Corrosion Science 111 (2016) 690702

Contents lists available at ScienceDirect

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

The effects of particle angularity on low-stress three-body

abrasion-corrosion of 316L stainless steel
Grazyna B. Stachowiak a , Mobin Salasi a, , William D.A. Rickard b , Gwidon W. Stachowiak a

a
Tribology Laboratory, Department of Mechanical Engineering, School of Civil and Mechanical Engineering, Curtin University, Bentley, WA 6102, Australia
b
John de Laeter Centre, Curtin University, Bentley, WA 6102, Australia

a r t i c l e i n f o a b s t r a c t

Article history: This study investigates the effects of particle angularity on the tribo-electrochemical behaviour of 316L
Received 26 May 2016 stainless steel under low-stress three-body abrasioncorrosion conditions. Different angularity particles
Accepted 8 June 2016 were delivered to the interface between the samples and a rotating rubber counterface, while changes
Available online 9 June 2016
in current with time were potentiostatically monitored. Based on a linear relationship existed between
tribological factors (load and speed) and the average current evolved, an empirical correlation is proposed
Keywords:
and possible underlying mechanisms are discussed. The effects of particle angularity on the surface and
A. Stainless steel
sub-surface damage are also demonstrated. It was found that increasing particle angularity improved
B. Polarization
B. Potentiostatic
their depassivation efciency.
C. Passive lms 2016 Elsevier Ltd. All rights reserved.

1. Introduction Abrasion-corrosion, the second category of tribocorrosion, is

Superior corrosion resistance of stainless steels is due to the
presence of a nano-scale passive lm on their surface [13]. The
passive lm is, however, prone to localised breakdown as a result
of chemical [46] and/or mechanical attack [79], leading to accel-
erated corrosion. Tribocorrosion, dened as the conjoint action
of tribological wear (mechanical damage) and corrosion (electro-
chemical dissolution), is a frequently encountered cause of the
passive lm breakdown [10,11]. Over the past two decades, various
tribo-electrochemical methods have been successfully applied to
elucidate how mechanical abrasion or erosion accelerate corrosion
of passive stainless steels [1219].
Tribocorrosion is usually divided into two distinct categories:
i) erosion-corrosion, and ii) abrasion-corrosion [10]. In erosion-
corrosion, the particles (entrained in a owing solution) disturb
the integrity of the passive lm by impacting the surface. Pub-
lished results revealed the following critical mechanical factors that
determine the severity and rate of erosion-corrosion: impact angle
[8,2022], kinetic energy of the particles (affected by particle size,
density and impact velocity) [2327], particle hardness [28], shape
and the shattering strength [29].
Corresponding author.
E-mail addresses: mobin.salasi@curtin.edu.au, mobin.salasi@gmail.com
(M. Salasi).
further divided into two contact types: i) two-body; where two
solid bodies slide against each other, and ii) three-body; where
abrasive particles are trapped between two surfaces in relative
motion. In both cases the surfaces are loaded and the motion occurs
in a corrosive environment.
Fundamentals of the dynamic two-body abrasion-corrosion of
passive materials have been studied by several research groups,
e.g. Refs [15,3033]. It was shown that in a two-body abrasion-
corrosion contact (reciprocating or pin-on-disc) where an inert
counterface slides on a passive metal, the depassivated area is gen-
erated at a rate that depends on the sliding velocity, v, and the
applied normal force, F.
Three-body abrasion-corrosion, involving interactions between
loosely held coarse abrasive particles forced against the passive
lm, is yet to be well understood. The particles (third bodies)
trapped between two surfaces under load, abrade the passivating
material (which could be one or both of the surfaces) by rolling
or sliding across the interface. Studies published so far either
quantied the synergy between three-body abrasion and corro-
sion [34,35], or investigated the interactions between three-body
abrasion and electrochemical corrosion [19,36]. Under identical
electrochemical conditions, the damage to the passive lm in three-
body abrasion-corrosion can be affected by several interconnected
variables such as the load exerted on the particles, frequency of
particle-surface interactions, load-bearing ability of the passive

http://dx.doi.org/10.1016/j.corsci.2016.06.008
0010-938X/ 2016 Elsevier Ltd. All rights reserved.
 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702
Fig. 1. Tribo-electrochemical test rig used in the three-body abrasion-corrosion experiments.
lm, hardness of the counteracting surfaces, and particles charac-
teristics such as hardness, size and angularity.
A signicant factor in controlling three-body abrasive wear rate
is angularity (or sharpness) of the abrading particles. This study
investigates how the particle angularity affects the electrochem-
ical corrosion of 316L stainless steel at different normal loads
and counterface tangential speeds, under low-stress three-body
abrasion-corrosion. It is believed that understanding the relation-
ship between the mechanical factors and the current evolved is
instrumental in developing a semi-mechanistic model for three-
body abrasion-corrosion contacts.
2. Experimental
2.1. Test rig and materials
Schematic of the tribo-electrochemical test rig is illustrated in
Fig. 1 while the details are described in Refs [19,36,37]. As shown
in Fig. 1, the test rig incorporates three electrodes necessary for
electrochemical testing, i.e. 316L stainless steel (SS316L) work-
ing electrode (WE1), platinum (Pt) counter electrode (CE) and
Pt pseudo-reference electrode (P-RE). A rubber ring counterface
(10 mm wide and 40 mm in diameter) rotates against a loaded at
wear sample (WE 1) while the abrasive slurry is delivered to the
contact interface. This contact conguration was found to simulate
low-stress three-body abrasion-corrosion, i.e. the particles remain
relatively intact [36].
Three types of abrasive particles were used: glass beads (GB),
silica sand (SS), and alumina (AO). Particle average size, Vickers
microhardness, and angularity (ASPQ ) of each abrasive type are
given in Table 1. Representative secondary electron SEM images
in Fig. 2a illustrate shape differences of the particles while the par-
ticle size distribution, measured by laser diffraction, is shown in
Fig. 2b.
It is generally accepted that the particle-size effect on abrasive
wear rate becomes pronounced for particles smaller than about
100150 m, and for larger particles abrasive wear is rather inde-
pendent of grit size [38,39]. This size effect was conrmed for
wear samples with different hardness and for different applied
loads. Therefore, small differences in the average sizes between the
three particle types (Table 1) are deemed not signicant in affect-
ing three-body abrasion-corrosion. On the other hand, the particle
shape, characterized in this study in terms of particle angularity, is
very important. For a given load on a particle, particle angularity
affects the groove area and the contact pressure.
691
The particles angularity (ASPQ ) was quantied using a spike
parameter-quadratic t (SPQ) method, developed by Hamblin and
Stachowiak [40]. For each particle, the method isolates the bound-
ary features protruding outside the average-radius circle of the
particle. These protrusions, called spikes, are most likely to cause
abrasion. Quadratic segments are then tted to both sides of each
spike and differentiated; the cosine of half the apex angle, called
spike value (SV), is used as a measure of the spikes angularity.
The SPQ value for a particle is averaged over all individual SVs [41].
For each type of abrasive particles used the SPQ values of at least 50
particles were calculated and the average value is denoted as ASPQ .
Particle hardness, another parameter considered in abrasion-
corrosion, signicantly affects the wear rate only when the ratio r
of particle hardness to wear sample hardness is between 0.8 and
1.5 [42,43]. Since in all tests the ratio r was higher than 2 (the hard-
ness of SS316L was 210 HV), the abrasive wear rate is presumed to
be independent of the particle hardness. Therefore, any differences
in the wear rates, or abrasion efciency, between the three types
of particles used should be affected predominantly by the differ-
ences in their angularity. The low-stress abrasion conditions were
selected to avoid fracture of the loaded particles that would alter
their original angularity values.
The SS316L wear samples, mounted in an epoxy resin, were
ground using SiC papers to a 1200 grit nish, resulting in an aver-
age surface roughness Ra of 0.03 m. Prior to testing, the samples
were rinsed with ethanol and ultrasonically cleaned for 3 min in
deionised water. A nominal surface area of 1.13 cm2 was exposed to
1.5 L of an open-to-air 3.5 wt% NaCl electrolyte. Temperature of the
solution was 24 (2) C. Unbuffered solution (pH 6.2) was used for
the tests. It was found, however, that the presence of GB increased
the pH of the unbuffered solution from 6.2 to 11. To investigate the
effect of pH change on the current evolved, verication tests were
conducted where the solution pH was buffered at 8.35 (0.03) by
adding 0.01 M sodium bicarbonate. No signicant differences in the
current evolved were found between the buffered and unbuffered
solutions indicating that the GB effect on the solution pH did not
change the tribo-electrochemical results.
2.2. Tribo-electrochemical tests
The tribo-electrochemical tests were conducted potentiostati-
cally while potential of the sample was monitored against a long
Pt wire used as a pseudo-reference electrode (P-RE) [44,45]. Prior
to selecting Pt wire as the pseudo-reference (P-RE), its potential
in 3.5 wt% NaCl was measured against a double junction Ag/AgCl
692 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702
Table 1
Characteristics of abrasive particles used in the three-body abrasion-corrosion experiments.
Abrasives Average size (m) a
Glass beads (GB) 242
Silica sand (SS) 305
Alumina (AO) 359
a
Laser diffraction was used to measure the average size of at least 6 g of particles.
b
Vickers microhardness of at least 50 particles was measured.
c
ASPQ is the average SPQ value from at least 50 particles.
Fig. 2. a) Representative secondary electron SEM images of the particles used in the three-body abrasion-corrosion tests: glass beads (GB), silica sand (SS), and alumina (AO);
b) size distribution of the particles measured by laser diffraction.
(sat.) reference electrode for 7000 s and the results are shown in
Fig. A1 in the Appendix. Since the P-RE potential depends on the test
conditions (such as chemical composition and temperature of the
solution), before each test its potential was rst measured against
the double junction Ag/AgCl reference electrode for at least 300 s
(an example is shown in Fig. A2 in the Appendix). Then, during
abrasion-corrosion the wear sample potential was held constant
against the P-RE at the potential approximate to 0.1 V/SSE (0.02).
After the tests were nished, the potential of P-RE was re-measured
and differences of less than 20 mV were found. Constant potential
of approximately 0.1 V/SSE was chosen based on two potentio-
dynamic polarization curves of SS316L, shown in Fig. A3 in the
Appendix, measured using either Ag/AgCl (sat) or Pt wire as the
standard and pseudo-reference electrodes, respectively. Since no
incidents of metastable pitting are seen in the potentiodynamic
curves in Fig. A3 at 0.1 V/SSE, and also previous studies [36] show
that this potential is below the repassivation potential of 316L
stainless steel, material repassivation after mechanical/chemical
lm breakdown can be assumed. The results of all these tests con-
rm stability of the P-RE potential under testing conditions and
suitability of using the Pt wire as the pseudo-reference.
The potentiostatic tests were preceded by 2 min of cathodic
pre-conditioning (at approx. 0.8 V/SSE) aimed at reducing the
thickness of oxide lm formed in air (note that in neutral solu-
tions complete cathodic stripping of the oxide is not practically
possible [46]). Following pre-conditioning, the sample was anod-
ically polarized at the desired potential, and once a stable anodic
current was attained, load was applied and the abrasion-corrosion
test was commenced. The current with time changes were recorded
at a sampling frequency of 5 Hz. The current evolved due to abra-
sion was calculated by subtracting the corrosion-only current from
the average current measured during at least 180 s of steady-state
abrasion-corrosion.
The mechanical test conditions were selected with aim to inves-
tigate the dynamic disruption and regeneration of the passive lm.
Nominal normal loads ranging between 0.5 and 2.5 N for GB and
between 0.5 and 10 N for SS and AO were applied to the wear
samples, as shown in Fig. 1. The counterface rotational speed
was controlled to achieve the targeted tangential speeds of 30,
61, 122, and 204 mm/s while the abrasives were delivered to the
Microhardness (HV0.1 ) SD b ASPQ SD c
490 23 0.03 0.05
1220 54 0.22 0.09
2020 130 0.52 0.16
contact interface. Biologic SP300 potentiostat was used for the
tribo-electrochemical measurements.
2.3. Surface and sub-surface examination
After abrasion-corrosion tests, all worn surfaces were exam-
ined using optical prolometry (Altisurf 530), followed by scanning
electron microscopy (SEM). The sub-surface deformation due to
abrasion-corrosion was studied using a Tescan Lyra focused ion
beam (FIB) and SEM dual-beam platform. Prior to FIB milling, a thin
platinum layer was deposited on the surface. Finally, an orthogonal
time-of-ight secondary ion mass spectroscopy (TOF-SIMS) (Tofw-
erk, Germany) system, utilising the FIB, was used to determine if
there was material transfer between the particles and the abraded
surface.
3. Results
3.1. Tribo-electrochemical tests
3.1.1. Effect of abrasion time on current
Current changes with time were recorded before, during and
after delivery of the abrasive particles to the interface. The abrasion-
corrosion tests were continued for approximately 1800 s (30 min)
to investigate the effect of abrasion time. The current with time
trends for the tests with GB, SS and AO particles are shown in Fig. 3;
the tests were conducted at 0.5 N load, 61 mm/s tangential speed,
while the potential was held constant against the P-RE (Pt wire)
to achieve approx. 0.1 V/SSE (0.02). It can be seen in Fig. 3 that
the curves exhibit two distinct stages: 1) running-in and 2) steady
state. Running-in is distinguished by a sudden current rise at the
onset of abrasion, followed by a gradual current decay until a steady
state is reached. The running-in stage was only prominent at 0.5 N
load; it was not visible at higher loads. Once unloaded, the anodic
current rapidly returns to the pre-abrasion values, indicating that
the samples have repassivated.
Although at the steady state the measured anodic current oscil-
lates frequently, the average current remains constant relative to
time. For the tests with GB particles, occasional higher amplitude
current spikes are also visible. Fig. 3 shows that the steady state
 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702 693

140
Abrasion (1) Alumina
(1)ca Sand
(2) Sili
started
120 (3) Glass beads

100 Abrasion

Current (A)
stopped
80

60
(1)

40

20 (2)

(3)
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200
Time (s)
Fig. 3. Current changes with time, measured potentiostatically in the presence of different abrasive particles at 0.5 N load and 61 mm/s counterface tangential speed.

a) 6 b) 2200
SS
2000 AO
5 Slope: 219.83
Slope: 2.39 1800 2
2 R =0.99
R =0.91 1600
4
Current (A)

Current (A)

1400
1200 Slope: 176.17
3 R2=0.99
1000
800
2
600

1 400
200
0 0
0.0 0.5 1.0 1.5 2.0 2.5 0 2 4 6 8 10
Load (N) Load (N)

Fig. 4. Relationship between the average current evolved and normal load measured potentiostatically under three-body abrasion-corrosion of SS316L at 61 mm/s tangential
speed in the presence of: a) glass beads (GB), b) silica sand (SS), and alumina (AO) particles.

amplitude of current oscillations and the average current evolved
are the functions of particle angularity.
3.1.2. Effect of load on current
Changes in the current evolved were measured at different loads
for the tests with GB, SS and AO particles. The values of the average
current evolved at each load at 61 mm/s tangential speed are plot-
ted in Fig. 4. For all three particle types, a proportional increase of
the average current with load is evident. At each load tested, abra-
sion by the angular AO particles has resulted in the highest amount
of current evolved.
3.1.3. Effect of tangential speed on current
Experiments were performed to determine the effect of tangen-
tial speed on the current evolved in the presence of GB, SS and AO
particles. The average currents evolved at constant loads are plot-
ted in Fig. 5 against tangential speed. For all three particle types,
a proportional increase of current with tangential speed was mea-
sured.
3.1.4. Contour maps of current
The combined effects of the mechanical variables (normal load
and tangential speed) on the current evolved for the three types of
particles tested are summarized in Fig. 6. It can be seen that the cur-
rent increases with increasing load and tangential speed for all the
particles; the value of current depends on the particle angularity.
3.2. Surface and sub-surface examination
In Fig. 7, the average roughness Ra values of the worn surfaces,
measured on conclusion of the potentiostatic abrasion-corrosion
experiments at 0.5 N load and 61 mm/s tangential speed (see Fig. 3),
are correlated with the particles ASPQ values. As shown in Fig. 7,
increasing the angularity of the abrasives resulted in rougher worn
surfaces. Small imperfections found on the GB surfaces (see Fig. 2a)
have contributed to the roughness increase.
Next, the effects of the abrasive particles angularity on the wear
surface morphology and sub-surface deformation of the SS316L
samples were studied using optical prolometry, SEM, FIB-SEM and
TOF-SIMS.
694 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702

a) 14 b) 700
0.5 N 0.5 N
1N 1N
12 Slope: 0.07 600 Slope: 3.56
2N 2N
R2=0.99 R2=0.97
10 500

Current (A)
Current (A)

Slope: 0.05
2
R =0.99
8 400
Slope: 1.63
2
6 300 R =0.99
Slope: 0.03
R2=0.97
4 200
Slope: 0.26
2 100 2
R =0.99

0 0
0 20 40 60 80 100 120 140 160 180 200 220 0 20 40 60 80 100 120 140 160 180 200 220
Speed (mm/s) Speed (mm/s)

c)

Fig. 5. Relationship between the average current evolved and tangential speed in three-body abrasion-corrosion of SS316L in the presence of: a) glass beads (GB), b) silica
sand (SS), and c) alumina (AO) particles.

3.2.1. Round GB particles
Representative secondary electron SEM images of the surfaces
worn during abrasion-corrosion tests at low and high loads are
compared in Fig. 8a and b, respectively. Shallow grooves and small
round indents generated by the GB particles at the low load of 0.5 N
can be seen in Fig. 8a. After an abrasion-corrosion test during which
the load was gradually increased to 2 N, surface grooving by indi-
vidual GB particles was no longer discernible, as shown in Fig. 8b.
Instead, there was evidence of superimposed plastic deformation
and delamination.
Optical prolometry measurements of the worn surface area
containing a shallow groove generated by a loaded GB particle are
shown in Fig. 9. A single prole across the groove in Fig. 9b indicates
that the round GB particle depressed the material underneath and
displaced it sideways. The prole shows a 1.7 m level difference
between the side and bottom of the groove.
Fig. 10 shows the section FIB milled across a shallow groove
generated by a round GB particle in the test using 0.5 N nor-
mal load and 61 mm/s tangential speed. The secondary electron
SEM image in Fig. 10a indicates that the original grinding marks
are attened inside the groove and there is evidence of material
smearing. Ion beam induced secondary electron (ISE) imaging was
used subsequently to study sub-surface deformation. ISE images
reveal orientation contrast between adjacent grains facilitating the
assessment of deformation to the crystal structure due to abra-
sion. The ISE image in Fig. 10b indicates that the crystal structure
beneath the groove has been deformed, especially in the middle of
the groove. Both Figs. 9 and 10 conrm that at the relatively low
load of 0.5 N the round GB particles deform the SS316L surface pre-
dominantly by shallow ploughing, with almost no actual material
removal.
For comparison, similar examination was carried out on the
SS316L surface after multiple passages of GB particles while the
load was increasing from 0.5 N to 2 N by steps (the corresponding
wear surface morphology can be seen in Fig. 8b at a lower mag-
nication). SEM image in Fig. 11a shows the crater FIB-milled in
the worn surface, and details of the sub-surface section are shown
in Fig. 11b and c. It is clear from Fig. 11 that prolonged abrasion-
corrosion by GB particles generated a highly deformed tribolayer
with a higher degree of deformation than that observed underneath
a single-passage groove at a lower load of 0.5 N (Fig. 10b). Evidence
of material compaction, extrusion, folding over and microfrac-
ture can be found in the images. Strain hardening decreased the
materials ductility, resulting in sub-surface microfracture and
delamination seen in Fig. 11b. The ISE channelling contrast image
in Fig. 11c indicates that repeated plastic deformation generated
highly-strained submicron-sized grains. The deformation extends
below the nely-grained layer, as revealed by the light-coloured
areas indicating distorted grains orientation.
Possible embedment of broken GB fragments in the worn sur-
face was investigated by mapping the concentration of silicon at
different depths below the surface using TOF-SIMS (within the FIB-
SEM instrument). It can be seen in Fig. 12 that the tribolayer is
enriched in Si which indicates material transfer between the GB
particles and the worn surface, or even an embeddment of very
small GB pieces.
 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702
Fig. 6. Current contour maps showing the combined effect of tangential speed and normal load in three-body abrasion-corrosion of SS316L in the presence of a) glass beads
(GB), b) silica sand (SS), and c) alumina (AO) particles.
Fig. 7. Correlation between the average roughness, Ra , values of the surfaces worn
during abrasion-corrosion and the angularity (ASPQ ) of the abrasive particles: glass
beads (GB), silica sand (SS), and alumina (AO).
695
3.2.2. Semi-round SS and angular AO particles
Representative secondary electron SEM images taken from the
centre of wear scars after abrasion-corrosion by SS and AO particles
are shown in Fig. 13. For both types of particles the SS316L surface
abraded at a low load of 0.5 N (Fig. 13a and b) exhibits irregular
morphology comprised of a mixture of indents and multidirectional
short grooves. Fig. 13c and d indicate that at higher loads grooving
wear dominates, especially for the angular AO particles. Since at the
higher loads the particles are pressed deeper into the rubber coun-
terface, which restricts their movement, the grooves align with the
direction of the rubber ring rotation.
To investigate sub-surface deformation generated by angular
AO particles, a cross-section was FIB-milled in the SS316L sur-
face after abrasion-corrosion at 0.5 N load and 61 mm/s tangential
speed. The SEM image of the FIB-milled sub-surface region in
Fig. 14a shows surface undulation, representing cross-sectional
proles of the grooves scratched by the AO particles, and an exam-
ple of wear particle detachment. The ISE image in Fig. 14b shows
a severely-strained top layer with nely-grained microstructure.
Grains distortion, characterized by the elongated white areas below
the nely-grained layer, indicates that plastic deformation extends
to about 4 m underneath the worn surface. Since microcutting
is not efcient at low loads, material is mainly removed by the
accumulation of plastic strain followed by microfracture.
696 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702

Fig. 8. Secondary electron SEM images of the SS316L surfaces after abrasion-corrosion by round glass beads at a) constant load of 0.5 N; and b) increasing load up to 2 N.

Fig. 9. Prolometry measurement of the SS316L surface abraded by round glass beads: a) worn surface containing a shallow groove; b) single prole across the groove (the
prole position is marked by an arrow in Fig. 9a). Test conditions: 0.5 N load and 61 mm/s tangential speed.

Fig. 10. a) Secondary electron SEM image of the FIB-milled section across a shallow groove generated by a single glass bead particle at 0.5 N load and 61 mm/s speed; b) ISE
image of the sub-surface area; (i) outside the groove, (ii) inside the groove, (iii) original grinding scratches, (iv) platinum layer deposited prior to FIB milling, (v) sub-surface
deformation due to abrasion-corrosion.

4. Discussion I = Rdep Qp (1)

4.1. Comparison of two- and three-body abrasion-corrosion According to a mechanistic model for two-body abrasion-
corrosion in a dynamic reciprocating contact, the rate of surface
Abrasive wear of stainless steel leads to destruction of the pro- depassivation is [30]:
tective passive lm by its displacement and/or removal from the  F n
surface. Under potentiostatic control, when the potential is held Rdep = kp v (2)
H
at the repassivation region, simultaneously with the destruction
of the lm the surface tries to repassivate. At the steady-state where kp is the proportionality factor, v is the sliding velocity, F
abrasion-corrosion, the anodic current evolved due to  abrasion,
 is the normal force, n varies between 0.5 and 1 depending on the
m2 contact type, and H is the hardness of the metal.
I (A), is equal to the rate of surface depassivation, Rdep s , mul-
 C
 For a rough hard counterface sliding on a at metal sample the
tiplied by the repassivation charge density, Qp , needed to model predicts a square root relationship between the depassiva-
m2
re-establish the destroyed passive lm: tion rate and the normal force [30,47]. However, when a rough
 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702 697

Fig. 11. a) Secondary electron SEM image of the crater FIB-milled in the worn surface after prolonged abrasion-corrosion by glass beads while the load was increasing up
to 2 N; b) higher magnication secondary electron image of the sub-surface area showing delamination; c) ISE image of the entire FIB-milled sub-surface section. The areas
shown in (b) and (c) are depicted by the rectangles in (a). (i) Platinum deposited layer, (ii) evidence of material delamination, (iii) nely-grained layer, (iv) distorted area and
(v) bulk material.

Fig. 12. a) ISE image of the worn surface (10 10 m) b) TOF-SIMS elemental map of Si at different depths beneath the worn surface.

Fig. 13. Secondary electron SEM images of the SS316L surface after abrasion-corrosion by: a) silica sand & b) alumina at 0.5 N; c) silica sand & d) alumina at the load increasing
up to 10 N; particles were delivered from the top.
698 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702

Fig. 14. FIB-milled cross-section of the SS316L surface after abrasion-corrosion by alumina particles at 0.5 N load and 61 mm/s tangential speed a) in-lens SEM image; b) ISE
image. (i) platinum deposited layer, (ii) sub-surface microcracks, (iii) nely-grained top layer, (iv) distorted region, (v) bulk material.

Fig. 15. Relationships between the average current evolved and the product of normal load and counterface tangential speed (FSt ) during abrasion-corrosion using a) glass
beads; b) silica sand; and c) alumina particles.

ceramic ball was sliding on a at stainless steel plate the experi-
mental data indicated that the n factor was equal to one [48]. In
such cases, since Qp is constant for a potentiostatically held pas-
sivating metal (at a certain potential and tribocontact condition)
[49], and hardness of the metal is assumed constant, Eq. (1) takes
the following form:
I = K  vF
where K
is the proportionality factor for the two-body recip-
rocating contact. The factor K  accounts for the efciency of
depassivation by an asperity, i.e. the probability of a contact caus-
ing the passive lm damage. The hardness of the metal and the
repassivation charge density are included in the K  factor.
In this study, under low-stress three-body abrasion-corrosion
conditions the average current evolved increased approximately
linearly with increasing both the load (normal force) and the coun-
terface tangential speed (see Figs. 4 and 5 in Results). The average
current data replotted in Fig. 15 against the product of load and tan-
(3)
gential speed, FSt , indicate linear correlations for all three abrasive
particles tested:
Iavg = KFSt (4)
 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702
a)
3.0
2.5
2.0
Slope
1.5
1.0
0.5
0.0
0.0 0.1 0.2 0.3 0.4
ASPQ
Fig. 16. Slopes of the lines in Fig. 15 (K factor) plotted against the particle angularity values (ASPQ ) on a) linear; b) log scales.
where Iavg is the average current evolved during the steady-state
three-body abrasion-corrosion, and K is a proportionality con-
stant representing the efciency of depassivation by the abrasive
particles. The depassivation efciency depends on the particle
angularity and the wear mechanisms acting. Similarly to the K  fac-
tor, hardness of the wearing surface and the repassivation charge
density are assumed constant and are included in the K factor.
The empirical correlation described by Eq. (4) is similar to that
formulated for a two-body abrasion-corrosion reciprocating con-
tact (Eq. (3)), despite differences in the phenomena occurring at
the contact interface. Some of the important aspects relevant to
the low-stress three-body abrasion-corrosion are outlined in the
following sections.
4.1.1. Oscillation of current around an average value
Oscillation of current was observed in the tribo-electrochemical
measurements of current evolved with time (see Fig. 3). During the
steady-state abrasion-corrosion, the average current evolved (Iavg )
was constant at each load and tangential speed tested. Although
the degree of depassivation by individual particles varies, it appears
that the average depassivation rate can be assumed constant at con-
stant tangential speed and normal load. The amplitude of current
oscillation was found to depend on the particles angularity; the
amplitude was reduced for the less angular abrasives such as GB or
SS compared to the highly angular AO particles. The current oscil-
lation can be related to the abrasives geometry and their alignment
inside the tribocontact. Since the particles are not xed in place but
just loosely gripped by the rubber counterface, they may approach
the wearing surface with varying attack angles. This variation in
the attack or approach angle is especially prominent in the case
of highly angular AO particles which have irregular geometry (see
Fig. 2a). Hence, the severity of depassivation would be different for
different AO particle alignment, giving rise to large oscillations of
the current measured. On the other hand, since all round GB parti-
cles contact the wearing surface at very similar attack angles, there
is little inuence of their alignment on the severity of depassiva-
tion. As a result, very small oscillations of the current measured are
observed in Fig. 3.
4.1.2. Wear mechanisms at different loads
The SEM images show that the abrasive particles have gener-
ated a mixture of grooves and indents on the worn SS316L surfaces.
At the lower loads the particles movement is less restricted and
the contribution of indentation to the overall damage is higher
than that at the higher loads. The grooves generated at the lower
699
b)
3.0
2.5
2.0
Slope
1.5
1.0
0.5
0.0
0.5 0.6 0.01 0.1 1
ASPQ
loads are short and randomly oriented. The dominance of grooving
wear increases with increasing the load. As the particles are forced
deeper into the rubber counterface, the grooves become longer and
better aligned with the direction of counterface rotation. At the
higher loads the constrained abrasive particles act in a way simi-
lar to the hard surface asperities in two-body abrasion. In grooving
wear, the passive layer and the underlying fresh steel surface are
displaced by ploughing or removed by microcutting. The dominat-
ing mechanism depends on the particle angularity and the load.
The particle angularity and the normal load affect the depth of par-
ticle penetration and the shape of the groove. Prolonged ploughing
results in the accumulation of plastic deformation and subsequent
delamination of the heavily-strained tribolayer. This is the domi-
nant depassivation mechanism when the round GB particles abrade
the SS316L surface.
The running-in stage shown in Fig. 3 was only observed at a
lower load of 0.5 N where the particles were less constrained. Pos-
sibly this stage reects the effect of the original surface roughness
on the surface damage and the disruption of passivity. At higher
loads the abrasives were more rmly gripped by the rubber and
were forced to groove the wearing surface. Therefore the inuence
of the wear samples initial surface roughness was less important
and the running-in stage was not observed.
4.1.3. Particle angularity
It is expected that the attack angle of abrasives (hence, particle
angularity) would have a major effect on the efciency of depassi-
vation (represented by K in Eq. (4)). Under the load conditions used,
the round GB particles either indent the SS316L surface or plough
it; both actions are accompanied by plastic ow. Therefore, the
low values of the steady-state current evolved under abrasion are
related to the passivity impairment or degradation due to repeated
material displacement by the GB particles. Accumulation of plas-
tic strain is needed for the GB particles to remove the material by
microfracture and delamination. The small imperfections on the GB
surface and/or minor microfracture of the tribolayer developed on
the surface could have given rise to the measured current spikes
(see Fig. 3).
The loaded angular AO particles, on the other hand, are fairly
effective in material removal. The material removal by microcutting
generates large fresh (unpassive) areas that give rise to a signi-
cant current discharge. Semi-round SS particles are less effective in
microcutting, especially at lower loads, but can generate grooves
deeper than those by GB particles. To determine the inuence
of particle angularity on the K factor, the slopes of the lines in
700 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702

100 become efcient and the substantial accumulation of plastic defor-

Glass beads mation is less likely. This means that changes to the original surface
Silica sand hardness and microstructure depend on the dominating wear
Aluminium oxide mechanisms. As a result, any mechanistic tribocorrosion model for-
mulated for three-body abrasion-corrosion should account for this
10 dependency.
Current (A)

5. Conclusions

1. Electrochemical corrosion of SS316L under three-body abrasion-

1 corrosion conditions, measured in terms of the average current
evolved, was proportional to the load and counterface tangential
speed. It was found empirically that a linear relationship exists
between the average current evolved and the product of normal
load and tangential speed.
2. As the angularity of the abrasive particles increased, the proba-
0.1
bility of passive lm removal/surface depassivation, represented
0.1 1 10 100
by the proportionality constant K, also increased. It was found
Time (s) that the proportionally constant K, exhibits a non-linear (loga-
rithmic) correlation with the particle angularity.
Fig. 17. Repassivation of the worn sample after stopping potentiostatic abrasion- 3. The time taken for surface repassivation after stopping abrasion-
corrosion using different angularity particles under the test conditions given in Fig. 3.
corrosion increases with increasing particle angularity.

Fig. 15 are plotted against the ASPQ values of the particles in Fig. 16. Acknowledgements
Comparison of the plots shown in Fig. 16a and b indicates that a log-
arithmic (rather than linear) correlation exists between the slope The authors wish to thank the Department of Mechanical Engi-
and the particles angularity (ASPQ ). The reason for this behaviour neering, Curtin University, for nancial support of this project. The
needs to be further investigated. authors acknowledge the use of Curtin Universitys John de Laeter
Centre, whose instrumentation has been partially funded by the
4.1.4. Mechano-chemical characteristics of the tribolayer University, State and Commonwealth Governments. Thanks also
Characteristics of the tribolayer, namely hardness and passiv- go to Dr. Tomasz Woloszynski for the surface prolometry mea-
ity, may inuence the depassivation and lm regeneration process. surements.
In the Results it was shown that a nely-grained tribolayer was
formed after prolonged abrasion-corrosion using either round GB
Appendix A
or angular AO particles. The formation of a nely-grained tribo-
layer was also found after erosion-corrosion [50,51] and two-body
abrasion-corrosion [52]. 0.30
It is possible that the dynamic passive lm formed during
abrasion-corrosion, had a modied structure and therefore differ-
0.25 pt vs Ag/AgCl
ent electrochemical properties. This can affect the rate of dynamic
depassivation/repassivation occurring under abrasion-corrosion. It
Potential (V vs Ag/AgCl)

is also evident from Fig. 17 that after stopping abrasion-corrosion 0.20

the time to repassivate the surface depends on the particle
angularitythe surface abraded by the particles with higher angu- 0.15
larity takes longer to repassivate.
The repassivated lm formed after abrasion-corrosion may 0.10
exhibit an improved or poorer pitting resistance. Past studies have
indicated that the nano-crystalline stainless steels exhibit better
0.05
passivity than stainless steels with larger grains [53], particu-
larly when exposed to passivating environments [54]. Different Ag/AgCl vs Ag/AgCl
0.00
mechanisms such as increased surface reactivity [55] or improved
diffusion of chromium through grain boundaries [56] are suggested.
Further research is, however, needed to determine whether a sim- -0.05
ilar phenomenon is relevant for the nano-crystalline tribolayer 0 1000 2000 3000 4000 5000 6000 7000

formed under three-body abrasion-corrosion conditions. Time (s)

Since the nely-grained tribolayer has lower ductility and
Fig. A1. a) Potential stability tests; the change in potential of Pt wire against the
higher hardness than that of the bulk material, its removal occurs
standard Ag/AgCl (sat) reference electrode compared to that of coupled Ag/AgCl
mainly by microfracture and delamination, especially when round (sat) standard reference electrodes.
GB particles are used. At higher loads and with the presence of
angular particles, however, material removal by microcutting can
 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702
0.290
0.288
Potential (V vs Ag/AgCl)
0.286
0.284
0.282
0.280
0 100 200 300 400 500 600
Time (s)
Fig. A2. Potential of the Pt wire measured against the Ag/AgCl (sat.) reference elec-
trode before starting each potentiostatic test.
Fig. A3. Potentiodynamic polarization curves of the 316L stainless steel tested in
3.5 wt% NaCl solution. The potential was swept at a scan rate of 1 mV/s against the
Ag/AgCl (sat.), and Pt reference electrodes from 0.1 V vs. OCP until the current had
reached 100 A. The arrows indicate the potential at which the potentiostatic tests
were conducted.
References
[1] C.R. Clayton, I. Olefjord, Passivity of austenitic stainless steels, in: P. Marcus, J.
Oudar (Eds.), Corros. Mech. Theory Pract, 1995, pp. 175199.
[2] C.O.A. Olsson, D. Landolt, Passive lms on stainless steels: chemistry,
structure and growth, Electrochim. Acta 48 (2003) 10931104.
[3] H.-H. Strehblow, P. Keller, XPS investigations of electrochemically formed
passive layers on Fe/Cr-alloys in 0.5 M H2SO4, Corros. Sci. 46 (2004)
19391952.
[4] D.E. Williams, M.R. Kilburn, J. Cliff, G.I.N.N. Waterhouse, Composition changes
around sulphide inclusions in stainless steels, and implications for the
initiation of pitting corrosion, Corros. Sci. 52 (2010) 37023716, http://dx.doi.
org/10.1016/j.corsci.2010.07.021.
[5] P.M. Natishan, W.E. OGrady, F.J. Martin, R.J. Rayne, H. Kahn, A.H. Heuer,
Chloride interactions with the passive lms on stainless steel, J. Electrochem.
Soc. 158 (2011) C7C10.
[6] H. Krawiec, V. Vignal, O. Heintz, R. Oltra, Inuence of the dissolution of MnS
inclusions under free corrosion and potentiostatic conditions on the
composition of passive lms and the electrochemical behaviour of stainless
steels, Electrochim. Acta 51 (2006) 32353243, http://dx.doi.org/10.1016/j.
electacta.2005.09.015.
[7] T. Shibata, Passivity breakdown and stress corrosion cracking of stainless
steel, Corros. Sci. 49 (2007) 2030, http://dx.doi.org/10.1016/j.corsci.2006.05.
031.
701
[8] G.T. Burstein, K. Sasaki, Effect of impact angle on the slurry erosioncorrosion
of 304L stainless steel, Wear 240 (2000) 8094, http://dx.doi.org/10.1016/
S0043-1648(00)00344-6.
[9] D. Rodriguez-Marek, D. Bahr, M. Pang, Mechanical measurements of passive
lm fracture on an austenitic stainless steel, Metall. Mater. Trans. A 34 (2003)
12911296.
[10] R.J.K. Wood, Tribocorrosion, in: T.J.A. Richardson (Ed.), Shreirs Corros.,
Elsevier, 2010, pp. 10051050.
[11] D. Landolt, S. Mischler, Tribocorrosion of passive metals and coatings, in: D.
Landolt, S. Mischler (Eds.), Tribocorrosion Passiv. Met. Coatings, Woodhead
publishing, Cambridge, UK, 2011, pp. xviixxi, http://dx.doi.org/10.1016/
B978-1-84569-966-6.50023-3.
[12] A. Berradja, D. Dforge, R.P. Nogueira, P. Ponthiaux, F. Wenger, J.-P. Celis, An
electrochemical noise study of tribocorrosion processes of AISI 304L in Cl
and SO4 2 media, J. Phys. D Appl. Phys. 39 (2006) 31843192, http://dx.doi.
org/10.1088/0022-3727/39/15/S08.
[13] S. Mischler, Triboelectrochemical techniques and interpretation methods in
tribocorrosion: a comparative evaluation, Tribol. Int. 41 (2008) 573583.
[14] K. Sasaki, G.T. Burstein, The generation of surface roughness during slurry
erosion-corrosion and its effect on the pitting potential, Corros. Sci. 38 (1996)
21112120.
[15] P. Ponthiaux, F. Wenger, D. Drees, J.P. Celis, Electrochemical techniques for
studying tribocorrosion processes, Wear 256 (2004) 459468, http://dx.doi.
org/10.1016/S0043-1648(03)00556-8.
[16] D. Dforge, F. Huet, R.P. Nogueira, P. Ponthiaux, F. Wenger, Electrochemical
noise analysis of tribocorrosion processes under steady-state friction regime,
Corrosion 62 (6) (2006) 514521.
[17] M.M. Stack, H. Jawan, M.T. Mathew, On the construction of micro-abrasion
maps for a steel/polymer couple in corrosive environments, Tribol. Int. 38
(2005) 848856.
[18] D. Sun, J.A. Wharton, R.J.K. Wood, Abrasive size and concentration effects on
the tribo-corrosion of cast CoCrMo alloy in simulated body uids, Tribol. Int.
42 (2009) 1595, http://dx.doi.org/10.1016/j.triboint.2009.03.018.
[19] M. Salasi, G.B. Stachowiak, G.W. Stachowiak, New experimental rig to
investigate abrasivecorrosive characteristics of metals in aqueous media,
Tribol. Lett. 40 (2010) 7184, http://dx.doi.org/10.1007/s11249-010-9640-2.
[20] D. Lpez, J.P. Congote, J.R. Cano, A. Toro, A.P. Tschiptschin, Effect of particle
velocity and impact angle on the corrosionerosion of AISI 304 and AISI 420
stainless steels, Wear 259 (2005) 118124, http://dx.doi.org/10.1016/j.wear.
2005.02.032.
[21] M.M. Stack, N. Corlett, S. Zhou, Impact angle effects on the transition
boundaries of the aqueous erosioncorrosion map, Wear 225229 (1999)
190198, http://dx.doi.org/10.1016/S0043-1648(99)00050-2.
[22] A.W. Hassel, A.J. Smith, Single particle impact experiments for studying
particle induced ow corrosion, Corros. Sci. 49 (2007) 231239, http://dx.doi.
org/10.1016/j.corsci.2006.05.021.
[23] K. Sasaki, G.T. Burstein, Observation of a threshold impact energy required to
cause passive lm rupture during slurry erosion of stainless steel, Philos. Mag.
Lett. 80 (2000) 489493.
[24] M.M. Stack, J.S. James, Q. Lu, Erosioncorrosion of chromium steel in a
rotating cylinder electrode system: some comments on particle size effects,
Wear 256 (2004) 557564.
[25] H. Meng, X. Hu, A. Neville, A systematic erosioncorrosion study of two
stainless steels in marine conditions via experimental design, Wear 263
(2007) 355362.
[26] S.S. Rajahram, T.J. Harvey, R.J.K. Wood, Evaluation of a semi-empirical model
in predicting erosioncorrosion, Wear 267 (2009) 18831893, http://dx.doi.
org/10.1016/j.wear.2009.03.002.
[27] C.G. Telfer, M.M. Stack, B.D. Jana, Particle concentration and size effects on the
erosion-corrosion of pure metals in aqueous slurries, Tribol. Int. 53 (2012)
3544, http://dx.doi.org/10.1016/j.triboint.2012.04.010.
[28] H. Arabnejad, S.A. Shirazi, B.S. McLaury, H.J. Subramani, L.D. Rhyne, The effect
of erodent particle hardness on the erosion of stainless steel, Wear 332333
(2015) 10981103, http://dx.doi.org/10.1016/j.wear.2015.01.017.
[29] A.V. Levy, Erosion and erosion-corrosion of metals, Corrosion 51 (1995)
872883, http://dx.doi.org/10.5006/1.3293564.
[30] D. Landolt, S. Mischler, M. Stemp, Electrochemical methods in tribocorrosion:
a critical appraisal, Electrochim. Acta 46 (2001) 39133929.
[31] M. Keddam, P. Ponthiaux, V. Vivier, Tribo-electrochemical impedance: a new
technique for mechanistic study in tribocorrosion, Electrochim. Acta 124
(2014) 38, http://dx.doi.org/10.1016/j.electacta.2013.08.186.
[32] P. Henry, J. Takadoum, P. Bercot, Tribocorrosion of 316L stainless steel and
TA6V4 alloy in H2SO4 media, Corros. Sci. 51 (2009) 13081314, http://dx.doi.
org/10.1016/j.corsci.2009.03.015.
[33] T. Adler, R. Walters, Corrosion and wear of 304 stainless steel using a scratch
test, Corros. Sci. 33 (1992) 18551876.
[34] J.O. Bello, R.J.K. Wood, J.A. Wharton, Synergistic effects of
micro-abrasioncorrosion of UNS S30403, S31603 and S32760 stainless
steels, Wear 263 (2007) 149159.
[35] A. Hayes, S. Shari, M.M. Stack, Micro-abrasion-corrosion maps of 316L
stainless steel in articial saliva, J. Bio- Tribo-Corrosion 1 (15) (2015), http://
dx.doi.org/10.1007/s40735-015-0015-y.
[36] M. Salasi, G. Stachowiak, G. Stachowiak, Tribo-electrochemical behaviour of
316L stainless steel: the effects of contact conguration, tangential speed, and
wear mechanism, Corros. Sci. 98 (2015) 2032, http://dx.doi.org/10.1016/j.
corsci.2015.05.009.
702 G.B. Stachowiak et al. / Corrosion Science 111 (2016) 690702
[37] M. Salasi, G.B. Stachowiak, G.W. Stachowiak, Three-body tribocorrosion of
high-chromium cast irons in neutral and alkaline environments, Wear 271
(2011) 13851396, http://dx.doi.org/10.1016/j.wear.2011.01.066.
[38] H. Sin, N. Saka, N.P. Suh, Abrasive wear mechanisms and the grit size effect,
Wear 55 (1979) 163190, http://dx.doi.org/10.1016/0043-1648(79)90188-1.
[39] D.V. De Pellegrin, G.W. Stachowiak, Assessing the role of particle shape and
scale in abrasion using sharpness analysis Part II. Technique evaluation, Wear
253 (2002) 10261034.
[40] M.G. Hamblin, G.W. Stachowiak, Description of abrasive particle shape and its
relation to two-body abrasive wear, Tribol. Trans. 39 (1996) 803810, http://
dx.doi.org/10.1080/10402009608983598.
[41] G.W. Stachowiak, Particle angularity and its relationship to abrasive and
erosive wear, Wear 241 (2000) 214219.
[42] K. Kato, K. Adachi, Wear mechanisms, in: Mod. Tribol. Handb., 2000, http://dx.
doi.org/10.1201/9780849377877.ch7.
[43] G.W. Stachowiak, A.W. Batchelor, Engineering Tribology, Elsevier, 2013,
http://dx.doi.org/10.1016/B978-0-12-397047-3.00018-7.
[44] K.K. Kasem, S. Jones, Platinum as a reference electrode in electrochemical
measurements, Platinum Met. Rev. 52 (2008) 100106, http://dx.doi.org/10.
1595/147106708X297855.
[45] Handbook of Reference Electrodes, in: G. Inzelt, A. Lewenstam, F. Scholz
(Eds.), Springer Berlin Heidelberg, Berlin, Heidelberg, 2013, http://dx.doi.org/
10.1007/978-3-642-36188-3.
[46] J.A. Bardwell, G.I. Sproule, D.F. Mitchell, B. MacDougall, M.J. Graham, Nature of
the passive lm on Fe-Cr alloys as studied by 18O secondary ion mass
spectrometry: reduction of the prior lm and stability to ex situ surface
analysis, J. Chem. Soc. Faraday Trans. 87 (7) (1991) 10111019, http://dx.doi.
org/10.1039/ft9918701011.
[47] M. Stemp, S. Mischler, D. Landolt, The effect of contact conguration on the
tribocorrosion of stainless steel in reciprocating sliding under potentiostatic
control, Corros. Sci. 45 (2003) 625640.
[48] M. Stemp, S. Mischler, D. Landolt, The effect of mechanical and
electrochemical parameters on the tribocorrosion rate of stainless steel in
sulphuric acid, Wear 255 (2003) 466475, http://dx.doi.org/10.1016/S0043-
1648(03)00085-1.
[49] P. Jemmely, S. Mischler, D. Landolt, Electrochemical modeling of passivation
phenomena in tribocorrosion, Wear 237 (2000) 6376.
[50] S.S. Rajahram, T.J. Harvey, J.C. Walker, S.C. Wang, R.J.K. Wood, Investigation of
erosioncorrosion mechanisms of UNS S31603 using FIB and TEM, Tribol. Int.
46 (2012) 161173, http://dx.doi.org/10.1016/j.triboint.2011.05.012.
[51] S.S. Rajahram, T.J. Harvey, J.C. Walker, S.C. Wang, R.J.K. Wood, G. Lalev, A study
on the evolution of surface and subsurface wear of UNS S31603 during
erosioncorrosion, Wear 271 (2011) 13021313, http://dx.doi.org/10.1016/j.
wear.2010.11.018.
[52] J. Perret, E. Boehm-Courjault, M. Cantoni, S. Mischler, A. Beaudouin, W. Chitty,
J.-P. Verno, EBSD, SEM and FIB characterisation of subsurface deformation
during tribocorrosion of stainless steel in sulphuric acid, Wear 269 (2010)
383393, http://dx.doi.org/10.1016/j.wear.2010.04.023.
[53] C. Pan, L. Liu, Y. Li, F. Wang, Pitting corrosion of 304ss nanocrystalline thin
lm, Corros. Sci. 73 (2013) 3243, http://dx.doi.org/10.1016/j.corsci.2013.03.
022.
[54] R.K. Gupta, N. Birbilis, The inuence of nanocrystalline structure and
processing route on corrosion of stainless steel: a review, Corros. Sci. 92
(2015) 115, http://dx.doi.org/10.1016/j.corsci.2014.11.041.
[55] R.K. Gupta, R.K. Singh Raman, C.C. Koch, Electrochemical characteristics of
nano and microcrystalline FeCr alloys, J. Mater. Sci. 47 (2012) 61186124,
http://dx.doi.org/10.1007/s10853-012-6529-5.
[56] X. Wang, D. Li, Mechanical and electrochemical behavior of nanocrystalline
surface of 304 stainless steel, Electrochim. Acta 47 (2002) 39393947, http://
dx.doi.org/10.1016/S0013-4686(02)00365-1.