Wear 438-439 (2019) 203053

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

The mechanism of erosion-corrosion of API X65 steel under turbulent slurry T

flow: Effect of nominal flow velocity and oxygen content
Javiera Aguirrea, Magdalena Walczaka,∗, Michael Rohwerderb
a
Department of Mechanical and Metallurgical Engineering, Escuela de Ingeniería, Pontificia Universidad Católica de Chile, Santiago, Chile
b
Max-Planck-Institut für Eisenforschung GmbH, Düsseldorf, Germany

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

Keywords: Erosion-corrosion (E-C) of carbon steel API 5L X65 by mineral quartz and alumina slurry is studied to understand
Erosion-corrosion the mechanism of corrosion-accelerated material's degradation. A rotating cylinder electrode is employed to
API 5L X65 steel simulate the E-C damage and for electrochemical characterization, observed at different flow velocities and
Erosion-enhanced corrosion contents of dissolved oxygen (DO). Microstructural inspection using SEM and EBSD reveals the extension of the
EBSD
work-hardened layer and other features of the sub-surface. The effect of nominal flow velocity is shown to be
associated with an increment in average particles' impact angle and frequency of impacts at the faster flow. The
increase in DO availability is shown to promote localized-type corrosion different from pitting, demonstrating
the mechanism of erosion-enhanced corrosion.

1. Introduction mechanism for practical purposes, but the simultaneous progress of the
wear mechanisms and possible interdependence of variables may con-
The hydraulic transport of solids (hydro-transport) in the form of siderably change the way of material removal and its contribution to
slurry is an essential technology for transporting large volumes of raw, the overall E-C rate. Thereby, a comprehensive understanding of the
process and waste materials in several industries such as oil and gas, erosion-corrosion mechanisms is important to establish a practical mi-
mining and metallurgy. From a practical point of view, carbon and low- tigation strategy.
alloyed steels are the material of choice when building a slurry pipeline In this context, numerous studies have been focused on the influ-
due to their good ductility-strength relation, well-known construction ence of individual variables on E-C wear such as the properties of the
procedures and availability of bulk material, especially where long erodent [14–18], particle impact angle [19–22], concentration and
transport distance is required [1]. However, although this type of velocity of particles [23–29], properties of the target material [30–33]
transport system provides an economical and efficient solution com- and slurry chemistry [31,34]. However, in this multivariable system,
pared to other alternatives, corrosion and erosion have constituted an not all the factors contribute with the same intensity to the material
essential threat to the integrity of the slurry transport system, posing degradation. Therefore, the extent of E-C rate in a determined system of
economic concerns associated with technical safety risks [2–5]. metal-electrolyte/slurry can be attributed to some key variables, which
Erosion-corrosion (E-C) of metallic materials describe the material have been suggested to be the particle velocity, the impact angle, and
degradation process where mechanical wear by solid particles, liquid, chemistry of the slurry [13,35–37]. For the specific system of API 5L
or the combination of both interact with the electrochemical corrosion X65 steel exposed to a slurry of quartz and silica mixture, flow velocity
process [6]. Extensive experimental works [5,7–13] had demonstrated and content of dissolved oxygen (DO) were found to be the most re-
that erosion and corrosion in flowing slurry would enhance each other levant by the statistical approach [38]. This work focuses on clarifying
resulting in a mass loss of material greater than expected for the sum- the mechanisms by which these two system parameters determine the
mation of mass losses due to pure erosion and pure corrosion. This rate of E-C damage.
additional mass loss is commonly referred to as positive synergism and The effect of slurry particle velocity has been extensively studied in
presents a major issue in understanding and predicting the rate of de- the last decades because of its dominant effect on the material removal
gradation due to a large number of the involved variables. Each vari- rate [22,28,30,37,39,40]. The erosion rate has been found generally
able is associated with a specific wear process, conceptualizes as a related to the particles’ flow velocity by a power-law relationship in

∗
Corresponding author.
E-mail address: [email protected] (M. Walczak).

https://doi.org/10.1016/j.wear.2019.203053
Received 30 April 2019; Received in revised form 22 July 2019; Accepted 17 September 2019
Available online 18 September 2019
0043-1648/ © 2019 Elsevier B.V. All rights reserved.
J. Aguirre, et al. Wear 438-439 (2019) 203053

which the power exponent for velocity varies in the range of 0.34–4.83
depending on the relative properties of the particles and material, as
well as specific conditions of the test [32,41–43]. Because flow velocity
correlates with the kinetic energy conveyed by each particle and with
the collision efficiency of particles with the target [40,44], the exten-
sion of produced damage can be associated with the intensity of these
effects. However, the contribution of slurry velocity can also sig-
nificantly accelerate corrosion reactions mass-transfer, mainly by in-
creasing local turbulence when particles interact with the boundary
layer, and by the increase in surface roughness [5,45–47].
Further, the effect of dissolved oxygen present in slurries was stu-
died by several authors [48–51] revealing the possibility of forming
protective scales, although most of these works were focused on oil and
gas industry and slurries containing also carbon dioxide [52]. Ac-
cording to Postlethwaite et al. [48], E-C in slurry pipelines carrying
aerated slurries of sand is under oxygen mass-transfer control, with
corrosion being the dominant mode of metal loss. However, the role of
oxygen in E-C by slurries is complex and remains unclear since only a
few works are published on this specific factor in biphasic flow systems
[48–50].
Our work aims at clarifying the mechanisms of wear on API 5L X65
steel subjected to erosion–corrosion by mining slurry, by providing a
fundamental understanding of the effect of the slurry velocity and Fig. 1. Representative images of X-65 steel microstructure taken at two in-
dependent locations: Gray areas – ferrite, red – perlite. (For interpretation of the
dissolved oxygen on the microstructure of the sub-surface. The ex-
references to color in this figure legend, the reader is referred to the Web
perimental work is implemented using a rotating cylinder electrode
version of this article.)
(RCE) to simulate the damage by E-C and mineral slurry of quartz and
alumina mixture in an acidic electrolyte. The analysis is carried out by
means of weight loss measurements and electrochemical analysis, cleaned by sonication in distilled water, rinsed in acetone, dried by
which is complemented by systematic image analysis. The surface in- blowing hot air, and stored in a desiccator until used for the experi-
spection is conducted on the top surface and sub-surface using scanning ment. Before the exposure, the weight of each sample was measured
electron microscopy (SEM) and electron backscatter diffraction tech- three times using an analytical balance (Sartorius BP221S, sensitivity of
nique (EBSD), respectively. The technique of EBSD featuring analysis of 0.1 mg).
local misorientations has been shown effective for studying the fine The erodent used for preparing the slurry was a mix of quartz and
microstructure of materials [53]; however, this is the first demonstra- alumina in the proportion of 4:1 in mass, respectively. This mixture is
tion of its application in the context of the erosion-corrosion process. representative of typical copper ore in Chile in terms of relative
abundance of these components. The shape of the particles as seen
under scanning electron microscope (SEM, Zeiss (Leo) 1550vp) and size
2. Experimental distribution are shown in Fig. 2. The circularity of the particles (ratio of
the area of the shape to the area of the circle) was calculated from the
2.1. Materials SEM by digital image analysis process to be about 0.33 ± 0.04. The
hardness of quartz and alumina particles was measured using na-
The material used in this study is a low alloyed carbon steel typi- noindentation technique (Bruker, Hysitron TI 950) to be 13.25 ± 1.15
cally used for constructing pipelines, API 5L X65, hereafter referred to and 31.13 ± 1.40 GPa, respectively.
as X65 [1]. The chemical composition of the material determined by
glow discharge emission spectrometry bulk analysis (Spectruma, GDA
2.2. Experimental set-up
750 HR) is presented in Table 1.
Fig. 1 presents the microstructure of the steel, consisting of 96.2% of
To simulate the process of E–C damage, a rotating cylinder electrode
ferrite and 3.8% of perlite determined by image analysis [54]. The steel
(RCE) was used. Fig. 3 shows a schematic representation of the set-up, a
hardness determined using a micro-hardness tester (LECO, M-400-H)
full description of which can be found in our previous study [38]. The
was found to be about 2.32 GPa. The test specimens were hollow cy-
cylindrical shaft allows exposing up to two samples located in its lower
linders of 15–10 mm external-internal diameter and 10 mm in height.
part, separated by two rubber washers to provide electrical insulation.
The material was extracted from a pipe, cut to shape by lathe ma-
The concentricity of the shaft is ensured by a cylindrical pivot bearing
chining in such a way that the axis of the pipe coincides with that of
at the bottom of the slurry container. The nominal flow velocity, de-
each sample cylinder.
termined as tangential flow at the cylinder's surface, is controlled by a
The external surfaces of the cylindrical samples, i.e., the surface to
frequency modulator set to an rpm number (from 0 to 24,000 rpm)
be exposed, were ground using SiC paper (from 320 to 1200 grit) to
necessary to produce the desired tangential velocity at the external
obtain a final surface roughness (Ra) of 0.073 ± 0.006 μm (PCE in-
surface of the sample. The four baffles improve homogeneity of the
struments, PCE-RT 1200). This procedure diminished cylinders dia-
flowing slurry by reducing the tendency of particles to settle.
meter by approx. 50 μm. After the final polishing, the samples were
The exposure set-up allows automatic control of temperature and
the content of dissolved oxygen (DO) with the precision of ± 1 °C
Table 1
Chemical composition in weight percentage (wt%) of the API 5L X65 steel used
and ± 1 ppm, respectively. The DO control system consists of two gas
in this study. hoses submerged in the slurry, one for nitrogen and one for oxygen, and
a DO probe (Atlas Scientific probe, DO Kit accuracy ± 0.05 ppm) for in-
Material C Si Mn P V S Fe
situ measurements. The oxygen level was managed through a micro-
API 5L X65 0.06 0.21 1.53 0.01 0.03 0.001 Bal. controller integrating the gas supply valves and the oxygen probe.
When the DO level exceeded the set value, nitrogen (purity 99,999 vol

2
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 2. Characteristics of the quartz/alumina erodent: a) SEM image; b) Granulometric particle size distribution.

4.7 cm2, a graphite electrode as counter electrode (CE) with effective

area of exposure 24.8 cm2, and silver|silver chloride electrode
(Ag|AgCl) as the reference electrode (RE). The CE and RE were located
parallel to the shaft at the distance of 2 cm from it. To avoid damage,
each electrode was protected with a glass tube allowing access of the
electrolyte from below. The conducting of the electrochemical signal
from the shaft was ensured by a carbon brush with conductive grease.

2.3. Experimental procedure

The experimental runs shown in Table 2 were designed using

parameters in their ranges observed in copper mining operations to be
aggressive. The particular values were selected based on their sig-
nificance in the E–C process identified in our previous study [38]. The
runs were implemented separately for weight loss, surface character-
ization and electrochemical polarization, each one using two samples,
except for electrochemical polarization in which only one sample was
for each polarization.
Each exposure was run for 75 min under open circuit potential
condition (no potential applied) to generate measurable weight loss
without significant distortion of the geometry of the test sample or
attrition of the erodent. The slurry was prepared by suspending the
erodent in the desired concentration, i.e. 65 wt%, in 0.1 M sodium
sulfate (Na2SO4). With the objective of study only the effect of nominal
flow velocity, hereafter referred to as slurry velocity or particle velo-
city, and DO in the wear rate of the steel, some parameters were kept
constants in all the experiments. The pH was kept at 5, controlled with a
pH-meter, and adjusted manually by adding drops of 0.3 M H2SO4 or
0.3 M NaOH when pH levels deviated by more than ± 0.5 units and the
temperature was kept at 45 ± 1 °C during the entire experiment. Each
E-C exposure was conducted using a freshly prepared particle mixture.

2.4. Weight loss measurements

Fig. 3. Schematic representation of the experimental set-up in the arrangement
for electrochemical measurements: a) Side view, b) Top view. For weight loss
For determining weight loss after E-C exposure, a three-stage
measurement caused by E–C two samples are exposed (no electrochemical
cleaning was carried out immediately after removing from the shaft.
connection), whereas, for electrochemical measurements, only one sample is
exposed.
Table 2
Summary of experimental runs.
%) was added to the slurry; analogously, when the DO was below the
Nominal flow velocity (m·s−1) Angular velocity (rpm) Set DO level (ppm)
set level, oxygen (purity 99,999 vol%) was added. It should be noted
that the particular condition of nominal DO of 0 ppm is not provided, as 3 3820 0a
in practice it corresponds to values between 0 and 0.3 ppm rather than 10b
the absolute absence of DO. 7 8912 0a
10b
Further, in-situ electrochemical measurements were implemented
using an arrangement of three-electrodes, with only one steel cylinder a
Effective value between 0 and 0.3 ppm.
exposed as working electrode (WE) with effective area of exposure b
Variation ± 1 ppm.

3
J. Aguirre, et al. Wear 438-439 (2019) 203053

Mechanical cleaning using water and a soft-bristle brush was followed Table 3
by an ultrasonic bath in acetone for 3 min to eliminate traces of slurry EBSD acquisition parameters.
and oxide scale adhered to the surface. Then, to remove oxide layer Accelerating voltage Working distance Step size Binning Scan size
strongly adhered to the surface, the samples were ultrasonically cleaned
for 5 min in an inhibiting acid solution of Clark (1000 mL of hydro- 15 keV 22 mm 50 nm 5×5 12 μm × 28 μm
chloric acid, 20 g of antimony trioxide (Sb2O3), and 50 g of stannous
chloride (SnCl2)) according to the ASTM standard G1-03 [55]. Finally,
the samples were rinsed with water, then sonicated in ethanol for 3 min obtained using the same SEM instrument described above, which was
and dried by blowing hot air. also equipped with a Hikari EBSD Camera provided by EDAX. The
The weight of each cleaned sample was measured three times using measurements were performed with the sample tilted at 70° towards
an analytical balance with a sensitivity of 0.1 mg (Sartorius BP221S). the EBSD detector, using the acquisition parameters summarized in
The effective E–C rate in mg⋅cm−2 h−1 was calculated from the weight Table 3. For statistical reproducibility, at least four mappings were done
loss average of the two samples exposed in the same experimental run. for each sample in different locations. The acquisition software was TSL
OIM, version 6.2, where only the ferritic phase (BCC) was indexed ac-
cording to the Miller indexes (hkl) for this type of steel.
2.5. Potentiodynamic polarization
The data were processed using the software OIM Analysis™ version
7.0 in order to analyze local misorientations by the Kernel Average
Potentiodynamic polarization for the desired exposure conditions
Misorientation (KAM) method. KAM is a calculation of the average
was conducted on independent samples, not used for weight loss
misalignment between the orientation corresponding to each pixel in
measurement. Prior polarization, a freshly prepared sample was ex-
the map and its nearest neighbors, which allows the analysis of the
posed to the flowing slurry for 1 h during which the value of open
degree of local deformation of the material. In this study, only first
circuit potential (OCP) was monitored using a potentiostat/galvanostat
neighboring pixels with a misorientation angle lower than a threshold
(Gamry Reference 600, Gamry Framework software). The anodic and
of 5° are considered in the calculation in order not to include mis-
cathodic curves were collected individually under slurry flow, polar-
orientations due to grain boundaries. KAM is directly related to the
izing the sample from OCP at the scan rate of 0.001 V⋅s−1 in either
density of geometrically necessary dislocations (GNDs) responsible for
anodic or cathodic direction. In order to verify the possibility of elec-
the lattice curvature [58]. An increase in GNDs due to non-uniform
trochemical pitting associated with the presence of oxygen, the anodic
plastic deformation contributes to the hardening of the material as
sweep was extended to +0.8 V vs. Ag|AgCl. The value of corrosion
explained by the strain gradient theory [59,60].
current density (jcorr) was determined via the Tafel extrapolation
technique using potential ranges corresponding to activation control
[56], and the corrosion rate in mm·yr−1 was quantified by the Faraday's 3. Results
laws according to ASTM G 102-89 [57] considering for the calculation
the following reaction: 3.1. E-C rate from weight loss measurements
Fe ↔ Fe2 + + 2e‾ . (1)
Fig. 4 shows the E-C rate determined from the weight loss data for
the different conditions of exposure. Depending on the oxygen content
2.6. Surface and sub-surface analysis in the slurry, the wear rate spans between 0.5 and 11.7 mg⋅cm−2h−1
for the two levels of velocity. For the runs of the same nominal flow
For analysis of surface and sub-surface, the samples were extracted velocity but different oxygen content, a significant difference in the E-C
from the set-up immediately after E-C exposure, sonicated for 3 min to rate is observed with an increase by a factor of at least two.
remove traces of particles, then dried with hot air and stored in a de-
siccator at vacuum conditions for further analysis.

2.6.1. Surface characterization

The surface morphology of the samples was analyzed by SEM (Zeiss,
Leo 1550 VP) inspecting surface area in the axial direction and selecting
a representative image for further analysis. SEM images were com-
plemented by elemental analysis using energy dispersive X-Ray spec-
troscopy (EDX, Oxford Instrument) coupled with the SEM.

2.6.2. Sub-surface characterization

For the study of sub-surface, the samples were prepared using
standard metallographic techniques, where the samples cross-sections
were embedded in conductive epoxy resin, then cured and the surface
ground using SiC papers with grit sizes 240 (52 μm), 320 (35 μm), 400
(22 μm), 600 (15 μm), 800 (13 μm) and 1500 (8 μm). The first grinding
step aimed to produce a flat surface; the duration of the subsequent
steps was 5 min each. The final SiC grinding step was followed by two
polishing steps using the polishing cloth and diamond suspensions of
3 μm and 1 μm for 5 min each. The final polishing step was conducted
using 0.04 μm of colloidal silica suspension for 30 min. The samples
were thoroughly rinsed in water between the different grinding/pol-
ishing steps and finally sonicated and cleaned in ethanol.
The sub-surface was analyzed at different magnifications first by
SEM and then by EDX. Finally, the microstructural analysis of the cross-
section of each sample was conducted using electron backscatter dif- Fig. 4. E-C rate determined from weight loss measurements for two flow ve-
fraction technique (EBSD). The EBSD maps of the sub-surface were locities and two levels of nominal DO.

4
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 5. Potentiodynamic polarization for variation of: (a) Slurry flow velocity, and (b) Content of DO in the slurry.

3.2. Potentiodynamic polarization an evident increase in the damage by plastic deformation, showing a
higher density of craters and lips on the surface. These craters become
The potentiodynamic polarization curves collected for analyzing the deeper and more random at 7 m⋅s−1, with an increased presence of
electrochemical behavior of steel by varying the slurry velocity and DO indentation–type craters on the surface.
are shown in Fig. 5. The values of OCP obtained during monitoring The cross-sections of the samples reveal the surface profiles, with
prior to polarization are shown in Fig. S1 (Supporting material). the depth of craters quantified to be on average 0.6 and 1.7 μm for the
The electrochemical parameters derived from the polarization velocities of 3 and 7 m⋅s−1, respectively. SEM images of the sub-surface
curves are presented in Table 4. It is observed that in static conditions damaged at 3 m⋅s−1 present a profile with craters more rounded and
(reference sample) there is a well-defined limiting current in the homogeneous in-depth, compared with samples exposed at 7 m⋅s−1 that
cathodic branch followed by an increase in current at overpotentials reveal irregular and sharp craters with areas larger than depths. At the
more negative than −0.9 V. Whereas, the anodic branch shows the edges of the indents, some extruded material and cracks towards the
active behavior of the material in the entire measuring range. inner of the steel are observed.
The flow condition of the electrolyte in the absence of oxygen A distintive type of surface morphology is formed after exposure to
(Fig. 5a) affects both the anodic and cathodic polarization currents, the slurry with a high content of DO (Fig. 7). The samples exposed to
observing a greater increase in the cathodic one at a higher velocity, the slurry with 10 ppm of DO have a surface with irregular grooves of
producing a higher corrosion current density (jcorr). On the other hand, large extension and depth, within which retained particles of erodent
the increase of DO in the slurry (Fig. 5b) shows a shift in the corrosion can be observed. The magnitude of the damage is not clear since some
potential (Ecorr) of almost 0.2 V towards anodic direction, associated particles, apparently of erodent, and deposits are found covering the
with the increase in the current density of the cathodic branch, almost surface of the steel. Whereas, exposure at nominal 0 ppm of DO pro-
doubling the jcorr of the system (Table 4). The anodic current density duces heavy plastic deformation with a surface covered by craters and
presents a slight increase at higher DO content, without significant lips analogous to those observed in Fig. 6 for 7 m⋅s−1.
change in the shape of the anodic branch. This contrasting difference can also be observed in the cross-section
of the samples presented in the same Fig. 7. After exposure to the slurry
3.3. SEM analysis of nominal 0 ppm DO, a surface profile is observed with irregular and
sharp craters characteristic to plowing and extrusion of the steel target,
Fig. 6 shows representative SEM micrographs of the samples ex- with an average crater depth of 1.6 μm. However, in samples exposed at
posed at low (3 m⋅s−1) and high (7 m⋅s−1) velocity at nominal DO of 10 ppm of DO shows some deep and wide pits through the sub-surface,
0 ppm. Visible damage by plastic deformation is observed on the top with an average depth of 5.4 μm, where it is possible to observe these
surface due to the impact of slurry particles on the steel. At both ve- pits filled with deposits and particles.
locities cutting–type-, plowing–type-, and indentation–type craters are Fig. 8 shows the EDX mapping on the top surface and cross-section
observed on the surface, along with the formation of lips and sparse of these last samples to verify a possible change in the surface chem-
presence of steel debris. As expected, the increment in velocity reveals istry. There is an evident presence of significant amounts of oxygen,
silicon, and alumina on the surface, where it is possible to differentiate
Table 4 two well-marked areas: one with a high concentration of oxygen and
Electrochemical parameters from potentiodynamic polarization. the other high in iron content from the base metal. Areas with a high
oxygen content are preferably located in the groove zones, along with
Exposure Ecorr (V vs jcorr (A·cm−2) Corrosion rate
Ag|AgCl) (mm·yr−1) high content of silica and aluminum.

Reference (0 m⋅s−1 – −0.642 6.49 × 10−6 0.08

0 ppm) 3.4. EBSD analysis
3 m⋅s−1 – 0 ppm −0.675 8.27 × 10 −5
0.97
7 m⋅s−1 – 0 ppma −0.599 1.76 × 10−4 2.06
Figs. 9 and 10 show EBSD analysis of the sub-surface as affected by
0 ppm – 7 m⋅s−1a −0.603 1.69 × 10−4 1.97
10 ppm – 7 m⋅s−1 −0.466 3.21 × 10−4 3.75 slurry velocity and oxygen content, respectively. The reference sample,
i.e., prior exposure to the slurry, has a similar sub-surface throughout
a
Independent experimental runs. the bulk, except the smooth layer of smaller grain sizes at the surface

5
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 6. Representative SEM micrographs of the surface (left-hand column) and cross-sections (right-hand column) of samples exposed at 3 and 7 m⋅s−1 of the slurry at
0 ppm of DO. Magnification 10,000× .

that is typical for surface preparation employed in this study. After the velocities of 3 and 7 m⋅s−1, respectively.
exposure to the slurry, the damage induced by impacting particles has a The effect of DO content on the sub-surface damage is shown in
different appearance revealing that the grain size of the sub-surface Fig. 10. The image quality reveals a significant difference in the ap-
decreases as the velocity increases. In the case of samples exposed at pearance of microstructures after exposure to the slurry in the nominal
3 m⋅s−1, there is a top layer of deformed grains of about 1.7 μm in absence (0 ppm) and the presence of DO (10 ppm). In the absence of
thickness, which is followed immediately by a typical grain structure of oxygen, the microstructure is analogous to that described in Fig. 9, with
the bulk. Whereas, exposure to 7 m⋅s−1 reveals a thinner top layer, of layers of distinctive grain size in the sub-surface, extending down to
about 2.1 μm thickness, followed by a layer of about 4.6 μm thickness 1.8 μm below the nominal surface of the smallest nanograins, followed
consisting of micro-grains below which the nominal bulk micro- by the transition layer consisting of grains in the micrometer length
structure is observed. scale extending further 3.7 μm and finally the grains characteristic of
The local misorientation analysis by the KAM method shown in the bulk material. On the other hand, the sub-surface of samples ex-
Fig. 9, allows assessing the qualitative magnitude of plastic deformation posed to 10 ppm of DO does not show an apparent change in the grain
in the sub-surface of the samples. The blue and red areas in the KAM size immediately under the worn surface. There is also no deformation
analysis indicate the extreme values of local misorientation, where 0° observed in sections of the craters. However, incomplete grains are
and 5° correspond to non-deformed and highly deformed locations in observed at the sides and bottom of the crater.
grain, respectively. The effect of flow velocity can be quantified by the The KAM images reveal strong plastic deformation for samples ex-
depth of the deformed layer, which extends to 2.1 μm and 4.5 μm for posed to the oxygen-free slurry, analogous to that observed in Fig. 9.

6
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 7. Representative SEM micrographs of the surface (left-hand column) and cross-section (right-hand column) of samples exposed at 0 ppm and 10 ppm of DO in
the slurry at the velocity of 7 m⋅s−1. Magnification 5000× .

Whereas, exposure to the oxygen-rich slurry results in slight plastic mainly explained by the kinetic energy conveyed by the erodent since
deformation of the top surface and no plastic deformation within the only particles with enough kinetic energy are capable of inducing sur-
dimples. In particular, the grains comprising the pit show no plastic face deformation and producing a removal of the material. Since the
deformation around its perimeter, which is consistent with the in- faster flow of slurry due to the increase in rotational frequency of the
complete grains described previously. electrode implies more kinetic energy of the solid phase, the results of
It should be noted that the relative location of the top surface in E-C rate observed here (Fig. 4) would be qualitatively explained.
Figs. 9 and 10, is not the same as suggested by the presentation of the However, since the process progresses in time, other side effects asso-
image, because of continuous removal of material shifts the actual lo- ciated with the change in particle velocity should be taken into con-
cation of the top surface concerning the rotation axis. sideration, such as turbulence of the flow, the frequency of particles’
impact, change in effective impact angle, evolution of surface rough-
ness in time and contribution of mass loss due to anodic dissolution.
4. Discussion The morphology of top- and sub-surfaces observed after exposure
(Figs. 6–10) correspond to the state of damage accumulated during the
4.1. Effect of slurry velocity on E-C mechanisms entire time of exposure consisting of successive impacts of particles
hitting the steel and degradation by electrochemical corrosion process.
The increase in the wear rate by the increment of slurry velocity The particle velocity vector of each particle, as introduced by Finnie
observed in Fig. 4 is consistent with studies by other authors in similar and Bitter [61–63], can be described as a sum of two components, a
systems [8,26,28,49]. This effect, as discussed by the other authors, is

7
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 8. Representative EDX analysis on the: (a) Top surface and (b) Cross-section of samples exposed to the slurry with 10 ppm of DO. Magnification 5000× . Carbon
element in (b) (pink color) corresponds to the composition of the support resin (background). (For interpretation of the references to color in this figure legend, the
reader is referred to the Web version of this article.)

horizontal one, associated with the plowing mechanism of surface re- Subsequent impacts of particles flatten the ridges, generating metal
moval, and a vertical one, associated with penetration/indentation and debris by fracturing them off [22,64], explaining the presence of steel
accumulation of fatigue damage. Even though in our RCE set-up it is not debris in our results.
possible to control the angle of impact (θ) of each particle, i.e., the In the case of the faster flow (Fig. 6) the impact velocity and fre-
angle between velocity vector of a particle and the target surface, scar quency of particle impacts both increase, while the distribution of
morphologies characteristic of low impact angles (θ < 45°) and high particles' impact angles is modified by the slurry turbulence [65]. In
impact angles (θ > 45°) could be identified at both velocities (Fig. 6). particular, characteristic craters of high impact particles are observed,
Thus, it is possible to distinguish between the dominant erosion me- shifting the mechanism of material's removal towards indentation at
chanisms at both the nominal flow velocities. normal incidence and micro-machining. However, the first mechanisms
The scars produced in exposure to the slower flow appear less deep of metal removal seem to dominate under high particle velocity con-
and more elongated as compared with the faster flow (Fig. 6), in- ditions as indicated by the morphology of the crater. In this case, the
dicating a greater magnitude of the horizontal component of velocity in indentation crater displaces the deformed material, forming an edge
this case. As an abrasive particle slides on the surface after being pro- around the crater due to the impact of the particle. The material can be
jected at a low angle of impact, the preferential material's removal is eliminated by deformation or fracture of these edges by the subsequent
expected to be plowing or cutting. As the particles slide on the surface, impact of particles as described for the normal impacts of angular
they squeeze the metal forward and sideways to form ridges. particles [22,66].

8
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 9. EBSD analysis of the sub-surface microstructure at different velocities of slurry flow at 0 ppm DO. Image quality (top row); KAM analysis (bottom row).

The preferential mechanism of material's removal is thus different at particles. All the three aspects of particle's impact translate into a higher
the two velocities providing a qualitative explanation for the difference absorption of energy towards the material's bulk. As the top layer of the
in the corresponding E-C rate (Fig. 4). In terms of flow parameters, this surface layer undergoes work-hardening, the load is progressively
difference can be understood as the totality of effects associated with transmitted to the bulk grains. This load is initially transformed into
the frequency of impacts, average kinetic energy conveyed per particle work resulting in the formation of micro-grains, and after enough dis-
and impact angle. Considering that in case of ferritic steels, for a given locations have been accumulated, the deformation of the bulk grains
nominal velocity of flow, the rate of target's removal for low impact would proceed. Formation of these layered structures of sub-surface by
angles is faster than that of higher impact angles [22]; it follows that grain refinement was previously described by Umemoto [67] and Ra-
the increase in particle energy and frequency of impacts are the main jahram et al. [68] who associated the effect with processes of localized
contributors to the E-C rate associated with nominal flow velocity. but severe plastic deformation. This explanation is consistent with ob-
The extension of plastic deformation in the sub-surface is observed serving sharp and irregular craters that at the high velocity (Fig. 6),
in the EBSD analysis (Fig. 9), visualizing the differences produced at the which can be attributed to strain hardening of the surface layer and
two velocities of exposure. Samples worn at 3 m⋅s−1 show marks of thus reduced the ability of the target material to deform at impact,
plastic deformation in a reduced section of the sub-surface, all of which favoring fracture over deformation.
have been produced by the impact of particles conveying enough ki- In addition to the erosive mechanisms of materials removal through
netic energy and also impacting at the proper orientation to deform the mechanical damage described above, the contribution of the slurry
material. Whereas, in case of surfaces worn at 7 m⋅s−1 the damage in- velocity through electrochemical corrosion should be considered. This
duced by plastic deformation appear to accumulate in the sub-surface, additional mechanism is evidenced by higher corrosion current density
which can be explained by the simultaneous increase of particle's ki- at the higher slurry velocity (Fig. 5), which results in an increment of
netic energy, the frequency of impact and the impact angle of the the corrosion rate of the material (Table 4). In this case, although the

9
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 10. EBSD analysis of the sub-surface microstructure at different levels of oxygen at 7 m⋅s−1: Image quality (top row); KAM analysis (bottom row).

nominal DO is 0 ppm, its effective value varies between 0 and 0.3 ppm, which might be explained by localized-type of corrosion as suggested
allowing to expect the main electrochemical reactions at the measured by the visual appearance of the resulting surface (Fig. 7). Localized
overpotentials to be oxidation of steel (Fe ↔ Fe2++ 2ē) and the re- corrosion in the presence of oxygen has been reported for active metals
duction of oxygen in the slurry (O2 + 4 H+ + 4ē ↔ 2H2O). Thus, on associated with the breakdown of stable corrosion product film of
the one hand, higher velocity implies faster transport of electroactive Fe2CO3 generated in slurries containing dissolved CO2 [51,52]; how-
species towards and from the electrolyte [64], shifting the reactions ever, this is not the case here. In this study, both physical and chemical
towards activation control. On the other hand, the increment in cor- modification of surface are observed in association with the presence of
rosion current can be explained by a greater surface roughness pro- oxygen. Figs. 7 and 8 show two clearly distinctive areas at the top
duced at the higher velocity providing a larger area of fresh material surface: i) dimples, several times larger than the indents observed after
and the possible preferential activation of the strain-hardened zones exposure without oxygen, characterized by predominantly high con-
[69,70]. Both explanations are plausible as indicated by the increase in centrations of oxides and embedded particles; and ii) remaining top
cathodic and anodic currents by a similar amount at given overpotential surface of scar morphology equivalent to that described previously but
in Fig. 5. with sparsely distributed fine erodent particles present in addition.
Finally, the effect of velocity can be summarized by the conceptual Because the grains present inside the dimples consist of incomplete
model shown in Fig. 11. grains, they are interpreted as caused by electrochemical dissolution
and are thus associated with anodic areas. The remaining surface is
then interpreted to be cathodic, giving rise to a localized-type of cor-
4.2. Effect of dissolved oxygen on E-C mechanisms rosion through a galvanic coupling of the macroscopic anodic and
cathodic areas. In addition to these macroscopic features indicative of
The presence of oxygen in the slurry promotes electrochemical galvanic coupling, another couple can be identified at the microscopic
corrosion of the material, increasing significantly the E-C rate (Fig. 4),

10
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 11. Schematic representation of the effect of slurry velocity on the mechanism of E-C at very low concentration of oxygen (nominal DO 0 ppm).

level, i.e., between eroded and uneroded top surface, which is clearly areas associated with particle impacts, which were deduced in the ni-
less affected by the corrosion process. trogen purged electrolyte. Although the impact of erosive particles
The increase of corrosion rate in the presence of oxygen (Table 4) is generates a work-hardened layer, rendering the top surface more sus-
explained by the same electrochemical reactions of oxidation (Fe ↔ ceptible to brittle fracture by the subsequent impingement of particles,
Fe2++ 2ē) and reduction (O2 + 4 H+ + 4ē ↔ 2H2O), however, in the abundant presence of oxygen no plastic deformation is observed
considering the higher concentration of oxygen. The polarization curves at the anodic sub-surface. The incomplete grains at the bottom of the
recorded in exposure to the flowing slurry (Fig. 5b) show increased dimple in Fig. 10 (10 ppm DO) suggest that the electrochemical dis-
cathodic current density in abundance of oxygen, while anodic current solution of grains is the case rather than extraction of entire grains as
density is not greatly affected. Thus, the reaction system can be con- suggested by Neville et al. (1999) [72]. On the other hand, the sub-
sidered under cathodic control through the availability of oxygen at the surface of the cathodic zones might be expected to be equivalent to that
electrode's surface. The relationship between the current density in observed at 0 ppm level of DO for the same conditions of erosion in
terms of diffusional limiting current for an RCE system (JLRCE) can be terms of particles' content, shape, size, and momentum, resulting in
estimated by the Levich relation [71], which was adapted for the geo- formation of layers of plastic deformation of decreasing grain size to-
metry used in this study given by Eq. (2): wards the bulk. However, the extent of plastic deformation in these
cathodic areas is greatly reduced, suggesting that corrosion accelerated
JLRCE = 0.069z·F·cb ·v −0.344 D0.644 U 0.7 (2) by the presence of oxygen prevents the formation of a work-hardened
layer on the surface. Such a mechanism could be attributed to elec-
where cb is the bulk concentration of electroactive species, D - diffusion trochemically induced modification in steel's hardness, facilitating the
coefficient of the species, z - charge number, F - Faraday constant, v - removal of the material as reported by Lu and co-workers [11,31]. In
kinematic viscosity of the electrolyte, and U - peripheral velocity of the this case, the kinetic energy of the particles is conveyed directly into
RCE. Then, at a constant velocity, greater availability of DO in the material's removal, rather than accumulating it as plastic deformation
slurry increases the diffusional limiting current resulting in a faster in the sub-surface.
material degradation by the electrochemical action. On the other hand,
the shape of the anodic branch is not modified significantly by a greater
DO content in the slurry, which suggests that the localized-type of 4.2.1. Localized corrosion in the presence of oxygen
surface damage is related to the erosive effect rather than electro- In order to elaborate on the nature of the localized corrosion, an
chemical degradation associated with possible breakdown of a film of additional exposure run was implemented, analogous with the high
corrosion products on the surface. This idea is discussed in detail in velocity and high oxygen content but without particles. In this run,
Section 4.2.1. degradation by corrosion only is the case, but no rate data can be ex-
Generation of the macroscopic anodic and cathodic areas seem to tracted because the flow conditions are not exactly the same, serving
involve mechanisms different from the microscopic anodic and cathodic only for qualitative evaluation of the damage type. Fig. 12 shows that

11
J. Aguirre, et al. Wear 438-439 (2019) 203053

Fig. 12. SEM micrographs of a sample exposed to pure corrosion: (a) Top surface and (b) Cross-section. Exposure conditions: V = 7 m⋅s−1, T = 45 °C, pH 5,
DO = 10 ppm, no particles.

the main observable damage is by uniform corrosion with a buildup of steel after exposure to the mineral slurry at different velocities and
corrosion products indicating that the localized-type corrosion at the contents of dissolved oxygen, along with electrochemical and mass loss
higher DO content in the slurry only occurs in the presence of particles. analyses, allows formulating the following insights into the mechanisms
The synergistic effect between erosion and corrosion is thus deduced to of erosion-corrosion:
be at least partially explained by the mechanism of erosion-enhanced
corrosion. However, the contribution of corrosion-enhanced erosion 1. The erosive effect of slurry velocity is associated with grain refine-
should not be discarded because the same increase of DO level causes a ment and localized accumulation of plastic deformation in the sub-
more significant increase of total E-C rate (Fig. 4), by the factor of 2.6, surface. The increment of flow velocity increases the frequency of
as compared with the increase in corrosion rate (Table 2), by the factor impacts and the energy conveyed per particle through the normal
of 1.4, under the same exposure conditions. However, in the present component of the effective impact velocity, both of which are the
study, it is not possible to evaluate the relative contributions of erosion- main contributors to removing the material. In addition, a syner-
enhanced corrosion and corrosion-enhanced erosion to the total E-C. getic contribution of velocity through corrosion is produced by
Xie et al. [73,74] have shown that plastically deformed areas are faster transport of electroactive species, the increased surface area
more anodic and therefore are more susceptible to corrosion than non- for the electrochemical reaction to take place and possible activation
deformed areas. Such an anodic activation could be expected con- of the strain-hardened zones.
sidering the change of the surface energy of the exposed area associated 2. The effect of dissolved oxygen in the slurry, which significantly in-
with strain-hardening. In this work, considering a highly corrosive creases the effective wear rate of steel, is associated with corrosion
environment, it is suggested in the initial stages of damage by erosion- of localized-type different from pitting. In the abundant presence of
corrosion, areas plastically deformed at random by the impact of par- oxygen, the accumulation of a work-hardened layer in the sub-sur-
ticles in the steel would create areas of localized charge on the surface, face is either suppressed or the layer is removed preferentially
originating the dissolution of these zones preferably. Thus, areas in- during exposure. Finally, the mechanism of E-C wear is explained, at
itially deformed plastically would be used as incubation zones for the least in part, by erosion-enhanced corrosion, in which plastic de-
growth of the dimples during the process. A schematic representation is formation induced by particles hitting onto the surface favors the
presented in Fig. 13. It is not clear if this behavior remains active during formation of anodic and cathodic sites. The anodic sites are an order
the entire period of exposure. Further investigation should be done to of magnitude larger than the area of the initial particle-target con-
clarify this behavior. tact.

5. Conclusions Acknowledgments

The inspection of microstructure formed in the sub-surface of X65 This work has been funded by Pontifica Universidad Católica de

Fig. 13. Schematic representation of the effect of dissolved oxygen contained in the slurry on the mechanism of E-C.

12
J. Aguirre, et al. Wear 438-439 (2019) 203053

Chile, Grant Puente N° P1814, and by CONICYT through the doctoral erosion-corrosion of pure metals in aqueous slurries, Tribol. Int. 53 (2012) 35–44,
scholarships CONICYT-PCHA/Doctorado Nacional/2015–21150171 https://doi.org/10.1016/j.triboint.2012.04.010.
[24] R. Dasgupta, B. Prasad, A. Jha, Effect of sand concentration on slurry erosion of
and FONDEQUIP EQM160091. The authors are thankful to . Remigiusz steels, Mater. Trans. 39 (1998) 1185–1190.
Michalczewski of ITeE-Radom (Sieć Badawcza Łukasiewicz) for fruitful [25] S. Turenne, M. Fiset, J. Masounave, The effect of sand concentration on the erosion
discussions made possible by CONICYT through REDES170071. Also, of materials by a slurry jet, Wear 133 (1989) 95–106, https://doi.org/10.1016/
0043-1648(89)90116-6.
Cauê Correa and Dr. Waldemar Krieger from MPIE are gratefully ac- [26] M.A. Islam, Z. Farhat, Erosion-corrosion mechanism and comparison of erosion-
knowledged for their indispensable contributions to part of the ex- corrosion performance of API steels, Wear 376–377 (2017) 533–541, https://doi.
perimental work. org/10.1016/j.wear.2016.12.058.
[27] B.R. Tian, Y.F. Cheng, Electrochemical corrosion behavior of X-65 steel in the si-
mulated oil sand slurry. I: effects of hydrodynamic condition, Corros. Sci. 50 (2008)
Appendix A. Supplementary data 773–779, https://doi.org/10.1016/j.corsci.2007.11.008.
[28] Y. Yang, Y.F. Cheng, Parametric effects on the erosion-corrosion rate and me-
chanism of carbon steel pipes in oil sands slurry, Wear 276–277 (2012) 141–148,
Supplementary data to this article can be found online at https://
https://doi.org/10.1016/j.wear.2011.12.010.
doi.org/10.1016/j.wear.2019.203053. [29] M.M. Stack, S. Zhou, R.C. Newman, Effects of particle velocity and applied potential
on erosion of mild steel in carbonate/bicarbonate slurry, Mater. Sci. Technol. 12
References (1996) 261–268, https://doi.org/10.1179/mst.1996.12.3.261.
[30] T. Alam, M. Aminul Islam, Z.N. Farhat, Slurry erosion of pipeline steel: effect of
velocity and microstructure, J. Tribol. 138 (2015) 021604, , https://doi.org/10.
[1] API-5L/ISO-3183, specification for line pipe, Am. Pet. Inst. (2007). 1115/1.4031599.
[2] R.J.K. Wood, 2.15 Tribocorrosion, Shreir's Corros, (2010), p. 1018 https://doi.org/ [31] B.T. Lu, J.L. Luo, Correlation between surface-hardness degradation and erosion
10.1016/B978-044452787-5.00041-X. resistance of carbon steel—effects of slurry chemistry, Tribol. Int. 83 (2015)
[3] A. Gnanavelu, N. Kapur, A. Neville, J.F. Flores, An integrated methodology for 146–155, https://doi.org/10.1016/j.triboint.2014.11.008.
predicting material wear rates due to erosion, Wear 267 (2009) 1935–1944, [32] G.R. Desale, B.K. Gandhi, S.C. Jain, Slurry erosion of ductile materials under normal
https://doi.org/10.1016/j.wear.2009.05.001. impact condition, Wear 264 (2008) 322–330, https://doi.org/10.1016/j.wear.
[4] C. Wang, A. Neville, S. Ramachandran, V. Jovancicevic, Alleviation of erosion- 2007.03.022.
corrosion damage by liquid-sand impact through use of chemicals, Wear 258 (2005) [33] Y.I. Oka, M. Matsumura, T. Kawabata, Relationship between surface hardness and
649–658, https://doi.org/10.1016/j.wear.2004.09.038. erosion damage caused by solid particle impact, Wear 162–164 (1993) 688–695,
[5] A. Neville, M. Reyes, H. Xu, Examining corrosion effects and corrosion/erosion https://doi.org/10.1016/0043-1648(93)90067-V.
interactions on metallic materials in aqueous slurries, Tribol. Int. 35 (2002) [34] H.X. Guo, B.T. Lu, J.L. Luo, Response of surface mechanical properties to electro-
643–650, https://doi.org/10.1016/S0301-679X(02)00055-5. chemical dissolution determined by in situ nanoindentation technique,
[6] S. Huth, Tribocorrosion of passive metals and coatings, in: D.,M.S. Landolt (Ed.), Electrochem, Commun. Now. 8 (2006) 1092–1098, https://doi.org/10.1016/j.
Tribocorrosion Passiv. Met. Coatings, 2011, pp. 265–295, , https://doi.org/10. elecom.2006.04.020.
1533/9780857093738.2.265. [35] H.C. Meng, K.C. Ludema, Wear models and predictive equations: their form and
[7] M.M.A. Stack, G.H. Abdulrahman, Tribology International Mapping erosion-corro- content, Wear 181–183 (1995) 443–457, https://doi.org/10.1016/0043-1648(95)
sion of carbon steel in oil exploration conditions : some new approaches to char- 90158-2.
acterizing mechanisms and synergies, Tribiology Int 43 (2010) 1268–1277, https:// [36] S.S. Rajahram, T.J. Harvey, R.J.K. Wood, Full factorial investigation on the erosion-
doi.org/10.1016/j.triboint.2010.01.005. corrosion resistance of UNS S31603, Tribol. Int. 43 (2010) 2072–2083, https://doi.
[8] B.T. Lu, J.F. Lu, J.L. Luo, Erosion-corrosion of carbon steel in simulated tailing org/10.1016/j.triboint.2010.05.012.
slurries, Corros. Sci. 53 (2011) 1000–1008, https://doi.org/10.1016/j.corsci.2010. [37] B.K. Gandhi, S.N. Singh, V. Seshadri, Study of the parametric dependence of erosion
11.034. wear for the parallel flow of solid-liquid mixtures, Tribol. Int. 32 (1999) 275–282,
[9] R.J.K. Wood, S.P. Hutton, The synergistic effect of erosion and corrosion: trends in https://doi.org/10.1016/S0301-679X(99)00047-X.
published results, Wear 140 (1990) 387–394, https://doi.org/10.1016/0043- [38] J. Aguirre, M. Walczak, Multifactorial study of erosion–corrosion wear of a X65
1648(90)90098-U. steel by slurry of simulated copper tailing, Tribol. Int. 126 (2018) 177–185, https://
[10] M.A. Islam, Z.N. Farhat, The synergistic effect between erosion and corrosion of API doi.org/10.1016/j.triboint.2018.04.029.
pipeline in CO2 and saline medium, Tribol. Int. 68 (2013) 26–34, https://doi.org/ [39] H.M. Clark, Particle velocity and size effects in laboratory slurry erosion mea-
10.1016/j.triboint.2012.10.026. surements OR... do you know what your particles are doing? Tribol. Int. 35 (2002)
[11] H.X. Guo, B.T. Lu, J.L. Luo, Interaction of mechanical and electrochemical factors in 617–624, https://doi.org/10.1016/S0301-679X(02)00052-X.
erosion-corrosion of carbon steel, Electrochim. Acta 51 (2005) 315–323, https:// [40] H.M. Clark, On the impact rate and impact energy of particles in a slurry pot erosion
doi.org/10.1016/j.electacta.2005.04.032. tester, Wear 147 (1991) 165–183, https://doi.org/10.1016/0043-1648(91)
[12] B.T. Lu, J.L. Luo, Synergism of electrochemical and mechanical factors in erosion - 90127-G.
Corrosion, J. Phys. Chem. B 110 (2006) 4217–4231, https://doi.org/10.1021/ [41] E. Akbarzadeh, E. Elsaadawy, A.M. Sherik, J.K. Spelt, M. Papini, The solid particle
jp051985f. erosion of 12 metals using magnetite erodent, Wear 282–283 (2012) 40–51,
[13] H. Meng, X. Hu, a. Neville, A systematic erosion-corrosion study of two stainless https://doi.org/10.1016/j.wear.2012.01.021.
steels in marine conditions via experimental design, Wear 263 (2007) 355–362, [42] H.M.I. Clark, The influence of the flow field in slurry erosion, Wear 152 (1992)
https://doi.org/10.1016/j.wear.2006.12.007. 223–240, https://doi.org/10.1016/0043-1648(92)90122-O.
[14] M. Lindgren, J. Perolainen, Slurry pot investigation of the influence of erodent [43] V. Javaheri, D. Porter, V. Kuokkala, Slurry erosion of steel – review of tests , me-
characteristics on the erosion resistance of austenitic and duplex stainless steel chanisms and materials Slurry erosion of steel – review of tests , mechanisms and
grades, Wear 319 (2014) 38–48, https://doi.org/10.1016/j.wear.2014.07.006. materials, Wear 408–409 (2018) 248–273, https://doi.org/10.1016/j.wear.2018.
[15] G.R. Desale, B.K. Gandhi, S.C. Jain, Effect of erodent properties on erosion wear of 05.010.
ductile type materials, Wear 261 (2006) 914–921, https://doi.org/10.1016/j.wear. [44] K.K. Wong, H.M. Clark, A model of particle velocities and trajectories in a slurry pot
2006.01.035. erosion tester, Wear 160 (1993) 95–104, https://doi.org/10.1016/0043-1648(93)
[16] H. Arabnejad, S.A. Shirazi, B.S. McLaury, H. Subramani, L.D. Rhyne, The effect of 90409-F.
erodent particle hardness on the erosion of stainless steel, Wear 332–333 (2015) [45] X. Hu, a. Neville, The electrochemical response of stainless steels in liquid-solid
1098–1103, https://doi.org/10.1016/0043-1648(91)90263-T. impingement, Wear 258 (2005) 641–648, https://doi.org/10.1016/j.wear.2004.09.
[17] H.M. Clark, R.B. Hartwich, A re-examination of the “particle size effect” in slurry 043.
erosion, Wear 248 (2001) 147–161, https://doi.org/10.1016/S0043-1648(00) [46] R. Malka, S. Ně, D.A. Gulino, Erosion – Corrosion and Synergistic Effects in
00556-1. Disturbed Liquid-Particle Flow vol. 262, (2007), pp. 791–799, https://doi.org/10.
[18] G.B. Stachowiak, M. Salasi, W.D.A. Rickard, G.W. Stachowiak, The effects of par- 1016/j.wear.2006.08.029.
ticle angularity on low-stress three-body abrasion-corrosion of 316L stainless steel, [47] S. Zhou, M.M. Stack, R.C. Newman, Characterization of synergistic effects between
Corros. Sci. 111 (2016) 690–702, https://doi.org/10.1016/j.corsci.2016.06.008. erosion and corrosion in an aqueous environment using electrochemical techniques,
[19] B.D. Jana, M.M. Stack, Modelling impact angle effects on erosion-corrosion of pure Corros. Sci. 52 (1996) 934–946.
metals: construction of materials performance maps, Wear 259 (2005) 243–255, [48] J. Postlethwaite, M.H. Dobbin, K. Bergevin, The role of oxygen mass transfer in the
https://doi.org/10.1016/j.wear.2005.02.012. erosion-corrosion of slurry pipelines, Corrosion 42 (1986) 514–521, https://doi.
[20] N. Andrews, L. Giourntas, A.M. Galloway, A. Pearson, Effect of impact angle on the org/10.5006/1.3583060.
slurry erosion-corrosion of Stellite 6 and SS316, Wear 320 (2014) 143–151, https:// [49] B. Yu, D.Y. Li, A. Grondin, Effects of the dissolved oxygen and slurry velocity on
doi.org/10.1016/j.wear.2014.08.006. erosion-corrosion of carbon steel in aqueous slurries with carbon dioxide and silica
[21] G.T. Burstein, K. Sasaki, Effect of impact angle on the slurry erosion-corrosion of sand, Wear 302 (2013) 1609–1614, https://doi.org/10.1016/j.wear.2013.01.044.
304L stainless steel, Wear 240 (2000) 80–94, https://doi.org/10.1016/S0043- [50] J. Jiang, Y. Xie, M.A. Islam, M.M. Stack, The effect of dissolved oxygen in slurry on
1648(00)00344-6. erosion–corrosion of En30B steel, J. Bio- Tribol-Corrosion. 3 (2017) 1–10, https://
[22] M.A. Islam, Z.N. Farhat, Effect of impact angle and velocity on erosion of API X42 doi.org/10.1007/s40735-017-0105-0.
pipeline steel under high abrasive feed rate, Wear 311 (2014) 180–190, https://doi. [51] Y. Hua, R. Barker, A. Neville, The effect of O2 content on the corrosion behaviour of
org/10.1016/j.wear.2014.01.005. X65 and 5Cr in water-containing supercritical CO2 environments, Appl. Surf. Sci.
[23] C.G. Telfer, M.M. Stack, B.D. Jana, Particle concentration and size effects on the 356 (2015) 499–511, https://doi.org/10.1016/j.apsusc.2015.08.116.

13
J. Aguirre, et al. Wear 438-439 (2019) 203053

[52] R.A. De Motte, R. Barker, D. Burkle, S.M. Vargas, A. Neville, The early stages of Heinemann, Oxford, 1992 First.
FeCO 3 scale formation kinetics in CO 2 corrosion, Mater. Chem. Phys. 216 (2018) [65] N. Molina, J. Aguirre, M. Walczak, Application of FFT analysis for the study of
102–111, https://doi.org/10.1016/j.matchemphys.2018.04.077. directionality of wear scars in exposure to slurry flow of varying velocity, Wear
[53] A. Schwartz, M. Kumar, B. Adams, D. Field (Eds.), Electron Backscatter Diffraction 426–427 (2019) 589–595, https://doi.org/10.1016/j.wear.2019.01.013.
in Materials Science, Second, Springer, 2009. [66] R.E. Winter, I.M. Hutchings, Solid particle erosion studies using single angular
[54] National institutes of health, ImageJ (n.d.), https://imagej.nih.gov/ij/. particles, Wear 29 (1974) 181–194.
[55] ASTM international, standard practice for preparing , cleaning , and evaluating [67] M. Umemoto, Nanocrystallization of steels by severe plastic deformation, Mater.
corrosion test, Significance 90 (1999) 1–9, https://doi.org/10.1520/G0001- Trans. 44 (2003) 1900–1911, https://doi.org/10.2320/matertrans.44.1900.
03R11.2. [68] S.S. Rajahram, T.J. Harvey, J.C. Walker, S.C. Wang, R.J.K. Wood, Investigation of
[56] G.S. Frankel, M. Rohwerder, Electrochemical techniques for corrosion, in: A. Bard, erosion-corrosion mechanisms of UNS S31603 using FIB and TEM, Tribol. Int. 46
M. Stratmann (Eds.), Encycl. Electrochem. Wiley, 2007, pp. 687–723, , https://doi. (2012) 161–173, https://doi.org/10.1016/j.triboint.2011.05.012.
org/10.1002/9783527610426 First. [69] S. Ghosh, A.K. Singh, S. Mula, P. Chanda, V.V. Mahashabde, T.K. Roy, Mechanical
[57] ASTM, Astm G102 - 89 standard practice for calculation of corrosion rates and properties, formability and corrosion resistance of thermomechanically controlled
related information from electrochemical measurements, ASTM Int. 89 (2010) 1–7, processed Ti-Nb stabilized IF steel, Mater. Sci. Eng. A 684 (2017) 22–36, https://
https://doi.org/10.1520/G0102-89R10.2. doi.org/10.1016/j.msea.2016.12.034.
[58] P.J. Konijnenberg, S. Zaefferer, D. Raabe, Assessment of geometrically necessary [70] W. Kingkam, C.-Z. Zhao, H. Li, H.-X. Zhang, Z.-M. Li, Hot deformation and corrosion
dislocation levels derived by 3D EBSD, Acta Mater. 99 (2015) 402–414, https://doi. resistance of high-strength low-alloy steel, Acta Metall. Sin. (Engl. Lett.) 32 (2018)
org/10.1016/j.actamat.2015.06.051. 495–505, https://doi.org/10.1007/s40195-018-0797-2.
[59] J. Jiang, T.B. Britton, A.J. Wilkinson, Measurement of geometrically necessary [71] F.C. Walsh, G. Kear, A.H. Nahlé, J.A. Wharton, L.F. Arenas, The rotating cylinder
dislocation density with high resolution electron backscatter diffraction: effects of electrode for studies of corrosion engineering and protection of metals—an illu-
detector binning and step size, Ultramicroscopy 125 (2013) 1–9, https://doi.org/ strated review, Corros. Sci. 123 (2017) 1–20, https://doi.org/10.1016/j.corsci.
10.1016/j.ultramic.2012.11.003. 2017.03.024.
[60] R.R. Shen, P. Efsing, Overcoming the drawbacks of plastic strain estimation based [72] A. Neville, T. Hodgkiess, H. Xu, An electrochemical and microstructural assessment
on KAM, Ultramicroscopy 184 (2018) 156–163, https://doi.org/10.1016/j. of erosion-corrosion of cast iron, Wear 233–235 (1999) 523–534, https://doi.org/
ultramic.2017.08.013. 10.1002/1521-4176(200201)53:1<5::AID-MACO5>3.0.CO;2-R.
[61] I. Finnie, Erosion of surfaces by solid particles, Wear 3 (1960) 87–103, https://doi. [73] J. Xie, a.T. Alpas, D.O. Northwood, Mechano-electrochemical effect between ero-
org/10.1016/0043-1648(60)90055-7. sion and corrosion, J. Mater. Sci. 38 (2003) 4849–4856, https://doi.org/10.1023/
[62] J.G.A. Bitter, A study of erosion phenomena. Part I, Wear 6 (1963) 5–21. B:JMSC.0000004404.53054.b4.
[63] J.G.A. Bitter, A study of erosion phenomena. Part II, Wear 6 (1963) 169–190, [74] J. Xie, A.T. Alpas, D.O. Northwood, The effect of erosion on the electrochemical
https://doi.org/10.1016/0043-1648(63)90073-5. properties of AISI 1020 steel, J. Mater. Eng. Perform. 12 (2003) 77–86, https://doi.
[64] I. Hutchings, Tribology: Friction and Wear of Engineering Materials, Butterworth- org/10.1361/105994903770343510.

14