Corrosion Science 135 (2018) 158–166

Contents lists available at ScienceDirect

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

Effects of thiosulphates and sulphite ions on steel corrosion☆ T

⁎
M. Cabrini, S. Lorenzi , T. Pastore
Department of Engineering and Applied Sciences, University of Bergamo, Viale Marconi 5, 24044, Dalmine (Bergamo), Italy

A B S T R A C T

The paper deals with the general corrosion of carbon steel in geothermal plants for conveying and re-injecting
condensates. The results of electrochemical and weight loss tests are discussed in function of sulphite and
thiosulphate concentrations typical of Italian geothermal power plants. Thiosulphates can inhibit the reactions
between oxygen and sulphites and increase the corrosion rate, but they control the acidification resulting from
the reactions of sulphites with oxygen. In anaerobic solutions, both species modify mainly the cathodic process.
In the absence of oxygen, the corrosion process leads to the formation of mackinawite scales, while iron sulphide
forms due to the reduction of thiosulphate and sulphite.

1. Introduction alloys, and stress corrosion [6] caused by sulphides in high strength
steels. Sulphur may be present in the form of numerous compounds that
In geothermal plants that generate electrical energy, the steam are characterized by different degrees of oxidation. Because of the in-
produced in the extraction wells is conveyed to steam turbines that stability of these compounds, they can become involved in oxidation
transform mechanical energy into electrical energy. The condensed and reduction processes that depend on the characteristics of the en-
water is then sent to the cooling towers and, finally, to a well for re- vironment in which they are dissolved. As a consequence, sulphur can
injection into the ground, to avoid environmental pollution by sub- initiate a large variety of corrosion phenomena that are characterized
stances present in the condensates. This also helps prevent subsidence by complex behaviour [5,6].
that might be induced by the extraction of significant quantities of In addition to sulphides and sulphates, sulphur can also exist as
water, while supporting steam pressure in the geothermal reservoir. sulphites. Of great importance to corrosion are the polythionic acids, of
The process releases into the atmosphere approximately 75% of the generic formula H2SxO6, which have at least one sulphur atom bridging
water flow [1]. two sulphonic acid units, and the thiosulphate group that consists of a
The extracted fluid also contains non-condensable gases (NCGs), in sulphate group bound to a sulphur atom. Thiosulphate ions display
a variety of concentrations depending on the reservoir, with extracts strongly acidic behaviour and, in water, they are normally dissociated.
from the Larderello geothermal wells having an NCG content of around These compounds are involved in most of the reactions involving
5% [2]. These fluids contain mainly carbon dioxide and significant polythionic acids, and can be formed by reduction on a metal surface
trace amounts of hydrogen sulphide [3]. To limit their release into the [6–8].
atmosphere, and to maintain their levels to values that are significantly The role of thiosulfates in corrosion has been the object of several
below those perceptible by the population, steam plants in Italy are studies, and their corrosive effects have been well known since the
equipped with abatement systems that oxidize hydrogen sulphide to 1940s, coinciding with the first accidents in the oil and paper industries
SO2. This gas is subsequently washed with the condensation water involving sensitized austenitic stainless steels [7] [4]. These failures
produced by the plant, and this dissolves the SO2 in the form of sul- were caused by stress corrosion promoted by polythionic acids and
phite. Subsequently, sulphites give rise to sulphates and thiosulphates thiosulphates, and were caused by the entry of air and moisture during
by further reactions with oxygen and sulphur [1,4]. the plant maintenance [9]. Similar problems occurred several years
Literature reports have highlighted the influence of the thiosulphate later, in pressurized-water nuclear reactors, where solutions containing
ion on various forms of corrosion, from that of generalized carbon steel, thiosulphates were used as emergency liquids for the absorption of
to the localized corrosion of stainless steels a [4] and nickel alloys, radioactive iodine. In the paper industry, the addition of sodium thio-
intergranular corrosion of sensitized stainless steels [5] and nickel sulphate, used as a bleaching agent, resulted in the corrosion of AISI

☆
Consorzio INSTM, UdR “Materials and Corrosion”, Via G.Giusti 6, Firenze, Italy.
⁎
Corresponding author.
E-mail addresses: [email protected] (M. Cabrini), [email protected] (S. Lorenzi), [email protected] (T. Pastore).

https://doi.org/10.1016/j.corsci.2018.02.046
Received 13 January 2017; Received in revised form 19 February 2018; Accepted 20 February 2018
Available online 02 March 2018
0010-938X/ © 2018 Elsevier Ltd. All rights reserved.
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

304 stainless steel [8]. Corrosion behaviour was studied by weight loss, and electro-
On corrosion-resistant alloys, thiosulphates tend to become reduced chemical measurements. Potentiodynamic polarization tests and
to sulphur in zones where the passive film breaks down, according to cathodic potentiostatic polarization tests were also performed by using
the following reaction: IVIUM Compactstat potentiostat. Weight loss measurements were per-
formed by means of Sartorius BP211D scale on the disc specimens,
S2 O32 − + 6H+ + 4e− = 2S + 3H2 O E 0 = 0.465 VSHE (1) described above. The surface of each disc was polished with emery
On the one hand, the sulphur that is deposited during this reaction paper of up to 2400 grit. The exposed surfaces were further polished to
stimulates anodic processes on the active surface, and on the other, it 1 μm using diamond paste, followed by rinsing and degreasing with
prevents the adsorption of hydroxide ions. In this way, it prevents acetone in an ultrasonic bath, and then drying. After exposure, the
passive film restoration. However, the formation of sulphur takes place specimens were pickled in diluted hydrochloric acid inhibited with
only on the active metal, since its formation is inhibited by the passive, hexamethylenetetramine, before being weighed. During these experi-
chromium oxide film, and is favoured by acidity. The action of thio- ments, two discs were immersed in a 400 mL closed glass container
sulphate ions occurs mainly on active steel or whenever the passive film filled with the de-aerated solution, under continuous nitrogen bubbling.
is damaged, i.e. at places of localized corrosion. Electrochemical tests were performed under stagnant conditions
Several reports agree that corrosion phenomena due to thiosul- using a rotating cylinder electrode EG&G mod. 616 RDE. In the first
phates can occur only in the presence of other substances that are able case, one litre ASTM cell was used, in which the exposed surface area of
to attack the passive film, or when the film is damaged by mechanical the specimen was 1 cm2. A double-junction Ag/AgCl/KClsat. from now
stress, i.e. at the apex of a crack formed by stress corrosion cracking or on indicated by Ag/AgCl reference electrode and a graphite counter-
similar [7,9–18]. Some reports [19–21] indicate that thiosulphate can electrode were placed in separate compartments with porous ceramic
itself promote localized corrosion, even in the absence or presence of diaphragms. A Huber-Luggin capillary was used. The flux of the gases
small amount of chloride. Many studies have been carried out on produced at the counter-electrode of the test cell was vented through a
general corrosion in neutral or weakly acid solutions containing thio- hole in the upper part of the compartment. In experiments using the
sulphate, but very few investigations have been carried out in solutions rotating electrode, similar equipment to that described above was used,
with high sulphite content. To the best of our knowledge, there are no and the rotating shaft was sealed against external air exposure.
data presented in the literature regarding the simultaneous presence of Before any electrochemical testing, the solution was de-aerated by
these two chemical species. nitrogen bubbling for at least twelve hours. Subsequently, a constant
The object of this research is the study of the general corrosion of flow of nitrogen was maintained throughout the test period.
the carbon steel used in geothermal plants for conveying and re-in- Potentiodynamic curves were plotted after about 5 min of immersion, at
jecting condensates containing thiosulfates and other species produced a scan rate of 10 mV/min, from the free corrosion potential to a po-
by the oxidation of sulphites. The paper reports the results from our tential of 100 mV higher, for anodic curves, or 700 mV lower, for
study into the corrosion mechanism of carbon steel in contact with cathodic curves. Ohmic drop compensation was applied to all po-
oxidation species of sulphur. tentiodynamic curves. The ohmic drop was evaluated by means of
electrochemical impedance measurements at a frequency of about
2. Materials and methods 103 Hz, with an amplitude of 10 mV around the free corrosion potential.
This correction is significant only for current densities above 5 mA/cm2
All experiments were carried out on API 5L grade X65 steel, with and 1 mA/cm2 for the disc electrode and the rotating cylinder electrode,
ferritic-pearlitic structure. Table 1 shows the chemical composition respectively.
determined by means of spectrographic analysis. Disk specimens of Potentiostatic experiments were carried out by detecting the value
15 mm diameter and 5 mm height, or cylinders of 12 mm diameter and of the stationary current, at −800 mV vs. Ag/AgCl/saturated KCl, on a
18 mm height were used. rotating electrode at increasing speeds between 0 and 6000 rpm. The
The composition of the test solutions used in this study is re- composition and morphology of the corrosion scale was characterized
presentative of that of the condensation water found in Italian geo- by electron microscopy and X-ray analyses of specimens immediately
thermal plants [19], namely 3000 ppm of SO42−, 300 ppm of sulphites, after testing. The specimen surfaces were washed with distilled water
150 ppm of thiosulphate, with pH ranging between 5 and 7. The pH was and dried under a flow of nitrogen. Analyses of the powdered corrosion
adjusted by the addition of sulphuric acid. All experiments were carried scale products were performed on specimens after long exposure.
out at room temperature. In some experiments, the sulphite and thio-
sulphate composition of the solution was modified in order to assess the 3. Results
effects of these ions on the dissoved oxygen concentration. All solutions
were prepared using sodium salts. 3.1. Oxygen–sulphite reaction
The stability of each solution was evaluated by measuring the
oxygen concentration and pH before and after each experiment. Oxygen The combined measurements of oxygen and pH permit the accurate
concentration was measured by means of WTW FDO® 925–optical description the rate of the reaction of sulphite with dissolved oxygen.
oxygen sensor and pH was measured by means of AMEL 2335 pH- Fig. 1 depicts the oxygen concentration of the of the solution at the
meter. Furthermore, monitoring was carried out for up to 24 h, with air beginning of the experiment. The addition of 300 ppm of sulphite, in
or nitrogen bubbling, with and without addition of sulphite and thio- the absence of thiosulphate ions, causes a rapid depletion of dissolved
sulphate. To evaluate the effect of the sulphite/thiosulphate ratio on oxygen, and this reaches zero in a matter of minutes. This reaction leads
oxygen content, the oxygen concentration of the solution exposed to air to a variation in pH, with acidity increasing with the amount of reacted
bubbling was measured after the progressive addition of 150 ppm of sulphite. In a solution containing 300 ppm of sulphite and continuous
sulphite, every 5 min. air bubbling, stable pH values between 3 and 4 were obtained (Fig. 2).

Table 1
Composition of steel used in this study (%w).

Type C Mn Si P S Ni Cr Mo Cu Nb Ceq

API 5L grade X65 0.09 1.64 0.24 0.003 0.002 0.02 0.03 0.01 0.01 0.049 0.366

159
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

Fig. 4. Variations in oxygen concentration following the addition of increasing quantities

of sulphites.

Fig. 1. Variations in dissolved oxygen concentration following the addition of sulphite to

solutions containing only sulphates.

Fig. 5. The dependence of oxygen reaction rate on the sulphite/thiosulphate ratio (de-
creases in dissolved oxygen content per time unit).
Fig. 2. Time-dependence of the pH of aerated solutions.

amount of sulphite until a steady state was reached. Data derived by

authors are in good agreement with those obtained by Mo et al. [27]. As
this ratio increases by a factor of 5, the rate is observed to increase by a
factor 6.5.

3.2. Weight loss experiments

Fig. 6 summarizes the results from the weight loss measurements, as

a function of pH and the number of days of exposure of the specimen, in
the de-aerated solution. The corrosion rate is expressed as mass varia-
tion per surface unit and exposure period. It is the average value re-
ferred to the whole exposure period. It is evident that the corrosion rate
decreases as the number of days of exposure increases. This decrease is
more noticeable for low pH values, while at pH 7.2 the rate is

Fig. 3. Variations in dissolved oxygen concentration following the addition of sulphite to

a solution of sulphate containing thiosulphate.

The addition of 150 ppm of thiosulphate to the solution, prior to the

addition of sulphite, significantly modifies the observed behaviour. The
concentration of oxygen was observed to remain at values very close to
its initial level (Fig. 3) and no acidification was observed. On the
contrary, a slight increase in pH was detected (Fig. 2). Fig. 4 depicts the
variation in oxygen concentration with further additions of fixed
amounts of sulphite, equal to 150 ppm each. The oxygen concentration
was monitored after 5 min, following the initial addition of sulphite.
The variation in oxygen concentration between additions becomes
greater with increasing sulphite content. Fig. 5 depicts the dependence
of the estimated reaction rate on the weight ratio between sulphite and
thiosulphate. The reaction rate was derived by the measurement of Fig. 6. Average corrosion rates from weight loss measurements as a function of exposure
oxygen concentration during time, after the addition of constant time and pH.

160
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

cathodic curve is not modified at potentials approaching the free cor-

rosion potential. At very low polarizations, Tafel behaviour is clearly
seen and gradually becomes more evident with increasing rotational
speed.
The cathodic polarization curves for the solution containing sulphite
and thiosulphate were significantly different from that of the de-aerated
solution containing only sodium sulphate (Fig. 10). Without sulphite
ions and thiosulphate, in the absence of oxygen, Tafel behaviour can be
observed at very low potentials, due to the dissociation of water. On the
other hand, in the presence of these ions, a complex modification of the
cathodic curve is observed with an increase in the cathodic current
density at less negative potentials, and greater overvoltage at very ne-
gative potentials. The Tafel curve for hydrogen reduction by water
dissociation becomes visible at potentials more negative than
150–200 mV, compared to the curve for the solution containing only
Fig. 7. Dependence of pH on weight loss in a solution without oxygen. sulphates.
The polarization resistance (Rp) was estimated through the Stern
approximately constant with time. The corrosion rates during the first and Geary relationship, by calculating the slope of the potentiodynamic
day of exposure at pH 5 and 7.2 differ by a factor of about eight, after curves at polarizations of ± 20 mV. The constant of Stern and Geary
which this difference tends to become smaller. At the point of extrac- relationship is assumed to be 26 mV for all experimental conditions.
tion, all specimens exhibited a thick black scale of poorly adhered This value has been estimated by taking into account that the slope of
corrosion products, except those exposed at pH 7.2. The data displayed the cathodic polarization curve is much higher than that of the anodic
in Fig. 7 show that pH increases slightly over time, except for samples curve (ba), and that the slope of the latter is very close to 60 mV (Fig. 8).
immersed in solutions at pH 6.5 and 7.2. This value correlates well with that derived from the measurements of
weight loss at day one.
Fig. 11 plots the dependence of corrosion rate, estimated from the
3.3. Potentiodynamic experiments
potentiodynamic experiments, on pH and rotational speed of the elec-
trode. The results for the stationary electrode refer to tests performed
Fig. 8 shows the polarization curves for the steel sample in stagnant
on both cylindrical and disc electrodes. The results confirm that the
solutions. The anodic curves remain substantially unaffected by pH, so
corrosion rate increases as pH decreases. Moreover, the effect of the
that trends are comparable. On the other hand, there is a clear shift of
rotational speed of the electrode decreases with increasing pH. This
the cathodic curve towards higher values of current density as pH de-
effect is almost negligible at pH 7.2, but becomes predominant at pH 5,
creases. At pH 7.2, the cathodic branch of the polarization curve is
where a significant increase in current density, at a rotational speed of
characterized by Tafel behaviour and limiting currents at lower po-
1000 rpm, is observed.
tentials. In the two most acid solutions, a significant increase in current
density can be observed at potentials slightly below the free corrosion
potential, with a progressive increase in slope on moving from pH 7.2 to 3.4. Potentiostatic experiments
pH 5. In addition, two different limiting processes can be easily iden-
tified. In these solutions, the curves tend toward a single Tafel slope, at Fig. 12 shows the cathodic current density measured during po-
around 180 mV/decade, at very negative potentials, due to reduction tentiostatic experiments at −800 mV vs. Ag/AgCl. A sharp increase in
reactions involving hydrogen produced by the dissociation of water. cathodic current is observed, in a relatively short time, as the rotational
The cathodic polarization curves are observed to change with speed increases. The stationary current density as a function of the
changes in the fluid dynamic conditions. Fig. 9 shows the effect of the rotational speed is shown in Fig. 13. The current density increases
rotational speed of the electrode. The higher the rotational speed, the follow a power law, with an exponent that progressively increases from
higher the cathodic limiting current density, in both the region close to less than 0.1 in solution at pH 7.2 to about 0.3 at pH 5. While the
the free corrosion potential, and at more negative potentials, con- kinetics of both anodic and cathodic processes, at pH 7.2, are affected
firming that transport phenomena are significant. In the solution at pH by rotational speed, the corrosion rate remains substantially un-
7.2, the effect is significant only at the lowest potentials, whilst the changed.

Fig. 8. Effect of pH on the polarization curves conducted with stationary electrodes. Anodic curve (left) and cathodic curve (right).

161
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

Fig. 9. Effect of pH and rotation speed on the cathodic potentiodynamic curves for the rotating cylinder electrode.

Fig. 11. Dependence of current density on pH and rotation rate.

Fig. 10. Effect of the addition of sulphite and thiosulphate on the cathodic polarization
curve.
chemical formula Fe(1+x)S, where x varies between 0.057 and 0.064
3.5. Scale morphology and composition [20] and it is commonly observed in H2S-containing environments. The
spectra obtained from the sample exposed for 10 days, and the powder
The corrosion scale from the weight-loss specimens was analysed obtained from the sample surface exposed for 30 days, reveal that the
immediately after the exposure period by means of X-ray diffraction peak corresponding to amorphous iron sulphide is less pronounced after
(XRD), directly on the sample, or on the corrosion scale powder ground longer exposure, whilst the mackinawite peak becomes more evident.
from the specimen surface after prolonged exposure. Fig. 14 shows the
time-dependent variation of the spectrum of the corrosion products 4. Discussion
after 3 days, 10 days and 30 days of immersion in the solution at pH 5.8.
The spectrum of the sample exposed for 3 days was compared those of In the test solutions considered in the present work, sulphur is
the aluminium sample-holder and bare steel. The presence of at least present in the form of oxidized species. The instability of sulphites and
four additional peaks can be noticed at 16.7°, 30°, 39° and 49.5°, and thiosulphates can lead to other chemical species through reduction or
are attributable to the amorphous phase of FeS and mackinawite. The disproportion, firstly to sulphur and hydrogen sulphide, and though
latter is a non-stoichiometric sulphide with a tetragonal structure and reactions with other chemical species, mainly sulphites, thereby chan-
ging the composition of the environment. Moreover, iron sulphide

162
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

complicated and can modify electrochemical processes.

4.1. Corrosion processes in neutral or weakly acid water

The corrosion of steel in neutral or acidic aerated solutions occurs

through anodic iron oxidation, supported by cathodic oxygen reduc-
tion. The cathodic process involving the direct reduction of the hy-
drogen ion becomes predominant with decreasing pH. In neutral aer-
ated water, the corrosion rate is equal by the oxygen limiting current.
Under these conditions, the corrosion rate increases proportionally to
the concentration of dissolved oxygen and the water flux, because
oxygen transport towards the metal surface increases with increasing
flux. The latter decreases as deposits form, both calcareous and corro-
sion scale. Temperature has a complex effect. In the range from ambient
to 40–50 °C, the corrosion rate generally increases, approximately
Fig. 12. Effect of rotational speed on the cathodic current density during the potentio-
static experiment at −800 mV vs. Ag/AgCl in solution at pH 5.8.
doubling every 20–25 °C.
Hydrogen can form according to different reactions. In acidic so-
lutions, the hydrogen reduction reaction is prevalent due to the high
concentration of hydrogen ions:
2H+ + 4e− → H2 (2)
In neutral water, the concentration of hydrogen ions is too low to
support a relevant corrosion process. Hydrogen can be formed, how-
ever, through the dissociation of water:
2H2 O + 2e− → H2 + 2OH− (3)
However, this process takes place only at very negative potentials,
well below those that are usual for iron corrosion.
In neutral or weakly acid solutions, in the absence of oxygen, the
corrosion rate of steel is negligible, in the order of tens of microns per
year. Higher corrosion rates are expected only in presence of substances
that can promote the supply of hydrogen. These processes are different
to those depicted in Eq. (2) and Eq. (3). Among these substances,
Fig. 13. Effect of rotation speed on the cathodic current density at −800 mV vs. Ag/AgCl, carbon dioxide and hydrogen sulphide can be considered as prime
as a function of pH in de-aerated solutions (3000 ppm SO42−, 300 ppm SO32−, 150 ppm candidates. The first is responsible for “sweet corrosion”, and is asso-
S2O32−).
ciated with high corrosion rates of carbon steel, well above those from
oxygen attack observed in the oil industry, mining and geothermal
corrosion scale is formed, which provides a much less protective scale environments. Carbon dioxide supports the cathodic development hy-
compared to iron oxide. The corrosion of carbon steel in waters that drogen, even at pH between 5 and 7, by direct reduction of carbonic
contain high concentrations of sulphites and thiosulphates is therefore acid, according to the reaction:

Fig. 14. Dependence of the XRD spectrum of the corrosion products as a function of exposure time.

163
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

2H2 CO3 + 2e− → H2 + 2HCO3− (4)

The development of hydrogen by the direct reduction of hydrogen

sulphide has been described by Bolmer [21] in the 1960s, and more
recent studies can be found [22]. This reaction takes place according to
Eq. (5):

2H2S + 2e− → H2 + 2HS− (5)

4.2. Effect of thiosulphate on oxygen-induced corrosion

Low amounts of sulphite, well below the concentrations considered

in this work, are normally used in industrial plants for oxygen-
scavenging purposes. Sulphites react with oxygen dissolved in the
water, causing its removal. The oxidation-reduction reaction, which is
Fig. 15. Effect of thiosulphate and sulphite on the corrosion rate, as a function of the
fast even at ambient temperature (Eq. (6) leads to the formation of rotational speed of the electrode, measured by the polarization resistance technique.
sulphates by the reaction:

2HSO3− + O2 → H+ + 2SO42 − (6) characteristic curves for both anodic and cathodic process, and increase
the kinetics, with a major effect on the cathodic reaction, both for po-
In industrial plants, sulphite is added, as a sequestering agent, in
tentials close to the free corrosion potential, and at much lower po-
proportion to the oxygen content of the water.
tentials [23]. They also modify the cathodic polarization curve, causing
The presence of thiosulphates affects the oxygen-scavenging reac-
the occurrence of a limiting process at low potential values. Similar
tion promoted by sulphites (Fig. 3). Ulrich et al. and Mo et al. studied
behaviour is observed in solutions containing sulphites (Fig. 10). In the
their inhibiting effect [26,27]. Thiosulphate acts as a free radical se-
presence of sulphites and thiosulphates, the increase in the kinetics of
questering agent involved in the reactions sequence in which the
the cathodic process is more evident with greater water flow and with
overall reaction is articulated. In particular, they have been studied as
lower the pH. As a consequence, the corrosion rate of steel increases by
inhibitors of the formation of sulphates in sulphur abatement implants
an order of magnitude on moving from pH 7.2–5 (Fig. 10 and 6).
for combustion gases. The reaction rate depends upon the nature of the
Kappes et al. [29] reported that the corrosion of carbon steel in-
cation and the acidity [26,27]. The inhibiting effect depends on the
volves the formation of sulphur and H2S according to the reactions Eq.
sulphite/thiosulphate ratio; the addition of increasing amounts of sul-
(7) and Eq. (8) or, alternatively, Eq. (9) and Eq. (10).
phite over thiosulphate hinders the effect of the thiosulphates, and
gradually increases the oxygen-scavenging rate. When the sulphites/ S2 O32 − + 6H+ + 4e− = 2S 0 + 3H2 O E 0 = 0.465 VSHE (7)
thiosulphates ratio is around 1 or 2, the rate of this sequestering reac-
tion becomes negligible or extremely low (Fig. 5). Therefore, the action
of thiosulphate increases the levels of oxygen, even in sulphite-con- S° + 2H+ + 2e− = H2S E0 = 0.142 VSHE (8)
taining solutions, and may promote substantial increases in the corro- S2 O32 − + H+ = S0 + HSO3− (9)
sion rate of carbon steel, in proportion to the residual oxygen content.
In addition, the oxidation of sulphite to sulphate causes a decrease HSO3− + 7H+ + 6e− = H2 S + 3H2 O E 0 = 0.366 VSHE (10)
in pH (Fig. 7), according to the reaction Eq. (6). In fact, the second
dissociation constant of sulphuric acid is greater than that of sulphurous At the free corrosion potential, the cathodic current mainly promotes
acid. The HSO3− ion is not substantially dissociated in the pH range of the reduction of thiosulphate, in sulphite-free waters. The formation of
the water considered in this research. Upon reaction, it is replaced by H2S is catalysed by the metal surface and otherwise it does not occur in
the sulphate ion, which is completely dissociated under these condi- bulk solution. It leads to the formation of a black film of iron sulphide,
tions. In the experiments using solutions without thiosulphates, the pH mainly mackinawite [22,23,29–31].
was observed to rapidly reach of 3.4 after all the sulphite (300 ppm) The formation of scale, associated with high corrosion rates, is re-
had reacted. In this manner, the acid attack of steel can occur. Thio- ported even in saline waters without thiosulphates, but containing an
sulphate ions tend to block the oxygen-scavenging reaction and hence, excess of sulphites, compared to the values usually added to control
they stabilize the pH, slowing down the production of acidity. However, oxygen-induced corrosion [32–35]. This reaction (Eq. (10) therefore
corrosive attack, supported by cathodic oxygen reduction can still represents an important process in waters where sulphites reach re-
occur. The pH tends to increase if oxygen entry is prevented during the levant concentrations, far greater than those of thiosulphates. The ag-
active corrosion process (Fig. 7). gressiveness of these waters may not be related only to the content of
thiosulphate, but also to that of sulphite.
4.3. Effect of thiosulphate on general corrosion, in the absence of oxygen The formation of hydrogen sulphide on the metal surface can sti-
mulate the cathodic process that produces hydrogen through two dif-
In de-aerated neutral solutions devoid of particularly aggressive ferent mechanisms. On the one hand, it can increase the availability of
species, the corrosion rate of carbon steel is practically negligible. The the hydrogen ion due to its dissociation on the metal surface, according
addition of sulphites and thiosulphates, at the levels considered in this to the reaction:
research work, results in an increase in the corrosion rate of about one 2H2S → H+ + HS− (11)
order of magnitude (Fig. 15). This attack is enhanced by the rotational
speed of the electrode. At 1000 rpm, the corrosion rate is almost double On the other hand, hydrogen evolution can occur owing to the direct
compared to that observed under stagnant conditions (Fig. 11). reduction of hydrogen sulphide [24,25]:
Ezuber [23] and Kappes et al. [24] studied the corrosion of carbon
2H2S + 2e− → H2 + 2HS− (12)
steels in solutions containing only thiosulphates, at pH levels compar-
able or lower than the solutions considered in this research work. The The mackinawite scale is conductive, poorly protective and it is a great
results of these studies highlighted that the corrosion rate of carbon cathode for hydrogen production, further stimulating the corrosion of
steel increases with thiosulphate concentration. These ions modify the the metal due to galvanic coupling [23,28,29]. The scale formed during

164
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

Fig. 16. Morphology of the corrosion products (solutions containing 3000 ppm SO42−, 300 ppm SO32−, 150 ppm S2O32−) as a function of pH and exposure time.

our laboratory experiments is fragile and multi-layered. Underneath the under conditions that promote high corrosion rates [25]. The weight
scale, the metallographic structure of the steel can be evidenced, with loss experiments (Fig. 7) substantially confirm this mechanism. The
the formation of localized attack sites (Fig. 16) after long exposure reactions depicted in Eq. (7), Eq. (8), Eq. (9) and Eq. (10) are favoured
times. The sulphur layer favours the formation of H2S on the metal in acidic solutions and become more aggressiveness as the pH di-
surface according to the reaction [25]: minishes. In particular, the corrosion rate significantly increases on
moving from pH 7.2 to 5, due to increases in the kinetics of the cathodic
(x − 1) Fe + Sy − 1 S 2 − + 2H+ = (x − 1) FeS + H2 S + Sy − x (13)
processes, and are associated with the more rapid growth of corrosion
The stability diagram for various sulphur compounds confirms that the scale (Figs. 6, 16 and 18).
free corrosion potential of steel falls inside the stability region of H2S
[22,36] (Fig. 17). The concentration of hydrogen sulphide on the sur- 5. Conclusions
face of steel in acidic solutions containing thiosulphates has been
evaluated by Kappes [29,30]. The author points out that the corrosion The corrosion of carbon steel in neutral or weakly acid solutions
rate of carbon steel in solutions containing thiosulphates is higher than containing sulphites and thiosulphates in significant concentrations
that found in the same solution devoid of these ions, but saturated with occur with different mechanisms. Thiosulphates can inhibit the reac-
an H2S partial pressure equal to that estimated on the surface in the tions between oxygen and sulphites and promote a substantial increase
presence of thiosulphates. in the corrosion rate, in proportion to the residual oxygen content. On
The reduction of thiosulphate, firstly to sulphur and then to hy- the other hand, they control the acidification resulting from the reac-
drogen sulphide, and that of sulphite, both reduce the concentration tions of sulphites with oxygen, which may, in turn, induce acid corro-
hydrogen ions and tend to increase pH on the metal surface, especially sion.
In anaerobic solutions, both sulphites and thiosulphates modify the
characteristic anodic and cathodic potentiodynamic curves, with a

Fig. 17. Stability regions of the oxidation species of sulphur [37], and the range (in red)
of the corrosion potential of steel, measured in solutions containing 3000 ppm SO42−,
300 ppm SO32−, 150 ppm S2O62−. Fig. 18. Dependence of current density on pH and electrode rotational speed.

165
M. Cabrini et al. Corrosion Science 135 (2018) 158–166

marked effect on the cathodic process at potentials approaching the free study, Corros. Sci. 39 (1996) 913–933.
corrosion potential, and under conditions of water flow. [15] P. Marcus, Surface science approach of corrosion phenomena, Electrochim. Acta 43
(1998) 109–118.
In the absence of oxygen, the corrosion process leads to the for- [16] D. Tromans, L. Frederick, Effect of thiosulfate on crevice corrosion of stainless
mation of thick scales of mackinawite. The iron sulphide forms as a steels, Nace 40 (1984) 633–639.
result of the reduction of thiosulphate and sulphite, leading to the [17] Y.-S. Choi, J.-G. Kim, Investigation of passivity and its breakdown on Fe3Al-Cr-Mo
intermetallics in thiosulfate-chloride solution, Mat. Sci. Eng. A 333 (2002) 336–342.
formation of hydrogen sulphide and the precipitation of the corrosion [18] Y.M. Liou, S.Y. Chiu, C.L. Lee, H.C. Shih, Electrochemical pitting behaviour of type
scale. 321 stainless steel in sulfide-containing chloride solutions, J. Appl. Electrochem. 29
(1999) 1377–1381.
[19] R. Newman, H. Isaacs, B. Alman, Effects of sulfur compounds on the pitting beha-
Acknowledgment vior of type 304 stainless steel in near-neutral chloride solutions, Corrosion 38
(1982) 261–264.
This research project was made possible by funding from the ITALY® [20] A. Garner, Thiosulfate corrosion in paper-machine white water, Corrosion 41
(1985) 587–591.
PROJECT of the University of Bergamo.
[21] R.C. Newman, Pitting of stainless alloys in sulfate solutions containing thiosulfate
ions, Corrosion 41 (1985) 450–453.
References [22] M. Cabrini, M. Favilla, S. Lorenzi, B. Tarquini, T. Pastore, R. Perini, Corrosione Da
Tiosolfati Negli Impianti Geotermici, Pietro Pedeferri e la scuola di corrosione e
protezione dei materiali al Politecnico di Milano, Milano, 2013.
[1] A. Baldacci, M. Mannari, F. Sansone, Greening of geothermal power: an innovative [23] N. Sridhar, D. Dunn, A. Anderko, M. Lencka, H.U. Schutt, Effects of water and gas
technology for abatement of hydrogen sulphide and mercury emission, Proceedings compositions on the internal corrosion of gas pipelines – modeling and experi-
World Geothermal Congress, Antalya, Turchia, 2005. mental studies, Corrosion 57 (2001) 221–235.
[2] M. Cabrini, S. Lorenzi, T. Pastore, Corrosion behavior of carbon steel in acidic so- [24] P.W. Bolmer, Polarization of iron in H2S.NaHS buffers, Corrosion 21 (1965) 69–75.
lutions containing thiosulfate and sulfiteions, Metallurgia Italiana 106 (11–12) [25] Y. Zheng, B. Brown, S. Nesic, Electrochemical study and modeling of H2S corrosion
(2014) 3–10. of mild steel, Corrosion 70 (2013) 351–365.
[3] V. Duchi, A. Minissale, M. Manganelli, Chemical composition of natural deep and [26] R.K. Ulrich, G.T. Rochelle, R.E. Prada, Enhanced oxygen absorption into bisulphite
shallow hydrothermal fluids in the Larderello geothermal fields, J. Volcanol. solutions containing transition metal ion catalysts, Chem. Eng. Sci. 41 (1986)
Geotherm. Res. 49 (1992) 313–328. 2183–2191.
[4] F. Sabatelli, M. Mannari, R. Parri, Hydrogen sulfide and mercury abatement: de- [27] J. Mo, Z. Wu, C. Cheng, B. Guan, W. Zhao, Oxidation inhibition of sulfite in dual
velopment and successful operation of AMIS technology, GRC Trans. 33 (2009) alkali flue gas desulfurization system, J. Environ. Sci. 19 (2007) 226–231.
343–347. [28] H.M. Ezuber, Influence of temperature and thiosulfate on the corrosion behaviour
[5] D. Macdonald, Critical Issues in the Use of Metals and Alloys in Sulphur-Containing of steel in chloride solutions saturated in CO2, Mater. Des. 30 (2009) 3420–3427.
Aqueous Systems, (1993). [29] M. Kappes, G.S. Frankel, N. Sridhar, R.M. Carranza, Corrosion behavior of carbon
[6] H. Horowitz, Chemical studies of polythionic acid stress corrosion cracking, Corros. steel in acidified, thiosulfate-containing brines, Corrosion 10 (2012) 872–884.
Sci. 23 (1983) 353–362. [30] M. Kappes, G. S.Frankel, N. Sridhar, R.M. Carranza, Reaction paths of thiosulfate
[7] R.C. Newman, K. Sieradzki, H.S. Isaacs, Stress corrosion cracking of sensitized type during corrosion of carbon steel in acidified brines, J. Electrochem. Soc. (2012)
304 stainless steel in thiosulfate solutions, Metall. Trans. A 13 (1982) 2015–2026. C195–C204.
[8] E.A. Abd El Meguid, N.A. Mahmoud, S.S. Abd El Rehim, The effect of some sulphur [31] J. Banas, U. Lelek-Borkowska, B. Mazurkiewicz, W. Solarski, Effect of CO2 and H2S
compounds on the pitting corrosion of type 304 SS Mater, Chem. Phys. 63 (2000) on the composition and stability of passive film on iron alloys in geothermal water,
63–74. Electrochim. Acta 152 (2007) 5704–5714.
[9] J.O. Park, M. Verhoff, R. Alkire, Microscopic behavior of single corrosion pits: the [32] T. Hemmingsen, H. Vangdal, T. Våland, Formation of ferrous sulfide film from
effect of thiosulfate on corrosion of stainless steel in NaCl, Electrochim. Acta 42 sulfite on steel under anaerobic conditions, Corrosion 48 (1992) 475–481.
(1997) 3281–3291. [33] T. Hemmingsen, The electrochemical reaction of sulphur-oxigen compounds – part
[10] H.-S. Kuo, H. Chang, W.-T. Tsai, The corrosion behavior of AISI 310 stainless steel in II. Voltammetric investigation performed on platinum, Electrochim. Acta 37 (1992)
thiosulfate ion containing saturated ammonium chrolide solution, Corros. Sci 41 2785–2790.
(1999) 669–684. [34] T. Hemmingsen, H. Lima, Electrochemical and optical studies of sulphide,
[11] T. Laitinen, Localized corrosion of stainless steel in chloride, sulfate and thiosulfate Electrochim. Acta 43 (1998) 35–40.
containing environments, Corros. Sci. 42 (2000) 421–441. [35] T. Hemmingsen, T. Våland, An investigation of the effect of bisulphite/sulphite on
[12] A.I. Almarshad, D. Jamal, Electrochemical investigations of pitting corrosion be- steel under anaerobic conditions, 11th Scandinavian Corrosion Congress,
haviour of type UNS S31603 stainless steel in thiosulfate-chloride environment, J. Stavanger, 1989.
Appl. Electrochem. 34 (2004) 67–70. [36] P. Marcus, E. Protopopoff, Thermodynamics of thiosulfate reduction on surfaces of
[13] W.T. Tsai, M.J. Sheu, J.T. Lee, The stress corrosion crack growth rate in sensitized iron, nickel and chromium in water at 25 and 300 °C, Corros. Sci. 39 (1997)
alloy 600 in thiosulfate solution, Corros. Sci. 38 (1996) 33–45. 1741–1752.
[14] C. Duret-Thual, D. Costa, W. Yang, P. Marcus, The role of thiosulfates in the pitting
corrosion of Fe-17Cr alloys in neutral chloride solution: electrochemical and XPS

166