., has cast some doubts on the previous understanding of
corrosion has been investigated for more than four decades with the the role of different cathodic reactions in CO2 corrosion (Tran et al.,
purpose of understanding its mechanisms and preventing or mitigating 2015). Carbonic acid is a weak acid; it partially dissociates and serves as
metal degradation. Modeling of the CO2 corrosion mechanism began a hydrogen ion reservoir for the so-called “buffering effect” governing
with de Waard and Milliams in 1975 (De Waard and Milliams, 1975) CO2 corrosion. The authors demonstrated that for pH values below 6, at
and continues to this day with fully mechanistic models (Institute for moderate partial pressures of CO2 (less than 10 bar) and moderate
Corrosion and Multiphase Technology, 2014), (Nyborg, 2002). temperatures (less than 80 °C), the dominant cathodic reaction is gov-
CO2 corrosion of carbon steel is an electrochemical process that erned by hydrogen ion reduction, Equation (7), where the dissociation
involves the anodic dissolution of iron and the cathodic evolution of of carbonic acid, Equation (4), and dissociation of bicarbonate, Equa-
hydrogen. The overall reaction can be expressed as: tion (5), provide H+ through a buffering mechanism. This demonstrates
Fe (s) + CO2(aq) + H2 O (l) → Fe(2aq+) + CO3(2 −aq) + H2(g ) (1) that the mechanisms related to direct reduction reactions by carbonic
acid and bicarbonate were negligible in their test conditions. To com-
A number of chemical, electrochemical, and transport processes plete the list of cathodic reactions in sweet conditions, it can also be
occur simultaneously in aqueous CO2 corrosion of mild steel. The first observed that the water reduction in Equation (10) plays an important
step is the dissolution of gaseous carbon dioxide in water, as described role when the pH is higher than 6 for low partial pressures of CO2
in Equation (2). The second step is the hydration of aqueous CO2, which (Nesic et al., 1996).
results in the formation of carbonic acid (H2CO3) as shown in Equation
(3). This process is relatively slow and is often the rate-determining step 2.1.2. FeCO3
in CO2 corrosion (Kahyarian et al., 2017). The sequence of reactions FeCO3 is the main corrosion product from mild steel in CO2 corro-
associated with the presence of CO2 in H2O is as follows: sion environments and has been proven to be protective under certain
CO2(g ) ⇌ CO2(aq) (2) conditions (Nesic, 2007). FeCO3 is considered as a corrosion product
rather than scale since its constituent cations (Fe2+) are derived from
CO2(aq) + H2 O (l) ⇌ H2 CO3(aq) (3) the corroding surface steel. The corrosion protection conferred by
FeCO3 is highly dependent on its growth rate, if the rate of corrosion is
H2 CO3(aq) ⇌ H(+aq) + HCO3(
−
aq) (4) higher than FeCO3 formation then it will not be protective; the ratio of
−
HCO3( ⇌ H(+aq) + CO3(2 −aq) precipitation rate to corrosion rate, labeled scaling tendency, can be
aq) (5)
used to qualify the layer protectiveness (Gao et al., 2011). FeCO3 is
formed from ferrous ions in solution that originates from corrosion and
2.1.1. Electrochemistry of mild steel corrosion in CO2 environments carbonate ions that are present due to CO2 dissolution in water.
The anodic reaction for oxidative iron dissolution is a key element in Equation (11) describes precipitation (forward reaction) and dissolu-
CO2 corrosion, which takes place at the steel surface. The mechanism of tion (backward reaction) of FeCO3 in aqueous solutions.
iron dissolution in a CO2 environment is complex and difficult to ex-
Fe(2aq+) + CO3(2 −aq) ⇌ FeCO3(s) (11)
plain (Kahyarian et al., 2016). A number of multi-step mechanisms
have been proposed by researchers for the anodic reaction of mild steel Precipitation occurs when the product of the ferrous ion and car-
in CO2 saturated aqueous media, which is outside the scope of this bonate ion activities exceeds the solubility limit (Ksp) of FeCO3 at
paper (Bockris et al., 1961; Dražić and Hao, 1982; Keddam, 2011). equilibrium condition, Equation (12).
However, anodic dissolution of iron can be simply expressed by Equa-
Ksp, FeCO3 = aFe2 + ∗aCO32 − (12)
tion (6):

288
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

where aFe2 + is the ferrous ion activity (known as effective concentra- when their saturation degree is greater than unity. Similar to FeCO3, the
tion), aCO32 − is the carbonate ion activity, and Ksp, FeCO3 is the solubility solubilities of CaCO3 (calcite in this paper) and MgCO3 decrease with
product of FeCO3 in the equilibrium condition. Ionic activity is related increasing temperature. The available correlations for calculation of Ksp
to concentration via Equation (13): of CaCO3 and MgCO3 do not take into account the effect of ionic
ai = γi Ci strength, unlike siderite, as shown in Equations (18) and (19), respec-
(13)
tively (Plummer and Busenberg, 1982), (Bénézeth et al., 2011).
while γi is the appropriate activity coefficient, which can be calculated Therefore, for calculation of the saturation degree of these carbonates,
by an extended Debye-Hückel relationship, Equation (14), as proposed the activity of ions should be considered rather than concentration.
by Davies, for solutions with ionic strength up to 0.5 M (Davies, 1962). Equations (20) and (21) express how to calculate saturation degree for
For complicated systems with higher ionic strength, the Debye-Hückel CaCO3 and MgCO3, accordingly.
theory is no longer accurate. In such aqueous systems, advanced models
2839.319
such utilizing Pitzer equations can provide more realistic activity logKsp, CaCO3 = −171.9065 − 0.077993∗Tk + + 71.595∗LogTk
Tk
coefficients (Pitzer, 1991):
(18)
I
log γi = −Az i2 ⎜⎛ − 0.3∗I ⎞⎟ 1476.604
⎝1 + I ⎠ (14) logKsp, MgCO3 = 7.267 − − 0.033918∗Tk
Tk (19)
the parameters associated with the above equations are:
a Ca2 + * a CO2 −
3
SCaCO3 =
ai : activity of species i, (mol/L) K sp, CaCO3 (20)
Ci : concentration of species i, (mol/L)
γi : activity coefficient of species i, (-) aMg2 + * aCO32 −
−3 SMgCO3 =
A = 1.82 × 106 (εT ) 2 ;
A ≈ 0.5 at 25°C (Stumm and Morgan, 2012) Ksp, MgCO3 (21)
ε: dielectric constant
FeCO3 has the lowest solubility compared to CaCO3 and MgCO3. It
T: solution temperature, (k)
means that in the same environmental condition, FeCO3 reaches sa-
z i : charge of the species i, (-)
turation prior to CaCO3 and MgCO3. Fig. 1 depicts how the solubility
I: ionic strength of the solution, mol/L, described in Equation (16)
product constant (Ksp) of such carbonates change with temperature
based on Equations (15), (18) and (19) (for consistency of the results,
Sun et al., proposed Equation (15) for calculation of Ksp, FeCO3 at
ionic strength is set to zero for Equation (15)). Earlier studies on MgCO3
different temperatures and ionic strengths (Sun et al., 2009):
solubility have been conducted at specific temperatures and then the
2.1963 results were extrapolated to other conditions. This approach has led to
logK SP, FeCO3 = −59.3498 − 0.041377∗Tk − + 24.5724∗Log (Tk )
Tk discrepancies in the reported solubility for magnesite at different tem-
+ 2.518∗I 0.5 − 0.657∗I (15) peratures (Kittrick and Peryea, 1986), (Pokrovsky et al., 2009). Béné-
zeth et al., have recently proposed Equation (19) for magnesite solu-
where Tk is the temperature (in Kelvin) and I is the ionic strength (in
bility which has been derived from extensive experimental solubility
mol/L), as determined by Equation (16):
data conducted on synthetic magnesite at temperatures from 50 to
1 200 °C in 0.1 mol kg−1 NaCl solutions with CO2 partial pressure of
I=
2
∑ ci z i2
i (16) 4–30 bars (Bénézeth et al., 2011).

where ci is the concentration of each species in solution and zi is the

2.2.1. Isostructurality of calcite and magnesite with siderite
charge of the species.
Being the main constituent of limestone, calcite is one of the most
Saturation degree (S) is a key parameter in scale or corrosion pro-
commonly observed minerals in sedimentary systems including oil re-
duct forming situations. Equation (17) expresses the saturation degree
servoirs. Calcite crystallizes with a hexagonal unit cell as shown in
with respect to FeCO3:
Fig. 2.
CFe2 + ∗ CCO32 − Calcite and magnesite are isostructural with siderite meaning that
SFeCO3 =
Ksp, FeCO3 (17) they share the same crystal structure; note the trigonal carbonate an-
ions packed between the calcium cations in Fig. 2. Each calcium is
where CFe2 + and CCO32 − are ferrous ion and carbonate ion concentra- shielded by six oxygens from six different carbonates (Ca-O distance ca.
tions. 2.4 Å). Single unit cells of these minerals are shown in Fig. 3, Fig. 4, and
The FeCO3 saturation degree greatly affects the corrosion behavior Fig. 5. Each unit cell is shown perpendicular to their xy-plane. Note the
since it is the main driving force for FeCO3 precipitation kinetics. When
the saturation degree is less than unity (i.e., unsaturated condition), a
higher corrosion rate is expected since a protective layer of FeCO3 is not
thermodynamically favored and may not be formed on the steel sub-
strate. When the saturation degree is greater than unity (i.e., super-
saturated solution), formation of FeCO3 layer is expected. This can re-
sult in lowering of the uniform corrosion rate as the non-conductive
FeCO3 crystals can cover the metal surface and hinder mass transfer of
relevant species for anodic and cathodic reactions (Nesic, 2007). When
the saturation degree is close to unity, the rate of formation and dis-
solution of FeCO3 can be similar, depending on the temperature.

2.2. CaCO3 and MgCO3

CaCO3 and MgCO3 are the most observed scales in oil production Fig. 1. Solubility products of siderite, calcite, and magnesite versus tempera-
systems, and often exist as mixed carbonates. Precipitation happens ture.

289
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

Fig. 2. Hexagonal crystal structure of calcite (CaCO3; blue = Ca, red = O,

black = C), four unit cells. (For interpretation of the references to colour in this Fig. 5. Hexagonal single unit cell of magnesite (MgCO3; yellow = Mg, red = O,
figure legend, the reader is referred to the Web version of this article.) black = C). (For interpretation of the references to colour in this figure legend,
the reader is referred to the Web version of this article.)

Fig. 3. Hexagonal single unit cell of siderite (FeCO3; tan = Fe, red = O,
black = C). (For interpretation of the references to colour in this figure legend,
the reader is referred to the Web version of this article.)
Fig. 6. Orthorhombic crystal structure of aragonite (CaCO3; blue = Ca,
red = O, black = C), four unit cells. (For interpretation of the references to
colour in this figure legend, the reader is referred to the Web version of this
article.)

(PbCO3) (Chai and Navrotsky, 2015). The key difference of aragonite-

type calcium carbonate from the calcite-type is that each calcium is
shielded by nine oxygens from five different carbonate ions (Ca-O dis-
tance 2.4–2.6 Å). Consequently, the pseudo-octahedral geometry
around each metal cation no longer applies. In addition, aragonite
calcium carbonate is likely metastable, readily converting to calcite.
Magnesite, siderite, and calcite have the same unit cell type and
similar cation radii and can co-exist in a solid solution (Chai and
Navrotsky, 2015). Ca and Mg can replace Fe in the crystal structure of
FeCO3 and form a mixed substitutional solid solution. These substitu-
tions have the potential to change the protective properties of FeCO3
against further corrosion, as reported by various researchers (Zhao
et al., 2005; Esmaeely et al., 2016; Ding et al., 2009). Table 2 shows key
Fig. 4. Hexagonal single unit cell of calcite (CaCO3; blue = Ca, red = O,
black = C). (For interpretation of the references to colour in this figure legend,
the reader is referred to the Web version of this article.) Table 2
Key structural parameters for magnesite, siderite, and calcite.
maintenance of isostructurality for each phase, with metal cations at Formula & Name Cation Unit Cell Type a (Å) c (Å) Density (g/
apices and within the unit cell. All the images related to crystal struc- Radius (Å) cm3)
ture are generated by CrystalMaker® software.
MgCO3 0.72 hexagonal 4.59 14.87 3.01
A less common orthorhombic CaCO3 is aragonite (Fig. 6). Heavier
(magnesite)
alkaline earth elements, as well as lead, crystallize carbonates with this FeCO3 (siderite) 0.78 hexagonal 4.72 15.46 3.94
structure, such as strontianite (SrCO3), witherite (BaCO3) and cerussite CaCO3 (calcite) 1.00 hexagonal 4.99 17.04 2.71

290
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

structural parameters of these carbonates.

3. Literature review of the effect of Ca2+ and Mg2+ on CO2

corrosion

Since a limited number of corrosion studies have been conducted

using Ca2+ and Mg2+, the corrosion mechanisms have not been
methodically characterized. However, conducted experiments on this
topic have, to a limited degree, elucidated CO2 corrosion in the pre-
sence of these alkaline earth cations. A review of the available literature
is provided below, which covers the main messages of each paper and
their shortcomings. In section 4, a conclusive discussion on the litera-
ture review and future experimental path are presented.
Eriksrud and Sontvedt conducted one of the earliest studies to
evaluate the effect of Ca2+ and Mg2+ along with other ions on CO2
corrosion behavior of API 5L X52 specimens (Navabzadeh Esmaeely Fig. 7. X-ray diffraction patterns of the surface layers formed on X65 carbon
et al., 2013). Potentiodynamic sweep were conducted for three different steel specimens at various pCO2 values, quiescent condition, 65 °C, 64 ppm
brines with 0, 400 and 1200 ppm of Ca2+ at 20 °C, 1 bar partial pressure Ca2+, and 72 ppm Mg2+ (Gaoet al., 2008).
of CO2 and over a pH range of 5.80–7.39. They concluded that for-
mation of protective FeCO3 layers enriched with Ca2+ resulted in a
and localized corrosion increased. However, in these experiments, the
drop in corrosion rate of mild steel while the effect of Mg2+ on the
authors also did not report solution pH.
general corrosion rate was reported to be negligible. One of the short-
X. Jiang et al., conducted experiments in a glass cell at 57 °C, in the
comings of this research was that experimental durations were rela-
presence of 1000 ppm Ca2+ and at different concentrations of Cl− to
tively short (2 days); therefore, possible phase transformation of the
distinguish the effect of Cl− and Ca2+ on pitting phenomena (Jiang
precipitated carbonates may not have been taken into account. CaCO3
et al., 2006). They carried out three experiments with different elec-
has three polymorphs, mineralogically named calcite, aragonite, and
trolytes, specifically 3 wt% NaCl, 3 wt% NaCl +1.5 wt% CaCl2, and
vaterite. Calcite is the most stable form of these CaCO3 polymorphs
4.6 wt% NaCl. Based on electrochemical impedance spectroscopy (EIS),
(Zhong and Mucci, 1989). There have been reports in the literature that
they reported pitting initiation times were 70 h for the experiment with
aragonite can transform to calcite after several days during corrosion
3 wt % NaCl, 41 h for the 3 wt% NaCl +1.5 wt% CaCl2 electrolyte, and
experiments, resulting in possible change of mechanical properties
23 h for 4.6% wt.% NaCl. Fig. 8 illustrates Nyquist plots over time for
(Navabzadeh Esmaeely et al., 2013).
the experiment with 3 wt% NaCl +1.5 wt% CaCl2; the specimens were
Zhao et al., claimed that the corrosion rate decreased in the “short
made of API N80 steel. The diameter of the Nyquist semicircles increase
term” in the presence of Ca2+ and Mg2+, but that there was no special
over time up to 41 h. The authors attributed this behavior to the for-
difference for “long-term exposure” (Zhao et al., 2005). They postulated
mation of protective layers on the steel surface and improvement of the
that the availability of Ca2+ and Mg2+ promoted the rapid formation of
protectiveness of the formed surface precipitates over time. However,
a protective layer in the short-term (before 72 h), which resulted in
the impedance noticeably decreased as the immersion time reached
lower general corrosion rates. On the other hand, without Ca2+ and
49 h. According to Fig. 8, the impedance, again, gradually increased
Mg2+, a protective layer could form only with longer exposure time
over the course of the experiment up to 96 h. The author related this
when Fe2+ concentration was sufficient for precipitation of corrosion
phenomenon to propagation and “repassivation”, which the authors of
products. They conducted experiments at 90 °C and CO2 pressure of
this review interpret as pit death, associated with pitting corrosion.
25 bars, with 1000 ppm Mg2+ and 6000 ppm Ca2+ in solution. In ad-
They also claimed that, while Cl− caused pitting, the presence of Ca2+
dition to the corrosion behavior, the corrosion product morphology and
postponed the pitting initiation time. While the authors did not propose
composition were changed by the presence of Ca2+ and Mg2+ in the
a mechanism for pitting attack by Cl−, their claim is in contrast with
aqueous electrolyte. Inspection of the XRD data indicated that the most
some other studies that suggest Cl− has no effect on pitting corrosion of
intense peak occurred at a position of less than 30° 2θ, which would be
carbon steels (Jiang et al., 2013), (Fang et al., 2013). Another
consistent with the larger Ca2+ being the major cation within the
formed metal carbonate lattice. Peak broadening also seems to have
occurred, consistent with inhomogeneity within the corrosion product-
scale hybrid surface layers. The authors described the primary corro-
sion product as “amorphous Fe(Ca,Mg) (CO3)2” which was thicker than
“ crystalline FeCO3”. Although pH plays a dominant role in corrosion
and scaling processes, it was unreported in this work.
Gao et al., conducted autoclave corrosion experiments with a si-
mulated brine from an oilfield to investigate the formation of surface
layers at different test conditions and their effect on CO2 corrosion of
mild steel (Gaoet al., 2008). The experimental conditions included
different flow rates (0, 0.5 and 1 m/s) and partial pressures of CO2 (0.1,
0.3 and 1 MPa) while the temperature was set to 65 °C, with 64 ppm
Ca2+, and 72 ppm Mg2+. Based on energy dispersive X-ray spectro-
scopy (EDS) and X-ray diffraction (XRD) patterns, they suggested the
formation of distinct phases of FeCO3, CaCO3, and MgCO3 at 0.1 MPa
pCO2, a solid solution of “(Fe,Ca,Mg)CO3” at 0.3 MPa CO2 partial
pressure, and a solid solution of “(Fe,Ca)CO3” at 1 MPa CO2 partial
pressure. Fig. 7 shows an example of X-ray diffraction patterns of the
surface layers formed on X65 carbon steel specimens at their quiescent Fig. 8. Nyquist plots versus time for API CT5 N80 materials in 3 wt% NaCl
condition. According to their work, at higher CO2 pressure both general +1.5 wt% CaCl2 electrolyte at 57 °C (Jiang et al., 2006).

291
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

Tavares et al., conducted long-term (28 days) autoclave experiments

to study the effect of CaCO3 on the corrosion behavior of low carbon
steel in high partial pressures of CO2 (Tavares et al., 2015). The test
electrolytes were saturated with CO2 and NaCl at 80 °C, the pCO2 was
15 Mpa. They designed two series of experiments, namely baseline and
CaCO3-saturated tests. The baseline test was carried out in the absence
of CaCO3. The CaCO3-saturated test was conducted in the presence of a
significant excess amount of solid CaCO3 in the electrolyte (0.5 mol/kg;
10 times higher than its solubility limit), ensuring the solution was
saturated with respect to CaCO3 over the course of experiments. They
claimed that general corrosion was predominant for both test series
rather than pitting corrosion. Indeed, they reported a decline in the
corrosion rate over time, as determined by mass loss. This behavior was
observed for both tests, the authors related their findings to the for-
mation of protective “corrosion scales”. Fig. 13 shows SEM images of
Fig. 9. Corrosion rate versus different concentrations of Ca2+ for X65 specimens “corrosion scale” morphologies after 72, 336, and 672 h of exposure to
at 75 °C and 10 bar pCO2 after 10 days of exposure (Ding et al., 2009). electrolytes with and without CaCO3. For the CaCO3-saturated solution,
they reported formation of a carbonate solid solution with an average
shortcoming of this work is that the authors did not measure the depth chemical composition of Fe0.79Ca0.21CO3, concluded from EDS data.
and width of pits; therefore, the magnitude of pitting rate was un- The corrosion product from the baseline condition was reported to be
quantified. In addition, the general corrosion rate for the different pure FeCO3, with apparently larger crystal size compared to the “cor-
electrolytes was unreported. rosion scales” formed in the presence of CaCO3 (more evident after 72
Ding et al., conducted 10-day autoclave experiments with X65 and 336 h; see Fig. 13). It is interesting that even at such low initial pH
specimens at 75 °C and 10 bar pCO2 with up to 512 ppm of Ca2+ and values (2.71 and 4.7, respectively, for electrolyte without/with CaCO3)
78 ppm Mg2+ to study the effect of Ca2+ concentration on corrosion carbonate layers precipitated. The authors did not measure or calculate
behavior (Ding et al., 2009). They reported that the corrosion rate in- pH at the end of their experiments. However, the solution water
creased as the concentration of Ca2+ increased, as shown in Fig. 9. They chemistry would have dramatically changed in the 1-liter autoclave
also reported that the presence of calcium ions changed the corrosion over the course of experiments and resulted in final pH values sig-
product layer's crystal size. The authors concluded that the addition of nificantly higher than for the initial conditions (Gao et al., 2017).
Ca2+ forced the corrosion product layer crystals into morphological Therefore, high pH and Fe2+ concentration (introduced to the solution
features that were “bigger” and “looser”; therefore, they did not act as a from the corroding surface) has favored precipitation of FeCO3 and
sufficiently protective barrier and, consequently, allowed corrosive FexCa1-xCO3 over time in these experiments.
species to more readily diffuse to the metal surface. They also showed The authors also reported that the average corrosion rate in the
XRD data, with shifted peaks from the FeCO3 peak position with in- presence of CaCO3 was lower when compared to the solution without
creasing Ca2+ concentration in the electrolyte. This shift is the result of CaCO3, as shown in Fig. 14. This was mainly due to a different initial
the change in the unit cell of the corrosion products with the formation pH of the two electrolytes rather than a direct effect of CaCO3 on sur-
of mixed carbonate layers (Fe1-xCaxCO3). It is noteworthy that magne- face layer protectiveness against corrosion. Indeed, the corrosion rates
sium was undetected in their corrosion product. However, this is un- after 336 and 672 h of exposure were almost identical for experiments
surprising as Mg2+ was present at a concentration close to its equili- with and without CaCO3, meaning that water chemistry of the system
brium concentration in relation to the solubility of MgCO3 in water at was unchanged after 336 h. This work was one of the few corrosion
the applied physicochemical conditions at the beginning of the ex- studies conducted in electrolytes saturated with CaCO3 and NaCl at a
periment. The authors also reported formation of a bilayer on the steel high partial pressure of CO2. However, the main shortcoming of this
surface. The concentration of calcium in the surface outer layer was research is that comparison of the two tests, with and without CaCO3,
higher than the inner layer, as confirmed by EDS and shown in Fig. 10. may be problematic due to the considerable difference in the initial
Esmaeely et al., studied the effect of Ca2+ on CO2 corrosion at low solution pH for each case; pH 4.7 for the solution with CaCO3 and pH
and high concentrations of Ca2+ in 7-day glass cell experiments at 80°C 2.71 for the solution without CaCO3. However, this work showed that
and pCO2 0.53 bar (Navabzadeh Esmaeely et al., 2013). They reported even in high concentrations of Ca2+ (around 2000 ppm; calculated
that for a low concentration of Ca2+ (up to 100 ppm) a protective layer from data provided by the authors), if the solution was saturated with
formed and the corrosion rate decreased with time. However, for the respect to CaCO3, a protective layer against further corrosion formed on
electrolyte with a high initial concentration of Ca2+ (1000 and the steel surface similar to the experiment in absence of CaCO3. This is
10,000 ppm), the formed layers were non-protective and corrosion rate in contrast with other studies, as described earlier, reporting that with
did not drop throughout the experiments; see Fig. 11. The authors also roughly similar Ca2+ concentrations, but different water chemistry
observed that the use of the electrolyte with 10,000 ppm Ca2+ led to conditions, a protective layer could not form (Ding et al., 2009),
severe pitting corrosion. In this research, the pH for the baseline test (Navabzadeh Esmaeely et al., 2013). What can be concluded from this
and the test with the presence of high concentration of Ca2+ were not study is that the saturation degree of the carbonates is a critical para-
identical. Upon adding such a high initial concentration of calcium ions meter rather than calcium and magnesium ion concentration in evalu-
(added as CaCl2.2H2O), massive precipitation of CaCO3 acidified the ating the effect of Ca/Mg-containing aqueous solutions on CO2 corro-
electrolyte according to Equation (22). Therefore, it can be postulated sion and solution pH measurements before and after the experiments
that the increase in the corrosion rate was due to the lowered solution are necessary to quantify this value.
pH values rather than increased calcium ion concentration. pH beha-
vior of electrolyte with different concentrations of Ca2+ is depicted in 4. Discussion and future experimental pathway
Fig. 12. It is worth noting that [Fe2+], [Ca2+], and pH were not kept
constant over the course of experiments. It is widely accepted that Ca2+ and Mg2+ do not directly participate
in electrochemical reactions involved in CO2 corrosion (George, 2003),
Ca2 + (aq) + H2 CO3 (aq) → CaCO3 (s ) + 2H+ (aq) (22) but research related to these cations in solution have shown an influ-
ence on the corrosion rate. It is probable that a solid solution of

292
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

Fig. 10. Elemental analysis of the presence of calcium in the bilayers formed on the steel surface. Note the higher concentration of calcium in the outer layer
(“surface scale”) compared to the inner layer (“inner scale”) when the test solution contained 512 ppm Ca2+ (Ding et al., 2009).

Fig. 11. Corrosion rate over time measured by LPR for different initial con- Fig. 12. Variation of pH over time for different concentrations of Ca2+ at 80 °C,
centration of calcium ions at 80 °C, pCO2 0.53 bar, 1wt % NaCl, and 10 ppm pCO2 0.53 bar, 1wt % NaCl, and 10 ppm Fe2+ (Navabzadeh Esmaeely et al.,
Fe2+ (Navabzadeh Esmaeely et al., 2013).1 2013).1 (note the difference in the electrolytes' pH caused by precipitation of
CaCO3 in high concentration of calcium ions).
FexCayMgzCO3 (where x + y + z = 1) would form in CO2 corrosion of
mild steel in the presence of Fe2+, Ca2+, and Mg2+ dependent upon the are also experiments suggesting magnesium ions have no effect on CO2
available mole fraction of its constituents cations in solution. Heur- corrosion processes (Eriksrud and Sontvedt, 1984). Some researchers
istically speaking, a higher concentration of magnesium into a solid claimed that calcium ions initiate pitting corrosion attack (Navabzadeh
solution of FexCayMgzCO3 would lead to an increase in solubility; this is Esmaeely et al., 2013), while others stated that the presence of Ca2+
because MgCO3 has a higher solubility than FeCO3 and CaCO3 (Fig. 1). would postpone the pitting tendency on carbon steel (Jiang et al.,
Therefore, the corrosion rate could dramatically increase with possible 2006). These discrepancies about the true effect of Ca2+ and Mg2+ on
onset of localized attack when small changes in solution chemistry CO2 corrosion necessitate development of systematic and well-designed
cause dissolution of Mg-bearing carbonates. procedures in corrosion testing for understanding the relevant issues
From a general literature review on this topic, one cannot con- surrounding CO2 corrosion in the presence of these ubiquitous alkaline
clusively describe the true effect of Ca2+ and Mg2+ on CO2 corrosion earth cations.
for different water chemistries and conditions. There are studies One of the important parameters that has caused discrepancies in
claiming that the general corrosion rate is higher in the presence of the open literature is the unknown and probably transient saturation
calcium ions (Ding et al., 2009), (Navabzadeh Esmaeely et al., 2013). degree of carbonates in the bulk solutions of autoclave, and glass cell,
Conversely, some researchers reported the exact opposite conclusions experiments. Saturation degree of the bulk solution is a more important
(Tavares et al., 2015), (Eriksrud and Sontvedt, 1984). In addition, there parameter than the individual ion concentrations when studying the
effect of Ca2+ and Mg2+ on corrosion mechanisms. Precipitation ki-
netics of carbonates is greatly influenced by the bulk saturation degree
1
Reproduced with permission from NACE International, Houston, TX. All
as the main driving force (Zhang et al., 2001), (Spanos and Koutsoukos,
rights reserved. Navabzadeh Esmaeely, Saba, et al. Effect of calcium on the 1998). Most researchers have ignored this important environmental
formation and protectiveness of iron carbonate layer in CO2 corrosion, characteristic, solely relying on initial ion concentration as the core
Corrosion, 69, 9, 2013. ©NACE International 1945. influencing parameter. Furthermore, the majority of available research

293
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

Fig. 13. SEM images of surface layers' morphology formed on low carbon steel after 72, 336, 672 h of exposure in CO2-saturated solutions (a, b, and c) without and
(d, e, and f) with CaCO3 (Tavares et al., 2015).

experimental setups were not appropriately studied and/or reported.

Corrosion processes and formation of corrosion products/scales are
influenced by the mass transfer of the relevant ions between the bulk
solution and the metal surface (Nesic, 2012). These important para-
metric characteristics will significantly affect results, making experi-
ments difficult to reproduce.
The current authors recommend that, for a precise study of the ef-
fect of Ca2+ and Mg2+ on CO2 corrosion, experiments be conducted in
relation to two distinct saturation scenarios with respect to CaCO3 and
MgCO3:

1. A non-scaling condition: when the solution is unsaturated, or at the

saturation point, with respect to the CaCO3 and MgCO3 that can
potentially form.
2. A scaling condition: when the solution is supersaturated with re-
Fig. 14. Corrosion rate, by weight loss, versus exposure time for test electrolytes
spect to CaCO3 and MgCO3 where it is expected that they will form.
with and without CaCO3 (Tavares et al., 2015).

It is worth mentioning that in corrosion studies the test solution,

on this topic has been conducted in conditions where the water usually with small volume, will eventually be saturated with respect to
chemistry of the systems was unstable (uncontrolled) and, in some FeCO3 due to the corrosion process and production of Fe2+. At such
cases, completely ignored (e.g., pH is unreported) (Zhao et al., 2005), condition, precipitation of FeCO3, as the corrosion product, is expected.
(Ding et al., 2009). In addition, the flow characteristics of most of the In the non-scaling condition, the following fundamental question

294
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

needs to be answered: References

• What is the effect of Ca 2+

and Mg2+ on nucleation, growth, and Alsaiari, H.A., Kan, A., Tomson, M., others, 2010. Effect of calcium and iron (II) ions on
the precipitation of calcium carbonate and ferrous carbonate. SPE J. 15 (02),
precipitation of FeCO3? There is report in the literature claiming the
294–300.
presence of calcium ions can hinder precipitation of FeCO3 (Alsaiari Bénézeth, P., Saldi, G.D., Dandurand, J.L., Schott, J., Jun. 2011. Experimental determi-
et al., 2010). However, such claims need to be further investigated nation of the solubility product of magnesite at 50 to 200 °C. Chem. Geol. 286 (1–2),
from the standpoint of corrosion. 21–31.
Bockris, J.O., Drazic, D., Despic, A.R., Aug. 1961. The electrode kinetics of the deposition
and dissolution of iron. Electrochim. Acta 4 (2), 325–361.
Another relevant scenario is when the solution is supersaturated Chai, L., Navrotsky, A., 2015. Synthesis, characterization, and energetics of solid solution
with respect to CaCO3 and MgCO3. In such conditions, precipitation of along the dolomite-ankerite join, and implications for the stability of ordered CaFe
(CO3)2. Am. Mineral. 81 (9–10), 1141–1147.
carbonates from the bulk solution is expected. Therefore, carbonates, Davies, C.W., 1962. Ion association. Butterworths 37–53.
with distinct boundaries, can form affecting the homogeneity and De Waard, C., Milliams, D.E., May 1975. Carbonic acid corrosion of steel. Corrosion 31
porosity of the layers. The main questions in this condition are: (5), 177–181.
Ding, C., Gao, K., Chen, C., Dec. 2009. Effect of Ca2+ on CO2 corrosion properties of X65

• Do such scales, without interference of ferrous ions, offer protec-

pipeline steel. Int. J. Miner. Metall. Mater. 16 (6), 661–666.
Dražić, D.M., Hao, C.S., Oct. 1982. The anodic dissolution process on active iron in al-
tiveness against further corrosion? There is almost no data in the kaline solutions. Electrochim. Acta 27 (10), 1409–1415.
literature about the protective properties of pure CaCO3 and MgCO3 Dresel, P.E., Rose, A.W., 2010. Chemistry and Origin of Oil and Gas Well Brines in
Western Pennsylvania. The Pennsylvania State University Open-File Report OFOG
generated in aqueous CO2 environments in studying the internal 10–01.0.
corrosion of pipelines. Eriksrud, E., Sontvedt, T., 1984. Effect of flow on CO2 corrosion rates in real and synthetic

• How does precipitation of CaCO3 and MgCO3 affect the formation of formation waters. Proc. Symp. CO2 Corros. Oil Gas Ind. NACE 1, 20–38.
Esmaeely, S.N., Young, D., Brown, B., Nesic, S., Nov. 2016. Effect of incorporation of
FeCO3? An early postulate is that precipitation of such scales on the calcium into iron carbonate protective layers in CO2 corrosion of mild steel.
steel surface can provide more sites for heterogeneous nucleation Corrosion 73 (3), 238–246.
and growth of the corrosion product, FeCO3, thus favoring its crystal Fang, H., Brown, B., Nesic, S., Mar. 2013. Sodium chloride concentration effects on
general CO2 corrosion mechanisms. Corrosion 69 (3), 297–302.
growth. However, such postulates require verification through rig- Gao, M., Pang, X., Gao, K., Feb. 2011. The growth mechanism of CO2 corrosion product
orous experimental design. films. Corrosion Sci. 53 (2), 557–568.
Gao, S., Jin, P., Brown, B., Young, D., Nesic, S., Singer, M., Mar. 2017. Corrosion behavior
of mild steel in sour environments at elevated temperatures. Corrosion 73 (8),
915–926.
5. Conclusions Gao, K., et al., Oct. 2008. Mechanical properties of CO2 corrosion product scales and their
relationship to corrosion rates. Corrosion Sci. 50 (10), 2796–2803.
George, K., 2003. Electrochemical Investigation of Carbon Dioxide Corrosion of Mild Steel
This review of literature data shows inconsistencies relating to the
in the Presence of Acetic Acid. Ph.D. Dissertation. Ohio University.
influence of Ca2+ and Mg2+ on CO2 corrosion of mild steel. This is Graver, D.L., Newton, L.E., Hausler, R.H., 1984. CO2 Corrosion In Oil And Gas Production:
mainly due to undefined water chemistry and flow characteristics for Selected Papers, Abstracts, and References. NACE International.
most of the studied systems. Most of the previous work had focused on Institute for Corrosion, Multiphase Technology, 2014. MULTICORPTM. Ohio University.
Jiang, X., Zheng, Y.G., Qu, D.R., Ke, W., Oct. 2006. Effect of calcium ions on pitting
the initial concentrations of Ca2+ and Mg2+ as the core influencing corrosion and inhibition performance in CO2 corrosion of N80 steel. Corrosion Sci. 48
parameter. This probably has resulted in completely different bulk sa- (10), 3091–3108.
turation degrees with respect to the associated carbonates; saturation Jiang, X., Nesic, S., Kinsella, B., Brown, B., Young, D., Jan. 2013. Electrochemical in-
vestigation of the role of Cl- on localized carbon dioxide corrosion behavior of mild
degree is the main driving force for precipitation of carbonates rather steel. Corrosion 69 (1).
than the individual ions. Moreover, the water chemistry of the systems Kahyarian, A., Singer, M., Nesic, S., Feb. 2016. Modeling of uniform CO2 corrosion of mild
was uncontrolled and unknown in most studies, meaning that the initial steel in gas transportation systems: a review. J. Nat. Gas Sci. Eng. 29, 530–549.
Kahyarian, A., Achour, M., Nesic, S., 2017. CO2 corrosion of mild steel. In: Trends in Oil
and final test conditions were not the same. However, the available and Gas Corrosion Research and Technologies. Elsevier.
literature is indeed invaluable for spotlighting the mechanism of CO2 Keddam, M., 2011. Anodic dissolution. In: Corrosion Mechanisms in Theory and Practice,
corrosion in the presence of these alkaline earth cations. The literature Third. CRC Press, pp. 149–215.
Kermani, M., Morshed, A., 2003. Carbon dioxide corrosion in oil and gas production-A
unanimously agrees that Ca2+ and Mg2+ do not participate in elec-
compendium. Corrosion 59 (8), 659–683.
trochemical reactions at the steel surface and that a mixed metal car- Kittrick, J., Peryea, F., Feb. 1986. Determination of the gibbs free-energy of formation of
bonate can form in presence of Fe2+, Ca2+, and Mg2+. Based on the magnesite by solubility methods. Soil Sci. Soc. Am. J. 50 (1), 243–247.
Mansoori, H., Mowla, D., Esmaeelzadeh, F., Mohammadi, A.H., 2013a. Case study: pro-
literature review, the current authors have suggested two separate
duction benefits from increasing C-values. Oil Gas J. 111 (6), 64–69.
scenarios in conducting future experiments on this topic, namely in Mansoori, H., Mirzaee, R., Mohammadi, A.H., Esmaeelzadeh, F., 2013b. Acid washes,
non-scaling and scaling conditions. Furthermore, for the sake of re- oxygenate scavengers work against gas gathering failures. Oil Gas J. 111 (7),
producibility of the results, the water chemistry of the system has to be 106–111.
Mansoori, H., Mirzaee, R., Esmaeilzadeh, F., Vojood, A., Soltan Dowrani, A., Dec. 2017.
defined and maintained, or at least documented, over the course of each Pitting corrosion failure analysis of a wet gas pipeline. Eng. Fail. Anal. 82, 16–25.
experiment. Also, the flow characteristics of the system should be well- Navabzadeh Esmaeely, S., Choi, Y.-S., Young, D., Nesic, S., May 2013. Effect of calcium on
defined and reported since it affects the corrosion and scaling processes. the formation and protectiveness of iron carbonate layer in CO2 corrosion. Corrosion
69 (9), 912–920.
Nesic, S., Dec. 2007. Key issues related to modelling of internal corrosion of oil and gas
pipelines – a review. Corrosion Sci. 49 (12), 4308–4338.
Acknowledgments Nesic, S., Jul. 2012. Effects of Multiphase flow on internal CO2 corrosion of mild steel
pipelines. Energy Fuels 26 (7), 4098–4111.
Nesic, S., Postlethwaite, J., Olsen, S., 1996. An electrochemical model for prediction of
The authors would like to thank the following companies for their corrosion of mild steel in aqueous carbon dioxide solutions. Corrosion 52 (4),
financial support during the writing of this review: Anadarko, Baker 280–294.
Nordsveen, M., Nesic, S., Nyborg, R., Stangeland, A., May 2003. A mechanistic model for
Hughes, BP, Chevron, CNOOC, ConocoPhillips, DNV GL, ExxonMobil,
carbon dioxide corrosion of mild steel in the presence of protective iron carbonate
M-I SWACO (Schlumberger), Multi-Chem (Halliburton), Occidental Oil films—Part 1: theory and verification. Corrosion 59 (5), 443–456.
Company, PTT, Saudi Aramco, Shell Global Solutions, SINOPEC (China Nyborg, R., 2002. Overview of CO2 corrosion models for wells and pipelines. In:
Presented at the CORROSION 2002.
Petroleum), TransCanada, TOTAL, and Wood Group Kenny. The au-
Nyborg, R., 2010. CO2 corrosion models for oil and gas production systems. In: Presented
thors are also grateful to Dr. Bert Pots for discussions relating to pre- at the NACE International.
paration of this manuscript. All figures from other journals, reproduced Oddo, J.E., Tomson, M.B., 1994. And others, “why scale forms in the oil field and methods
herein, have been done so after securing permission from the original to predict it. SPE Prod. Facil. 9 (01), 47–54.
Ogundele, G., White, W., Nov. 1987. Observations on the influences of dissolved
publishers.

295
H. Mansoori et al. Journal of Natural Gas Science and Engineering 59 (2018) 287–296

hydrocarbon gases and variable water chemistries on corrosion of an API L80 steel. Sun, W., Nesic, S., Woollam, R.C., 2009. The effect of temperature and ionic strength on
Corrosion 43 (11), 665–673. iron carbonate (FeCO3) solubility limit. Corrosion Sci. 51 (6), 1273–1276.
Pitzer, K.S., 1991. Ion interaction approach: theory and data correlation. Act. Coeff. Tavares, L.M., da Costa, E.M., Andrade, J. J. de O., Hubler, R., Huet, B., Dec. 2015. Effect
Electrolyte Solut. 75–154. of calcium carbonate on low carbon steel corrosion behavior in saline CO2 high
Plummer, L.N., Busenberg, E., Jun. 1982. The solubilities of calcite, aragonite and vaterite pressure environments. Appl. Surf. Sci. 359, 143–152.
in CO2-H2O solutions between 0 and 90°C, and an evaluation of the aqueous model Tran, T., Brown, B., Nesic, S., 2015. Corrosion of mild steel in an aqueous CO2 en-
for the system CaCO3-CO2-H2O. Geochem. Cosmochim. Acta 46 (6), 1011–1040. vironment–basic electrochemical mechanisms revisited. In: Presented at the NACE
Pokrovsky, O.S., Golubev, S.V., Schott, J., Castillo, A., Jul. 2009. Calcite, dolomite and Corrosion Conference.
magnesite dissolution kinetics in aqueous solutions at acid to circumneutral pH, 25 to Zhang, Y., Shaw, H., Farquhar, R., Dawe, R., Apr. 2001. The kinetics of carbonate scaling-
150°C and 1 to 55 atm pCO2: new constraints on CO2 sequestration in sedimentary application for the prediction of downhole carbonate scaling. J. Petrol. Sci. Eng. 29
basins. Chem. Geol. 265 (1–2), 20–32. (2), 85–95.
Salman, M., Qabazard, H., Moshfeghian, M., Jan. 2007. Water scaling case studies in a Zhao, G.X., Jian-Ping, L., Shi-Ming, H., Xiang-Hong, L.U., 2005. Effect of Ca2+ and Mg2+
kuwaiti oil field. J. Petrol. Sci. Eng. 55 (1), 48–55. on CO2 corrosion behavior of tube steel. J. Iron Steel Res. Int. 12 (1), 38–42.
Spanos, N., Koutsoukos, P.G., 1998. Kinetics of precipitation of calcium carbonate in Zhong, S., Mucci, A., Dec. 1989. Calcite and aragonite precipitation from seawater so-
alkaline pH at constant supersaturation. Spontaneous and seeded growth. J. Phys. lutions of various salinities: precipitation rates and overgrowth compositions. Chem.
Chem. B 102 (34), 6679–6684. Geol. 78 (3), 283–299.
Stumm, W., Morgan, J.J., 2012. Aquatic Chemistry: Chemical Equilibria and Rates in Zoltai, T., Stout, J.H., 1984. Mineralogy: Concepts and Principles. Burgess Publishing
Natural Waters. John Wiley & Sons. Company.

296