Received: 19 October 2021 | Accepted: 25 February 2022

DOI: 10.1002/maco.202112871

ARTICLE

Marine atmospheric corrosion of carbon steel in the

tropical microclimate of Port Louis

Yashwantraj Seechurn1,2 | Baboo Y. R. Surnam1 | Julian A. Wharton2

1
Mechanical and Production Engineering
Department, Faculty of Engineering, Abstract
University of Mauritius, Réduit, Moka, The atmospheric corrosion performance of S235 carbon steel is assessed when
Mauritius
2
exposed to the tropical/marine aerosol pollutants of Port Louis, Mauritius. Port
School of Engineering, National Centre
of Advanced Tribology at Southampton Louis is situated on the northwest coast and has a distinct microclimate, leading
(nCATS), University of Southampton, to major atmospheric corrosion concerns. In addition, sulphur dioxide emissions
Southampton, Hampshire, UK
from three heavy oil‐fired power stations complicate corrosion prediction within
Correspondence the region. Mass‐loss analysis shows distinctive corrosion progression behaviours
Yashwantraj Seechurn, Mechanical and linked to the environmental and geographical variables. The formation of porous
Production Engineering Department,
surface layers leads to a continuous increase in metal dissolution during the
Faculty of Engineering, University of
Mauritius, Réduit 80837, Mauritius. electrochemical process. The surface analysis reveals lepidocrocite (γ ‐FeOOH)
Email: [email protected] and goethite (α‐FeOOH) as the predominant rust phases, with low percentages
of magnetite (Fe3 O4 ), akageneite (β ‐FeOOH) and jarosite (KFe3 (OH)6 (SO4 )2) .
A key insight in the S235 atmospheric corrosion involves two stages: the early
rapid corrosion stage followed by a slow stage of linear or power‐law corrosion
kinetics due to the formation of stable rust phases.

KEYWORDS
atmospheric corrosion, carbon steel, rust, SEM, weight loss, X‐ray diffraction

1 | INTRODUCTION complex, airborne chloride levels and sulphur dioxide are

key to establishing corrosivity estimates. The actual
Atmospheric corrosion is known to be severe in coastal corrosion rates in coastal environments can be difficult
zones worldwide; however, the mechanism of corrosion in to predict due to variable chloride deposition rates.
atmospheres associated with high humidity, airborne Likewise, sulphur dioxide plays an important role in
natural marine aerosols and heavy industry contaminants atmospheric corrosion of steel infrastructure in urban and
is not fully understood.[1] Carbon steels are by far the most industrial areas, leading to the formation of sulphur
important, multifunctional and most adaptable structural compounds, such as rozenite, melanterite and jarosite.[2,3]
material. Corrosion management strategies for carbon Mauritius is a small island (land area: 1865 km2) with
steels are essential to maximise the resilience of all 330 km of coastline and is situated 865 km east of
manners of sustainable infrastructure development, espe- Madagascar (Figure 1). It has a tropical climate with
cially for countries with emerging/small economies such abundant rainfall (2010 mm annually), high humidity
as Mauritius. Characterisation of the dominant environ- levels (mainly above 80%) and only two seasons: winter
mental parameters, such as the temperature‐humidity and summer.[4] The island is also under the influence of
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
© 2022 The Authors. Materials and Corrosion published by Wiley-VCH GmbH.

1474 | www.matcorr.com Materials and Corrosion. 2022;73:1474–1489.

 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1475

steady south‐easterly trade winds (Figure 2) carrying pollution from power industries makes the atmospheric
natural airborne sea aerosols inland. All these environ- conditions in PL distinct from the conditions prevailing
mental conditions combined, lead to atmospheric corro- over the rest of the island. Major salinity differences may
sion with deleterious effects on infrastructure, particu- also exist between the shores of one side of a country and
larly in the coastal regions of the island. the other side as each region varies in terms of its
Coastal atmospheric corrosion is of particular concern characteristics pertaining to the corrosion behaviour of
for Port Louis (PL) (20.2°S, 57.5°E), the harbour and materials.[7] In Mexico, for example, the corrosion rate for
capital city on the north‐west coast, with an industrial the northern coast is 362.4 µm annum−1 (µm a–1), which is
zone comprising three power plants, namely Saint‐Louis, significantly higher than the corrosion rate of 91.3 µm a–1
Fort‐Victoria and Fort George (FG), of 100, 107 and obtained for the southern coast.[8]
134 MW total effective capacity, respectively.[5] The power In addition, PL is situated on the leeward side of the
stations generate electricity from the combustion of heavy Moka mountain range (Figure 2), which contributes to its
fuel oil and are large emitters of gases, such as SO2 and microclimate, hindering the prevailing rain patterns, caus-
NOx among others. In 2016, stack gases of 16157 mg m–3 ing the land to remain dry and producing less moisture in
NO2 and 6537 mg m–3 SO2 contents were released from the air.[9] Atmospheric corrosion data from a previous
FG.[6] The presence of natural atmospheric marine study undertaken at a specific location within PL have
aerosols coupled with anthropogenic contaminants/ shown that the region is highly corrosive.[10] However,
given the spatial variability of atmospheric conditions in
PL, corrosivity estimates determined from a single
exposure location are not necessarily representative of
the wider region. Moreover, the environmental conditions
prevailing at other island locations are not similar to those
of PL. Therefore, this study uses additional outdoor
exposures to obtain a more comprehensive corrosivity
assessment of PL, for more effective long‐term corrosion
prediction, by determining the nature of rust layers and
the mechanism of transformation from one phase to
another.[11] Similar investigations have often utilised SEM
and XRD for the analysis and characterisation of corrosion
products on metallic surfaces.[12–15] However, the data
FIGURE 1 Map showing Mauritius in the Indian Ocean available on the specific morphology of rust phases, with

FIGURE 2 Map of Mauritius and Port Louis showing critical locations and test sites
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1476 | SEECHURN ET AL.

respect to identified chemical components within the TABLE 1 Chemical composition of the S235 carbon steel
corroded layers is limited. As such, morphologies associ- Steel C Mn P S Si Cu
ated with rust phases need to be justified by determining
S235 0.17 1.4 0.04 0.04 0.05 0.55
the specific formation mechanism in the prevailing
environment.
150 mm × 100 mm × 3 mm, were exposed to the atmo-
sphere in appropriate racks according to ISO 8565[18] from
2 | MATERIALS A ND METHODS/ June 2018 to June 2019. The expected main corroding
EXPERIMENTAL source at PL being the sea salt aerosol, all the racks were
facing the ocean, while the rack at MC was oriented
2.1 | Field exposure towards the north (Supplementary Information: Appendix).
Before exposure, the metal plates were immersed in a
To determine the atmospheric corrosion performance of solution of 10% wt./vol. hydrochloric acid (HCl specific
S235 carbon steel, three sites (Figure 2) were selected in PL, gravity 1.19) to remove the mill scale and oxides from the
namely Fort George (FG), Rajiv Gandhi (RG) and Mutual surface. The pickled plates were rinsed thoroughly with
Aid (MA), having elevations above sea level of 8, 41 and distilled water, hot air‐dried using a specimen dryer and
11 m, respectively. Moreover, a control atmosphere (Med- then stored in a desiccator. Removal of quintuplicate plates
ine Camp de Masque—MC), around 283 m above sea level, were carried out at each test site after 2, 5, 8 and 11 months
was also considered as shown in Figure 2. FG is 50 m of exposure. From the five plates at each time point, two
downwind of a heavy‐oil power plant chimney and 350 m were used for corrosion product characterisation (optical/
inland from the coast. It has the characteristics of a marine‐ electron microscopy and X‐ray diffraction [XRD]) and the
heavy industrial environment. Sulphur dioxide (SO2) remaining three were taken for mass‐loss analysis as per
emissions from the combustion of heavy fuel oil are likely ISO 8407.[19] Each plate was weighed to three decimal
to be the main contributor to corrosion in this area, with places using the Kern precision balance PB 400‐3.
sulphate aerosols also expected to naturally occur due to
the photooxidation of dimethylsulphide present in seawater
(accounting for approximately 50% of the total biogenic 2.2 | Corrosion rate determination
sulphur flux to the atmosphere, and linked to diurnal and
annual variations).[16] MA, situated in the city centre, is The corrosion rate (rcorr ) was calculated as per ISO 9226
203 m from the coastline and pollutants in the atmosphere using the following Equation (1): [20]
are representative of a marine‐urban environment. RG is
∆m
situated away from the city centre but close to an industrial rcorr = , (1)
At
zone, comprising a power station and other industries.
Data collected from this site should give an indication of where ∆m is the mass loss in g, A is the surface area in
the influence of marine salts on metal surfaces exposed in m2, and t is the exposure time in years (a).
the outskirts of PL and at a distance (896 m) further away It is well known that rcorr does not remain constant
from the coast. MC is a rural area located 17 km from the after the initial stages of surface degradation since the
east coast, closer to the island centre and upwind from any accumulation of corrosion products with time leads to a
industrial pollutants. The prevailing south‐easterly trade decrease in corrosion rate, depending on the nature of the
winds mean that the MC atmosphere will not be free from rust phases. However, rcorr , which is often determined
exposure to either chlorides (Cl–) or SO2. from first‐year data, can also be used to estimate the
Deposition rates of Cl– and SO2 levels in the atmosphere corrosion deterioration (plate thickness loss) for long‐term
were determined by the wet candle method and lead oxide exposures.[21] According to ISO 9224, ‘total attack’ (D),
sulfation plates (shown in Appendix), respectively, as per assessed using Equation (2), is useful for approximating
ISO 9225.[17] Two monthly measurements, each involving a the progress of corrosion damage.[22] The exponent b is a
single wet candle and two sulfation plates, were taken in complex association of the type of atmosphere, the
each season (winter and summer) to determine seasonal variation of climatic factors such as time‐of‐wetness and
variation. The Cl– and SO2 concentrations obtained using any other prevailing conditions at the exposure site.
the mercuric thiocyanate method and a turbidimetric Equation (2) is similar to the established expression of the
method, respectively, were validated using a standard form M = at b, where M is the mass loss, t is the number
solution.[17] Bare S235 carbon steel plates, with the of exposure years with a and b being constants.[23] A b
composition shown in Table 1, having the surface value around 0.5 suggests that the atmosphere is mildly
roughness measure Ra of 4.2 µm and dimensions of polluted, whereas b greater than 1 represents a coastal
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1477

region irrespective of the concentration of chloride.

Specific values of rcorr and b obtained for each material
and the associated environment could be used to
determine D for extended exposures.[24–26]
D = rcorr t b, (2)

where t represents years of exposure (a), rcorr is the

corrosion rate in g m–2 a–1 and b is a time exponent
indicative of the metal and environment.
In addition to corrosion rate, climate and pollution
parameters, such as airborne salinity and SO2 levels,
were also used in categorising the atmospheric environ-
F I G U R E 3 Mean relative humidity versus time for the Port
ments based on a corrosivity classification method
Louis (PL) and Medine Camp de Masque (MC) locations[28]
provided in ISO 9223.[27] The weather information used
in this study, such as monthly rainfall, wind direction
and mean monthly wind speed, relative humidity (RH)
and temperature, are all quality controlled observed data
provided by the Mauritius Meteorological Services and
also derived from monthly Meteorological Observations
and Climatological Summaries.[28] A parameter often
used instead of RH is time‐of‐wetness (TOW), defined in
ISO 9223 as the time during which RH exceeds 80% and
temperature is above 0°C, which frequently results in a
wet surface.[27] TOW is, therefore, a useful indicator of
atmospheric corrosivity since a high TOW would imply a
long‐lasting effect of the electrolyte on the metal surface.

F I G U R E 4 Rainfall versus time for the Port Louis (PL) and

2.3 | Corrosion product characterisation Medine Camp de Masque (MC) locations[28]

The Zeiss Merlin scanning electron microscope (SEM)

with an energy dispersive X‐ray spectrometer (EDS) was considered a dry region compared to MC, with an
used for examination of the surface morphology and the average RH value of 68% and always lower than MC
Bruker D2 Phaser XRD machine was used as a material during the whole exposure period. Furthermore, Figure 4
characterisation tool. Loose rust deposits were collected shows that precipitation was also lower at PL, with a
from the outer corroded surface and typically in marked increase for both regions on December 2018 and
powdered form for XRD analysis. The diffractometer is April 2019. A washing effect of the rain was likely to be
equipped with a Cu X‐ray tube and Lynxeye (1D mode) significant during these periods, thereby preventing the
detector. The analysis was carried out from 10° to 60° accumulation of corrosion products that would otherwise
with a step width of 0.03° and data were collected for 0.5 hinder corrosion progression. Figure 5 shows that PL is a
s at each step. The diffraction pattern of each sample was hotter region compared to MC. However, variation of
compared to a database containing reference patterns to temperature throughout the year is the same for both
identify the phases that are present. regions with the period January 2019 to March 2019
being the hottest.
Table 2 shows the average Cl– and SO2 deposition
3 | RESUL TS rates at each site, the TOW for MC and PL and their
corresponding ISO categories. The SO2 level is at the
3.1 | Variation of climatic factors lowest category P0 for all sites, even at FG, despite its
proximity to a power station releasing SO2 gases into the
From Figure 3, it is evident that RH was highest during atmosphere. The Cl– deposition rates vary significantly
the summer season (November 2018 and April 2019) for between the various sites. In PL, the highest salinity rate
both the PL and MC locations. However, PL can be is 57.25 mg m–2 day–1 at RG, which is close to the upper
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1478 | SEECHURN ET AL.

limit of 60 mg m–2 day–1 for the S1 category (3 < S ≤ 60). FG texture of the skyward faces of plates exposed at FG, MA,
is in the mid‐category while MA has a low value. The mid‐ RG (the three PL locations) and MC. Following the first
island atmosphere (MC) has the highest TOW and a removal after 2 months, the appearance of rusted carbon
significant Cl– content (higher than FG and MA). The latter steel differs at the various locations, with notable orange
confirms that variations in Cl– deposition rate, as would be and brown rust colours, consistent with the formation of
expected for a small island, cannot be differentiated by lepidocrocite (γ − FeOOH) and goethite (α − FeOOH)
distance from the shoreline in this study. Factors such as rust phases, respectively.[3,32–34] The orange colour at MC
the prevailing weather patterns and orography may is strong evidence of the presence of lepidocrocite, which
predominately affect the Cl– deposition rate around the turns from orange to reddish‐brown as the rust layer
island. In comparison, in an atmospheric corrosion study in thickens.[35,36] The lighter colour of the rust at FG is
the tropical climate of Ecuador, considered one of the most characteristic of a saline environment and there is a
corrosive in South America, six atmospheric sites investi- uniform texture and colour on the top faces across all sites,
gated in the marine zone of Manabí were categorised as with rougher surface textures observed for all the 11‐month
being S2 (60 < S ≤ 300) with a TOW of 4889 h a−1.[29] In corroded plates due to extensive corrosion damage.[14] In
another study in Male, the capital of the Maldives and a contrast, the rear faces (shown in Figure 8) are of
tropical island in the Indian Ocean, a high Cl– deposition heterogeneous appearance, with a reddish‐brown colour
rate of S3 (300 < S ≤ 1500) was observed, with a TOW of evident of the stable haematite phase (α − Fe2 O3 ),[33,37,38]
3067 h a–1.[30] and some dark/grey patches. The latter is an indication of
the severity of corrosion on the groundward faces, which
are not subjected to the washing effect of the rain and
3.2 | Visual appearance of corroded therefore have thicker rust layers.
specimens

The colouration of rust layers varies with material, time 3.3 | Variation of corrosion rate
and according to the corrosivity of the atmosphere.[14,31]
Figures 6 and 7 show the marked differences in colour/ The mass loss analysis was performed on metal plates
subjected to 2, 5, 8 and at least 11 months of exposure such
that the results are representative of the first year corrosion
loss. The corrosion rate (rcorr ) was calculated for each site
at specific time intervals using Equation (1) and then
substituted in Equation (2) to obtain the total attack (D).
Table 3 lists the form of the power equation of D versus t
and the corresponding b exponent and R2 for each site. The
value of b is suggestive of the corrosion products evolution;
between 0.5 and 1.0 indicates a low protection performance
of the corrosion layer, while less than 0.5 indicates
corrosion product growth according to a parabolic function,
due to corrosion products blocking aggressive species from
reaching the underlying metal surface.[13] Therefore, based
on the b value, the rust layer formed at each test site can be
F I G U R E 5 Mean temperature versus time for the Port Louis ranked from more protective to less protective in the
(PL) and Medine Camp de Masque (MC) locations[28] following order: FG–RG–MA–MC.

TABLE 2 Assessment of environmental conditions at the three Port Louis locations and the MC site
TOW Cl– level SO2 level ISO 9223 Class
Site (h a–1) (mg m–2 day–1) (mg m–2 day–1) TOW Cl– SO2
RG 1798 57.25 1.43 ± 0.54 τ3 (250 < τ ≤ 2500) S1 (3 < S ≤ 60) P0 (P ≤ 4)
FG 41.56 1.90 ± 0.09
MA 14.49 1.36 ± 0.33
MC 3272 51.37 1.55 ± 0.54 τ4 (250 < τ ≤ 5500)
Abbreviations: FG, Fort George; MA, Mutual Aid; MC, Medine Camp de Masque; RG, Rajiv Gandhi; TOW, time‐of‐wetness.
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1479

F I G U R E 6 Two‐month exposed skyward surfaces (from left to right: FG, MA, RG and MC). Dimensions: 150 × 100 mm. FG, Fort
George; MA, Mutual Aid; MC, Medine Camp de Masque; RG, Rajiv Gandhi [Color figure can be viewed at wileyonlinelibrary.com]

F I G U R E 7 Eleven‐month exposed skyward surfaces (from left to right: FG, MA, RG and MC). Dimensions: 150 × 100 mm. FG, Fort
George; MA, Mutual Aid; MC, Medine Camp de Masque; RG, Rajiv Gandhi [Color figure can be viewed at wileyonlinelibrary.com]

F I G U R E 8 Eleven‐month exposed groundward surfaces (from left to right: FG, MA, RG and MC). Dimensions: 150 × 100 mm. FG, Fort
George; MA, Mutual Aid; MC, Medine Camp de Masque; RG, Rajiv Gandhi [Color figure can be viewed at wileyonlinelibrary.com]

Table 4 shows the yearly corrosion rates, as calculated is, between 10 and 200 g m−2 a−1. In the present study,
using Equation (1), for the different test sites and the the lowest corrosion rate is 77.6 g m–2 a–1 and the highest
corresponding corrosivity categories as per ISO 9223.[27] being 189 g m–2 a–1, is close to the upper limit for this
All sites, apart from MC, are in the C2 category that category.
represents low corrosivity. Overall, a broad range of Figure 9 (D vs. t ) shows that the data fits a linear
corrosion rates are associated with the C2 category, that function very well for all sites, except FG where the
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1480 | SEECHURN ET AL.

TABLE 3 Mass loss data at the three Port Louis locations and the MC site
Time interval (days) Mass loss (g)
Site 1st 2nd 3rd 4th 1st 2nd 3rd 4th Equation R2
MA 66.9 151.9 243.76 348.83 1.16 1.55 1.90 2.33 D = 51.04t 0.4162 0.99
MC 65.6 155.6 246.03 346.01 2.47 3.86 5.10 6.82 D= 148.77t 0.5972 0.99
RG 61.9 153.9 245.89 350.93 1.75 2.48 2.92 3.65 D= 79.35t 0.4092 0.99
FG 61.8 152.7 244.66 349.78 3.10 4.83 5.07 5.70 D= 132.26t 0.3452 0.92
Abbreviations: FG, Fort George; MA, Mutual Aid; MC, Medine Camp de Masque; RG, Rajiv Gandhi.

T A B L E 4 Corrosion kinetics at the three Port Louis locations 3.4 | SEM/XRD

and the MC site
Total exposure Mass rcorr Corrosivity The specific rust phases that form in the diverse
Site (years) loss (g) (g m–2 a–1) category environmental conditions are also highly indicative of
MA 0.9531 2.33 77.6 C2 atmospheric corrosivity. Considering the main rust
phases (lepidocrocite, goethite, magnetite and akaga-
MC 0.9454 6.82 229 C3
neite), the strongest peaks from the XRD spectra
RG 0.9588 3.65 121 C2 shown in Figure 10 relate to the goethite and
FG 0.9557 5.70 189 C2 lepidocrocite phases. Magnetite (Fe3 O4 ) and akagen-
Abbreviations: FG, Fort George; MA, Mutual Aid; MC, Medine Camp de
eite (β − FeOOH) are present, but in small quantities,
Masque; RG, Rajiv Gandhi. as observed from the small peaks. Jarosite—
KFe3 (OH)6 (SO4 )2 , rozenite (FeSO4 ∙4H2 O) and melan-
terite (FeSO4 ∙7H2 O) can also be observed from the XRD
power‐law function shown in Table 3 is favoured. There spectra, confirming the presence of sulphate.[40,41]
is a marked corrosion loss in the first few months of While most of the corrosion product surface morphol-
exposure at FG, followed by a decrease in corrosion rate ogies were straightforward to identify and were just
until the end of the exposure period. This suggests that a various forms of lepidocrocite and goethite (Figure 11),
coherent corrosion product layer forms rapidly, hinder- there were some unique morphologies of the minor
ing further permeation of aggressive species to the constituent phases (not reported in previous studies).
underlying metal surface with time. Whereas at RG, MA They were mainly globular and tubular formations,
and MC, the corrosion rate (D / t ) is lower after two which are features of adherent rust.[38]
months, but then remains constant until the end of Lepidocrocite appeared as sandy crystals, bird's nests
exposure indicating that the rust layer was a more (Figure 11a), flowery structures with thin laminas or
porous/permeable barrier and the underlying steel petals and clumps of grass structure (Figure 11b). At all
substrate remained exposed. Given the urban environ- four sites, goethite occurred as whiskers forming on the
mental conditions at MA, effects of NOx (mainly periphery of globules (Figure 11c) and bar‐shaped
originating from vehicle emissions) predominate.[39] laminar structures with thorny appearance (Figure 11d).
However, the latter show little involvement in acceler- The spinel form of magnetite occurred as a black sphere
ating the corrosion of carbon steel in this area, with a or crystal‐like formations with isolated areas of a whitish
low mass loss at the first removal and subsequently a colour (Figure 11e,f). Akaganeite was also found in
low corrosion rate. Due to a high TOW at MC, the various tubular formations (cigar‐shaped) as shown in
reduced species can penetrate the permeable rust layer Figure 11g,h. Lath structures, typical of jarosite mor-
easily. There most probably is the formation of soluble phology, could be observed at FG as the EDS spectrum
corrosion products, which do not interfere with redox also shows significant potassium and sulphur content
reactions (Equations 3 and 4) occurring in the wet/dry (Figure 12). At RG and MA, bird's nest and flowery
cycles, as corrosion proceeds faster. crystals formed in circular nets. The latter was quite
distant apart, with a relatively smooth surface at MA and
Fe → Fe2+ + 2e−, (3)
a noticeable percentage of Cl– and sulphur as shown in
the EDS spectrum (Figure 13). The formation of surface
O2 + 2H2 O + 4e− → 4OH−. (4) cracks, observed from the SEM images, is due to the use
15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1481

X‐ray diffraction spectra for 11‐month exposure at (a) Rajiv Gandhi, (b) Fort George, (c) Mutual Aid and (d) Medine Camp
|

(b)

(d)
corrosion) versus time [Color figure can be
F I G U R E 9 Comparison of corrosion
performance, total attack (the degree of

viewed at wileyonlinelibrary.com]
ET AL.

F I G U R E 10
(a)

(c)

de Masque
SEECHURN
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1482 | SEECHURN ET AL.

F I G U R E 11 Corrosion product surface morphologies after an 11‐month exposure at the three Port Louis locations and the Medine
Camp de Masque (MC) site: (a, b) lepidocrocite, (c, d) goethite, (e, f) magnetite, (g, h) akaganeite [Color figure can be viewed at
wileyonlinelibrary.com]

of the ultrahigh vacuum SEM and is also a characteristic 4 | DISCUSSION

of the rust layer formed. The presence of cracks results in
a less protective barrier to electrolyte flowing to the metal The atmospheric salt concentration is generally signifi-
substrate, even with the accumulation of corrosion cant for a few hundred metres from the coast before a
products on the surface. rapid decrease towards inland[14] and decreasing
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1483

F I G U R E 12 Jarosite morphology identified at Fort George from scanning electron microscope image and energy dispersive X‐ray
spectrometer spectrum [Color figure can be viewed at wileyonlinelibrary.com]

F I G U R E 13 Elemental composition of the rust layer at Mutual Aid [Color figure can be viewed at wileyonlinelibrary.com]

corrosion kinetics with distance from the sea is usually sea but on the downwind side. This explains the low Cl–
observed.[8] However, the complex combination of concentration at the various locations in PL, even when
location, exposed conditions, seasonality and the prevail- situated relatively close to the shoreline at FG. In this
ing weather patterns can markedly affect the corrosion study, the mid‐island atmosphere, MC, has a salinity
performance of bare S235 steel.[42] Figure 14 shows a plot level due to ocean generated sea aerosol,[43] which is
of the wind speed versus time for PL and the average Cl– higher than the two marine sites in PL: FG and MA.
deposition rates obtained from measurements carried out Therefore, the dependence of Cl– deposition level solely
in the winter and summer. The Cl– measurement periods based on distance from the shoreline is not correct given
coincide with the periods of high wind and low wind and that locations close to the sea could result in lower
the difference between the two sets of values is corrosion rates than sites further inland, due to the
significant. complexity of chloride distribution in a marine‐island
The wind direction in PL is generally west/north‐ atmosphere.[44] This has been reported in various studies,
west, that is, the wind direction is from land to sea. whereby influencing factors, such as obstacles and
Therefore, the exposed specimen surfaces are facing the surface washing from rainwater have been noted.[21,45,46]
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1484 | SEECHURN ET AL.

Furthermore, higher Cl– deposition rate should not deceleration of corrosion at these sites is not pronounced
always be associated with a higher degree of corrosion due to the porosity of the oxide film.
since D can be higher in an atmosphere with lower From Figure 9, a power relation between D and t is
salinity as shown in the graph of D versus Cl– level observed at FG. The power law is valid in cases of
(Figure 15a), plotted from the results obtained in the diffusion‐controlled mechanisms, whereby the corrosion
current study. Similarly, D does not necessarily increase layer takes a consolidated form and species permeate
with SO2 level (Figure 15b). This implies that Cl– and SO2 through this layer to react with the steel substrate at the
do not primarily govern an increasing trend of corrosion metal/rust interface.[1] In contrast, the rust layer within
acceleration within the regions investigated. For the the first year of exposure at all the other sites grows at a
present study, airborne salinity is relatively uniform from constant rate, as shown by the linear D versus t
the coast to far inland, and RH is the only factor relationship (Figure 9). Nevertheless, there is a signifi-
significantly different between PL and MC. The higher cant difference between the corrosion rates of MA, RG
RH at MC leads to a predominance of the lepidocrocite and MC, which could be attributed to the different
rust phase at this site, as shown in the XRD spectrum climatic conditions existing at the test locations. Fluctu-
(Figure 10), which reduces the protective capacity of ations in RH and temperature give rise to the continu-
the corrosion layer.[31,47,48] This is in agreement with ously repeating wetting and drying cycles, which
the value of the power exponent b at MC, which is in the influence the corrosion rate.[49] The type, rate of
range 0.5–1.0, associated with low corrosion protection occurrence and length of these wet and dry periods have
performance. FG, MA and RG have b values below this a noticeable effect on the constitution and morphology of
range, but closer to the threshold of 0.5, implying that the corrosion products formed on the metal surface. RH
is known to increase corrosion rates and can be the main
factor in low salinity and nonpolluted regions. MC is in
the highest TOW category, with frequent condensation
aided by the porous nature of the surface, causing metal
surfaces to remain wet for a longer time. This explains
the higher yearly corrosion rate obtained for this site.
Furthermore, the formation of less adherent corrosion
products such as lepidocrocite in high humidity condi-
tions leads to enhanced diffusion of oxygen and water
through the rust layer, which is in accordance with data
obtained at other locations from previous studies in
Mauritius.[50] However, there is no indication of any
highly significant increase in corrosion kinetics during
F I G U R E 14 Effect of wind regime in Port Louis[28] the most humid months for all sites.

(a) (b)

F I G U R E 15 Effect of atmospheric
pollutants on corrosion: (a) D versus Cl− level
and (b) D versus SO2 level
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1485

Several factors, such as marine salts and gaseous (KFe3(OH)6(SO4)2) at FG, due to the presence of
pollutants may also contribute to an acceleration of the potassium in the emissions from the burning of fuel in
electrochemical process. Both Cl– and SO2 are major the heavy oil‐fired power plant.
contributors to atmospheric corrosion.[21,39,51] SO2 has
SO2 + H2 O + 1/2O2 → H2 SO4 , (5)
also been found to have a more damaging effect on
carbon steel than other metals such as copper.[39] In
4H2 SO4 + 4Fe + 2O2 → 4FeSO4 + 4H2 O. (6)
chloride‐rich atmospheres, the presence of a salt solution
on metal surfaces can cause condensation at a lower
humidity level, which would otherwise occur at 100% In addition, fragments of spinel magnetite were also
RH.[52] Corrosion due to deliquescence of sea salt found at FG. The latter shows a gradually decreasing
particles, mainly occurring as NaCl, deposited on metal corrosion rate after subsequent periods of exposure,
surfaces is very common in marine atmospheres.[53] indicating an increase in thickness of the rust layer,
However, the RH of PL is below the deliquescence which reduces the contact of the underlying surface with
relative humidity (DRH) for NaCl (75%) and is, therefore, electrolyte. This favours the oxidation of the existing rust
unable to induce deliquescence except for the minor phase instead, without an adequate supply of oxygen to
constituents, such as MgCl2 and CaCl2 with lower DRHs keep the same rate of dissolution of metal ions. Never-
of 33% and 20%, respectively.[54,55] Being hygroscopic, Cl– theless, the build‐up of corrosion products with time
and sulphate cause a prolonged TOW from an increase in would lead to the onset of cracks in the rust layer causing
moisture condensation. This explains the highest mass the corrosion rate to rise again.[50] This may lead to
loss at FG in the first few months of exposure (Figure 9), flaking of the outer rust layer, which is generally an
due to a higher SO2 concentration in the atmosphere at agglomeration of crystalline phases, such as lepidocrocite
the nearby power station, leading to corrosivity that is and goethite. The residual rust layer is mainly amor-
nearly at the level of C3. The beginning of exposure phous FeOOH, which is compact and may contain
coincides with the low rainfall period of the year magnetite and akaganeite. The latter forms under the
(Figure 4), causing minimal washing off of pollutants significant presence of Cl–, which acts as a catalyst in the
deposited on the metal surface of FG. Nevertheless, it process.[14] The presence of Cl– increases the electrolyte
should be noted the SO2 deposition measured at FG is conductivity, which may lead to the rupture of any
low, despite being downwind of the power plant, due to protective layer.[21,31]
the plume dispersing in the atmosphere without any A time profile schematic for the evolution of corrosion
significant deposition near ground level closer to the products formation under the region effects is shown in
stack. Aggressive environments give rise to more protec- Figure 16. Before exposure, the uncontaminated surface is
tive rust phases, especially under moderate SO2 condi- covered with a passivating oxide or hydroxide layer of
tions.[13] It is common to find sulphate as a constituent of thickness in the order of a few nanometres. A few weeks
the corrosion layer formed on exposure of steel to an later, a brown layer of lepidocrocite (γ − FeOOH) starts to
industrial atmosphere.[2,3] Ferrous sulphate is likely to appear, as an electrolyte film comprising of chlorides and
form if the SO2 deposits are not washed off by rainwater. sulphur dioxide from various sources sets on the metal
Given that the site at FG is very much exposed to surface. In the next stage of rust development, part of the
precipitation, the influence of SO2 is minimal and the lepidocrocite quickly transforms into the more stable
enduring electrolyte on the metal surface causes the only goethite (α − FeOOH) layer and the rusts grow as a
sparse formation of ferrous sulphate in small pits.[56] porous membrane on the surface. The permeable rust
The latter, if commonly present, leads to the rupture of layer, containing pores of up to 15 nm in diameter,[2]
even thicker corrosion layers formed in harsh atmo- allows the ingress of pollutants‐containing electrolytes to
spheres. It is known that sulphuric acid, which is an the substrate, causing corrosion to proceed at the same
intermediate in the formation of ferrous sulphate from rate. This porous layer separating the substrate and the
SO2 (Equations 5 and 6), reappears again as constant electrolyte is not able to prevent oxygen from migrating to
contact with water causes further hydrolysis of ferrous the underlying steel, which also explains the absence of a
sulphate, thereby causing continuous corrosion. In this consequent amount of the magnetite rust phase (Fe3O4).
study, the thin electrolyte layer on the metal surfaces In the absence of oxygen, water becomes the oxidising
exposed at the various sites undoubtedly comprises both agent in the formation of Fe3O4, in a reaction that also
Fe2+ and Fe3+ ions, in the form of metal complexes of produces gaseous hydrogen (Equation 7).[57] Hence, the
iron bonded to water molecules and possibly to the Cl– diffusion process is mainly absent in the corrosion
and sulphate ions present. This caused the precipitation occurring at the different sites, although to a lesser extent
of a small amount of the jarosite crystalline compound at FG. Chlorides in solution are attracted to Fe2+ being
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1486 | SEECHURN ET AL.

F I G U R E 16 Evolution of corrosion product formation [Color figure can be viewed at wileyonlinelibrary.com]

released through the pores, possibly forming iron oxy- penetrate the rust layer as there is insufficient sorption
chlorides (FeClO), with chloride also leading to some and precipitation for electrolyte formation, leaving a
akaganeite (β‐FeOOH) formation. The latter is present in relatively dry metal surface.
low percentages and, therefore, is not able to cause a Although at a slower rate compared to MC, the
significant rise in corrosion rate. corrosion loss at MA and RG also increases linearly with
time, indicating that the corrosion process persists during
3Fe + 4H2 O → Fe3 O4 + H2 . (7) the yearlong exposure, and the rust layer is not
protective. Due to the build‐up of corrosion products,
The Cl– deposition rate is in the S1 category for all spalling of the rust or an increase in friability of the
regions, with a relatively higher deposition rate at RG resulting fragments with time will mean that corrosion
(57.25 mg m–2 day–1). The latter, being at a higher will accelerate. However, the fact that the first‐year
elevation than the other sites in PL, is situated at the corrosion rate does not decrease for MC, MA and RG is a
foot of the Moka mountain range, which is likely to cause rare phenomenon, which is not observed in other studies.
increased salt deposition due to favourable airflow As per ISO 9223,[27] all locations in PL are of S1P0
patterns on the ridge slope.[58] Based on results from a environmental classification (Table 2) and C2 corrosivity
corrosion study involving field exposures of carbon steel category as determined from corrosion rate results
in Chile (nontropical) and Venezuela (tropical), there has (Table 4). Field exposures carried out in other countries
been a rapid increase in corrosion rate beyond a Cl– with a similar atmosphere (S1P0), but not necessarily the
deposition rate of 50 mg m–2 day–1; however, greater than same TOW, are listed in Table 5. Corrosivity categories
600 mm rain washes the Cl– content of the surface, within ‘C2–C4’ have been obtained for each of the 17 test
reducing the damage.[42] As shown in Figure 4, this stations of S1P0 category and high TOW (at least τ4) in the
rainfall threshold has not been attained at PL and MC. Ibero‐American area.[51] A location in Barranquilla
From the EDS spectrum in Figure 13, MA was the site (Columbia) has even been associated with the highest
with a clear presence of Cl– and sulphate in the corrosion corrosivity category (C5) despite having such a low level
products (on the outer surface), although it has the of pollution.[13] In Venezuela, C3 or C4 has been
lowest corrosion rate and Cl– deposition level (14.49 mg associated with the S1P0 regions having a quite high
m–2 day–1). The presence of building infrastructure in the TOW (τ4).[42] However, low corrosivity (C2) has been
immediate environment and a narrow bay led to a observed in Campeche (Mexico), although the humidity
reduced salinity level compared to the other sites.[59] is high.[60] Therefore, S1P0 regions are represented by a
Furthermore, the specimens are shielded from the effect wide range of corrosion rates, which cannot be explained
of rain, which therefore does not contribute to the by differences in TOW. A previous study carried out in
washing‐off of chlorides or sulphates deposited on the an S1P1 location in PL, lead to a C4 corrosivity
surface of the metal. However, these particles do not classification.[50] Despite being situated on the seashore,
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1487

T A B L E 5 Corrosivity of atmospheres in S1P0 category regions phases, which normally require high chloride and
in tropical countries acidic contents. However, small amounts of these
ISO 9223 corrosivity compounds are present and are not formed at the
Location TOW category metal/rust interface under limited oxygen conditions.
Campeche (Mexico) τ4 C2 2. Marked corrosion rate differences were observed at
the three PL test sites, implying that location‐specific
Ibero American τ4, τ5 C2–C4
conditions, such as proximity to heavy‐oil fired power
Venezuela τ4 C3, C4 stations and natural or manmade forms, were having
Barranquilla τ5 C5 an influence on corrosion kinetics.
(Columbia) 3. Despite its marine atmosphere, corrosion rates in PL
Port Louis (Mauritius) τ3 C4 are also dependent on wind direction. All three
locations had chloride deposition rates in the S1
Abbreviation: TOW, time‐of‐wetness.
category, attributable to the predominant wind direc-
tion from land to sea.
Cl– deposition rate recorded at this site was much lower 4. Chloride deposition rate at the PL sites was highly
than the other sites investigated in the current study. seasonal (roughly 19 and 37 mg m2 day–1, for the
However, the SO2 level was relatively high, which could summer and winter, respectively) and as such linked
be the cause for the more corrosive atmosphere. to wind speed. Thus, demonstrating airborne salinity
Research on marine atmospheric corrosion in tropical was highly dependent on prevailing weather conditions.
regions have attracted interest from several researchers 5. At MC, the airborne salinity (51.4 mg m2 d−1) was
through the years. Atmospheric corrosion tests were markedly higher than two of the PL sites, implying
performed in the Caribbean area comprising of Cuba, that significant chloride levels exist over the in‐land
Mexico and Venezuela, all with a maritime climate.[8] regions of Mauritius.
Pollutants released from factories and other sources did 6. The importance of RH in accelerating atmospheric
not have any noticeable effect on corrosion rate. Even corrosion has been highlighted, especially for small
with a higher Cl– deposition level, sites closer to the sea tropical islands such as Mauritius. The industrial marine
may be less aggressive than sites further inland and sites in PL have demonstrated corrosion rates in the C2
highly humid. The results of field exposures in other category as opposed to the more humid central region
regions in Mauritius showed that the corrosivity of the associated with a higher corrosivity category.
Mauritian atmosphere, in general, is at category C3,
which was mainly attributed to the high RH and, hence, AC KNOW LEDGEM ENTS
the high TOW prevailing around the island.[10] It should The authors express their gratitude to the Central
be noted that airborne salinity and SO2 levels were low in Electricity Board, Mauritius Civil Service Mutual Aid
the regions investigated. In general, pollution level Association Ltd. and Rajiv Gandhi Science Centre for
remains low in Mauritius, apart from PL, especially allocating the required space in their premises for
closer to the power plants. In the current study, there is the field tests. The authors would also like to thank
the formation of a rust layer in a marine atmosphere in the Mauritius Meteorological Services for providing the
which the airborne salinity is low. Coastal regions are climatic data. This study received funding under the
known to cause higher corrosion rates;[14,23,45] however, Africa Collaboration Grant to conduct X‐ray diffraction
this is not observed here. analysis, scanning electron microscopy and energy‐
dispersive X‐ray spectroscopy using equipment available
at Stellenbosch University. Correction added on 4 April
5 | CONCLUSIONS 2022, after first online publication: Online open access
has been added. Open access funded by JISC ‐
S235 carbon steel specimens were exposed at three UNIVERSITY OF SOUTHAMPTON.
specific locations (MA, RG and FG) in the marine
industrial environment of PL and one mid‐island rural C O NF L I C TS O F IN T E R E S T
atmosphere (MC) for a period of nearly 1 year. Key The authors declare no conflicts of interest.
insights into the Mauritius marine atmospheric corrosion
of carbon steel are as follows: A U T H O R C O N TR I B U T I O N S
Yashwantraj Seechurn was involved in conceptualisa-
1. The atmospheric conditions at the four sites limited tion, methodology, investigation, funding acquisition,
the formation of significant akaganeite and magnetite formal analysis, visualisation, writing—original draft,
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
1488 | SEECHURN ET AL.

review and editing. Baboo Y. R. Surnam and Julian A. [19] BS EN ISO 8407, Corrosion of metals and alloys—Removal of
Wharton were involved in supervision and writing— corrosion products from corrosion test specimens, BSI
review and editing. Standards Limited 2014.
[20] BS EN ISO 9226, Corrosion of Metals and Alloys—Corrosivity
of atmospheres ‐ Determination of corrosion rate of standard
DATA A VAILABILITY S TATEMENT
specimens for the evaluation of corrosivity, BSI Standards
The data that support the findings of this study are Limited 2012.
available from the corresponding author upon reasonable [21] S. Sabir, A. A. Ibrahim, Corros. Eng. Sci. Technol. 2017, 52,
request. 276.
[22] BS EN ISO 9224, Corrosion of metals and alloys—Corrosivity
ORCID of atmospheres—Guidance values for the corrosvity catego-
Yashwantraj Seechurn http://orcid.org/0000-0002- ries, BSI Standards Limited, 2012.
1733-0642 [23] M. Benarie, F. L. Lipfert, Atmos. Environ. 1986, 20,
1947.
Baboo Y. R. Surnam http://orcid.org/0000-0002-
[24] Q. Yu, C. Fang Dong, Y. Hua Fang, K. Xiao, C. Yun Guo,
3941-8135
G. He, X. Gang Li, J. Iron Steel Res. Int. 2016, 23, 1061.
Julian A. Wharton http://orcid.org/0000-0002- [25] D. de la Fuente, I. Díaz, J. Simancas, B. Chico, M. Morcillo,
3439-017X Corros. Sci. 2011, 53, 604.
[26] Y. Ma, Y. Li, F. Wang, Corros. Sci. 2010, 52, 1796.
REFERENCES [27] BS EN ISO 9223, Corrosion of metals and alloys—Corrosivity
[1] M. Morcillo, B. Chico, J. Alcántara, I. Díaz, J. Simancas, of atmospheres—Classification, determination and estima-
D. De La Fuente, Mater. Corros. 2015, 66, 882. tion, BSI Standards Limited 2012.
[2] C.L. Wallin, J. Tidblad, T. E. Graedel, Atmospheric Corrosion, [28] Mauritius Meteorological Services, Meteorological Observa-
John Wiley & Sons, Inc., Hoboken, NJ 2016. tions and Climatological Summaries 2019. http://metservice.
[3] B. Jegdic, S. Polic‐Radovanovic, S. S. Ristic, A. B. Alil, Sci. intnet.mu/publication.php (accessed: June 2021).
Tech. 2011, 61, 50. [29] J. C. Guerra, A. Castañeda, F. Corvo, J. J. Howland,
[4] Mauritius Meteorological Services, Climate of Mauritius, J. Rodríguez, Mater. Corros. 2019, 70, 444.
http://metservice.intnet.mu/climate-services/climate-of- [30] W. Wu, X. Cheng, J. Zhao, X. Li, Corros. Sci. 2020, 165,
mauritius.php (accessed: June 2021). 108416.
[5] CEB, power stations, https://ceb.mu/our-activities/power- [31] J. Alcántara, de la D. Fuente, B. Chico, J. Simancas, I. Díaz,
stations accessed: June 2021. M. Morcillo, Mater. (Basel) 2017, 10, 406.
[6] P. Tatayah, Air quality—Fort George, Mauritius, 2017. [32] H. Tamura, Corros. Sci. 2008, 50, 1872.
[7] J. F. Ríos Rojas, D. Escobar Ocampo, E. A. Hernández García, [33] A. Momber, Mater. Corros. 2012, 63, 333.
C. E. Arroyave Posada, DYNA. 2015, 82, 128. [34] A. C. Scheinost, U. Schwertmann, Soil Sci. Soc. Am. J. 1999,
[8] F. Corvo, C. Haces, N. Betancourt, L. Maldonado, L. Véleva, 63, 1463.
M. Echeverria, O. T. De Rincón, A. Rincón, Corros. Sci. 1997, [35] J. F. W. Bowles, Encycl. Geol 2021, 442.
39, 823. [36] R. Balasubramaniam, A. V. R. Kumar, P. Dillmann, Sci.
[9] C. G. Staub, F. R. Stevens, P. R. Waylen, Appl. Geogr. 2014, Archaeol. Archaeomaterials 2005, 85, 275.
54, 222. [37] J. Alcántara, B. Chico, J. Simancas, I. Díaz, D. de la Fuente,
[10] B. Y. R. Surnam, C. V. Oleti, Corros. Eng. Sci. Technol. 2012, M. Morcillo, Mater. Charact. 2016, 118, 65.
47, 446. [38] M. Morcillo, B. Chico, J. Alcántara, I. Díaz, R. Wolthuis,
[11] Y. Liu, Z. Wang, Y. Wei, Int. J. Electrochem. Sci. 2019, 14, 1147. D. de la Fuente, J. Electrochem. Soc. 2016, 163, C426.
[12] S. W. Dean, G. Hernandez‐duque, J. B. Bushman, Marine [39] K. Kreislova, D. Knotkova, Mater. (Basel) 2017, 10, 394.
Corrosion in Tropical Environments, ASTM, West Consho- [40] T. E. Graedel, J. Electrochem. Soc. 1990, 137, 2385.
hocken, PA 2000. [41] U. Schwertmann, R. M. Cornell, Iron oxides in the laboratory,
[13] J. G. Castaño, C. A. Botero, A. H. Restrepo, E. A. Agudelo, preparation and characterization, WILEY‐VCH Verlag GmbH,
E. Correa, F. Echeverría, Corros. Sci. 2010, 52, 216. Federal Republic of Germany, Weinheim 1992.
[14] J. Alcántara, B. Chico, I. Díaz, D. de la Fuente, M. Morcillo, [42] R. Vera, O. T. de Rincón, M. Bagnara, N. Romero, R. Araya,
Corros. Sci. 2015, 97, 74. S. Ossandón, Mater. Corros. 2018, 69, 614.
[15] A. Gismelseed, S. H. Al‐Harthi, M. Elzain, A. D. Al‐Rawas, [43] I. S. Cole, D. A. Paterson, W. D. Ganther, Corros. Eng. Sci.
A. Yousif, S. Al‐Saadi, I. Al‐Omari, H. Widatallah, Technol. 2003, 38, 129.
K. Bouziane, Hyperfine Interact. 2006, 167, 753. [44] G. R. Meira, C. Andrade, C. Alonso, I. J. Padaratz, J. C. Borba,
[16] A. Taalba, H. Xie, M. G. Scarratt, S. Bélanger, M. Levasseur, Corros. Sci. 2008, 50, 2724.
Biogeosciences 2013, 10, 6793. [45] F. Corvo, T. Perez, L. R. Dzib, Y. Martin, A. Castañeda,
[17] BS EN ISO 9225, Corrosion of metals and alloys—Corrosivity E. Gonzalez, J. Perez, Corros. Sci. 2008, 50, 220.
of atmospheres—Measurement of environmental parame- [46] S. Pintos, N. V. Queipo, O. Troconis De Rincón, A. Rincón,
ters, BSI Standards Limited 2012. M. Morcillo, Corros. Sci. 2000, 42, 35.
[18] ISO 8565, Metals and alloys—Atmospheric corrosion testing— [47] S. Hœrlé, F. Mazaudier, P. Dillmann, G. Santarini, Corros.
General requirements 2011. Sci. 2004, 46, 1431.
 15214176, 2022, 9, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/maco.202112871 by Readcube (Labtiva Inc.), Wiley Online Library on [16/02/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SEECHURN ET AL. | 1489

[48] H. Antony, L. Legrand, L. Maréchal, S. Perrin, P. Dillmann, [58] I. S. Cole, W. D. Ganther, D. A. Paterson, G. A. King,
A. Chaussé, Electrochim. Acta 2005, 51, 745. S. A. Furman, D. Lau, Corros. Eng. Sci. Technol. 2003, 38, 259.
[49] C. G. Soares, Y. Garbatov, A. Zayed, G. Wang, Corros. Sci. [59] I. S. Cole, D. A. Paterson, W. D. Ganther, A. Neufeld,
2014, 51. B. Hinton, G. McAdam, M. McGeachie, R. Jeffery,
[50] B. Y. R. Surnam, Anti‐Corrosion Methods Mater 2015, 62, 246. L. Chotimongkol, C. Bhamornsut, N. V. Hue,
[51] E. Almeida, M. Morcillo, B. Rosales, Mater. Corros. 2000, 51, S. Purwadaria, Corros. Eng. Sci. Technol. 2003, 38, 267.
865. [60] D. C. Cook, A. C. Van Orden, J. J. Carpio, S. J. Oh, Hyperfine
[52] D. Landolt, Corrosion and surface chemistry of metals, EPFL Interact. 1998, 113, 319.
Press, Lausanne, Switzerland 2007.
[53] I. S. Cole, D. Lau, D. A. Paterson, Corros. Eng. Sci. Technol.
2004, 39, 209. SU PPOR T ING INF OR MATI ON
[54] R. M. Katona, J. C. Carpenter, A. W. Knight, C. R. Bryan, Additional supporting information may be found in the
R. F. Schaller, R. G. Kelly, E. J. Schindelholz, Corros. Sci. online version of the article at the publisher’s website.
2020, 177, 108935.
[55] E. Schindelholz, R. G. Kelly, Corros. Rev. 2012, 30, 135.
[56] I. Matsushima, in Uhlig's Corrosion Handbook (Ed: R. W.
Revie), John Wiley & Sons, Inc., Hoboken, NJ 2011. How to cite this article: Y. Seechurn, B. Y. R.
[57] D. De La Fuente, I. Díaz, J. Alcántara, B. Chico, J. Simancas, Surnam, J. A. Wharton, Mater. Corros.
I. Llorente, A. García‐Delgado, J. A. Jiménez, P. Adeva, 2022;73:1474–1489.
M. Morcillo, Mater. Corros. 2016, 67, 227. https://doi.org/10.1002/maco.202112871