Tribology International 35 (2002) 669–679

www.elsevier.com/locate/triboint

Modelling the tribo-corrosion interaction in aqueous sliding

conditions
J. Jiang a,∗, M.M. Stack b, A. Neville c
a
Integrated Manufacturing Technologies Institute, National Research Council Canada, 800 Collip Circle, N6G 3A8 London, ON, Canada
b
Department of Mechanical Engineering, University of Strathclyde, James Weir Building, 75 Montrose St., Glasgow G1 1XJ, UK
c
Department of Mechanical and Chemical Engineering, Heriot-Watt University, Edinburgh EH14 4AS, Scotland, UK

Received 6 August 2001; received in revised form 10 October 2001; accepted 22 October 2001

Abstract

In corrosion wear, one of the most commonly observed phenomena is the synergism where both corrosion and wear are signifi-
cantly increased by the tribo-corrosion interactions, leading to much greater material losses than produced by the sum of losses by
either process alone. However, mechanisms for the synergistic effect are generally not well understood and there has been little
effort in modelling synergistic processes. In this paper, mechanisms for the interactions between corrosion and sliding wear of
metals are proposed and a mathematical model is developed which incorporates the various factors affecting corrosion wear, includ-
ing experimental and environmental conditions and material properties. Many of the observed phenomena in corrosion wear can
be explained. It also provides a good basis for wear map construction and would be especially useful in carrying out dimensional
analyses for constructing such maps.  2002 Elsevier Science Ltd. All rights reserved.

Keywords: Corrosion wear; Synergistic effect; Tribo-corrosion interactions

1. Introduction systems both directly and indirectly in terms of material

loss and associated equipment downtime for repair and
In engineering situations, wear can occur with sig- replacement of worn and corroded components.
nificant chemical or electrochemical reactions with the One of the most common features in corrosion wear
environment. The wear can be in the forms of erosion, is the synergistic effect, having been widely observed in
abrasion, sliding, or fretting. Such wear is often referred abrasion [3–7], erosion [8–17] and sliding wear [18–27].
to as corrosion wear [1]. Typical examples include wear The total material loss in the presence of corrosive media
of mechanical components working in mining and min- is significantly higher than that in the absence of
erals processing, in the marine atmosphere, in paper and corrosion. The total corrosion wear loss, Vcw, is also
pulping operations, in food handling and in fossil burn- higher than the simple superposition of material loss due
ing systems where corrosive products are generated. to pure corrosion, Vc, without the influence of wear and
Under some other conditions where the environments are the material loss due to wear, Vw, measured in the
seemingly non-deleterious, corrosion wear has also been absence of corrosion, i.e.
found to cause serious problems. For example, the gen-
Vcw ⫽ Vw ⫹ Vc ⫹ ⌬Vs (1)
eration of corrosive decomposition products in lubricants
due to synthesised additives such as sulphuric acid has This effect, ⌬Vs, is usually significantly greater than
been shown to accelerate wear [2]. The damage to zero and can be attributed to the combined effect of
materials due to the coexistence of wear and corrosion corrosion-induced wear, ⌬Vw, and wear-induced cor-
has a significant impact on economics of engineering rosion, ⌬Vc:
⌬Vs ⫽ ⌬Vw ⫹ ⌬Vc (2)
∗
Corresponding author. Fax: +1-519-430-7064. In spite of the many studies that have been carried out
E-mail address: jiaren.jiang@nrc.ca. (J. Jiang). to characterise the corrosion wear behaviour of various

0301-679X/02/$ - see front matter  2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 3 0 1 - 6 7 9 X ( 0 2 ) 0 0 0 5 8 - 0
670 J. Jiang et al. / Tribology International 35 (2002) 669–679

Nomenclature
a length of a micro-crack
a0 average size of potential micro-crack initiation sites on a wear surface
a0c size of a potential micro-crack initiation site in the presence of corrosive environment
ac critical size above which a micro-crack can propagate to generate a wear debris particle
⌬ac increase in the average size of potential micro-crack initiation sites due to the presence of a corrosion
environment
ar average size of potential crack initiation sites in the absence of corrosion in a reference environment
Aa apparent area of contact
Aatom atomic weight of the dissolving metal during electrochemical corrosion
Ar real area of contact
D average diameter of wear debris particles
f contact frequency between encountering surfaces
F the Faraday’s electrochemical equivalent
gcfs ⬅ (da / dn)cfs maximum attainable crack growth rate under given corrosion fatigue conditions
gr ⬅ (da / dn)r average crack growth rate during the generation of a wear debris particle in the reference
environment
gs ⬅ (da / dn)s crack growth rate in the corrosive environment under sustained stresses
H hardness of the metal
i transient repassivation current density on fresh metal surface
i0 maximum current density on fresh metal surface at the start of repassivation
ia average anodic current density passing through the apparent area of the specimen excluding the real
area of contact
ir average anodic current density passing through the real area of contact
is anodic current density at the steady state of corrosion on the passivated metal surface
Ia,w anodic corrosion current under the influence of wear
k⬘a,ka,kcw proportionality constants
Kw proportionality factor
l length of wear track
m empirical constant used to describe growth dynamics of potential micro-crack initiation sites in the
presence of corrosion
n number of contact cycles at a given crack site
Ñ number of debris particles generated in unit time on a unit apparent area of the wear surface
N0 number of potential sites for the initiation of micro-cracks for the generation of wear debris particles
on a unit apparent area of contact
Nc number of critical micro-cracks on a unit area of the wear surface that are ready to propagate
⌬Nc increment of the number of critical micro-cracks due to the presence of corrosion
Nr number of critical micro-cracks that are ready to propagate to generate wear debris in a reference
environment
s total sliding distance
t total sliding time
tc critical time for generating a wear debris particle with an average size, D, from an existing micro-
crack nucleus
v sliding velocity
Vc material loss due to ‘pure’ corrosion without the effect of mechanical wear
Vcc material loss due to corrosion during corrosion wear
Vcw total material loss during corrosion wear due to both corrosion and wear
Vw pure wear loss in otherwise the same environment as in the corrosion wear process but without
corrosion
 J. Jiang et al. / Tribology International 35 (2002) 669–679 671

Vwc material loss due to wear during corrosion wear

⌬Vc wear-induced corrosion
⌬Vs synergistic effect
⌬Vw corrosion-induced wear
W normal load
ze average discharge valence of the metal for anodic reactions during corrosion
l temporary integration variable
r density of the metal
sa standard deviation for the size distribution of potential micro-crack initiation sites on a wear surface
t instant time
t0 characteristic passivation time of a metal during repassivation from a fresh surface
ti characteristic reaction time for the reaction controlling a corrosion fatigue process

materials under various conditions, the current under- Generally speaking, the total material loss during cor-
standing of the corrosion wear process is still very poor. rosion wear, Vcw, is the sum of material losses due to
Mischler et al. [28] presented a model to describe the wear, Vwc, and due to corrosion, Vcc, respectively. This
corrosive behaviour of passive metals sliding against a can be represented by reorganising Eqs. (1) and (2) into
hard insulating body. The basic assumptions in the the following form:
model included that (a) the real area of contact is determ-
ined by load and the hardness of the metal in the same Vcw ⫽ Vw ⫹ Vc ⫹ ⌬Vw ⫹ ⌬Vc ⫽ Vwc ⫹ Vcc (4a)
way as in deriving the Archard’s adhesive wear theory where,
[29,30]; (b) the passive film at the real area of contact
is damaged at each cycle of sliding to expose a fresh Vcc ⫽ Vc ⫹ ⌬Vc (4b)
metal surface which is then rapidly depassivated; and (c)
corrosion outside the real area of contact is negligible. and
The wear component of material loss was not modelled. Vwc ⫽ Vw ⫹ ⌬Vw (4c)
An expression was derived for the relationship between
anodic corrosion current under the influence of wear,
Ia,w, and experimental parameters such as load, W, slid- 2.1. Material loss due to corrosion in the presence of
ing frequency, f, wear track length, l, and hardness of wear, Vcc
the metal, H:
1

冕
/f The material loss due to corrosion, Vcc, during cor-
Ia,w ⫽ Kwlf(W / H) 1/2
idt (3) rosion wear is higher than the pure corrosion in the
absence of wear, Vc, because the metal surface is acti-
0
vated by the wear actions (either by hard particle intrud-
The proportionality factor, Kw, cannot be theoretically
ing or by sliding) by: (a) removing [14,15,17,28,31] or
determined. By choosing a fitted value for Kw, good
damaging [31] the protective passive films on the sur-
agreement was shown between experiments and the
face; and (b) increasing internal energy of the metal sur-
theoretical predictions [28].
face due to the plastic, normally very severe defor-
If the performance of materials in corrosion wear sys-
mation. From the electrochemical point of view, such
tems is to be improved, it is important to identify and
actions will lead to a negative shift of the corrosion
understand the regimes and mechanisms of the corrosion
potential, positive shift of critical and steady passivation
wear process. In this paper, a model is presented for slid-
potentials and increased current densities of anodic dis-
ing wear of metals in corrosive aqueous solutions, which
solution and passivation current, as indeed have been
considers both the effects of corrosion on wear and wear
widely observed [19,21,26,27,31–33].
on corrosion. Factors affecting corrosion wear are dis-
The total material loss due to corrosion, Vcc, is related
cussed on the basis of the model and issues that require
to the anodic corrosion current under the influence of
attention in the study of corrosion wear are pointed out.
wear, Ia,w, by the Faraday law and can be calculated
according to the following equation
2. Modelling the corrosion wear—basic processes
and expressions Vcc ⫽ Vc ⫹ ⌬Vc ⫽ Ia,wtAatom / (zeFr) (5a)
According to the evidences available from the litera- where,
ture, the following model for sliding corrosion wear can
be envisaged. Ia,w ⫽ Arir ⫹ (Aa⫺Ar)ia (5b)
672 J. Jiang et al. / Tribology International 35 (2002) 669–679

2.2. Material loss due to wear in the presence of ber of potential micro-crack initiation sites, Nc. During
corrosion, Vwc a wear process, the wear surface is, on a microscopic
scale, always loaded and deformed unevenly. Damage
The wear volume of a specimen is equal to the product to any surface films that may exist will also be non-
of the average volume of a single particle and the total uniform. Thus, localised corrosion attack of the wear
number of wear debris particles generated within a slid- surface will occur when a corrosive solution is present.
ing time, t, which can be expressed by This is similar to situations in corrosion fatigue and in
stress corrosion cracking where such localised attack can

冕
t
πD3 promote the initiation of short cracks. As a result, the
V⫽ ÑAadt (6a) density of active crack initiation sites, Nc, on a unit
6
0 apparent area of the wear surface will be different from,
and is usually higher than, that in a reference environ-
In steady state sliding, the integration of Eq. (6a) gives
ment where corrosion is prohibited. Second, the propa-
πD3 πD3s gation of the micro-crack during the formation of wear
V⫽ ÑAat ⫽ ÑA (6b)
6 6 v a debris particle is via the debonding at the crack tip (Fig.
1). If a reactive species, A, is present at the tip, and can
Thus, anything affecting either of the two factors, D and act to weaken the bonding, it will promote the propa-
Ñ, will result in a change in wear rate of the material. gation of the crack (increasing da / dn) [35,36], accelerat-
There can be many detailed mechanisms for the gener- ing the generation of wear debris particles (by reducing
ation of a wear debris particle. However, it can be gener- tc). The reaction between the reactive species and the
ally described as a crack initiation and propagation pro- bonds at the crack tip is not necessarily a chemical
cess, resembling micro-fatigue (either high cycle or low
one [37].
cycle). Fatigue is a very broad concept. For example,
removal of material by one-pass contact (such as cutting
or adhesion–breaking) can be regarded as a one-cycle
fatigue failure. This low cycle fatigue concept has been
successfully applied to analyse material removal frac-
tions during single particle scratching tests under various
particle tip penetration depths [34]. According to this
analysis, it was shown that the volume of material that
is actually removed as a fraction of the groove volume
formed in a single scratch is closely related to the low
cycle fatigue behaviour of the material. Thus, it is
assumed here that during sliding wear, a micro-crack is
initiated on the wear surface and will propagate for a
certain distance after each contact with an asperity from
the counter surface. A wear debris particle is generated
after a certain number of cycles of contact at the same
spot with the encountering surface.
The critical time, tc, for generating a wear debris par-
ticle with an average size, D, from an existing micro-
crack nucleus is related to the micro-crack growth rate,
da / dn, by

冕 冕
tc tc

da da
D⫽ dt ⫽ fdt (7)
dt dn
0 0

If Nc active micro-crack sites are present on a unit

area of the wear surface that are ready to propagate, then
the generation rate, Ñ, of wear debris particles can be
calculated by
Fig. 1. Schematic diagram showing the micro-crack propagation
Ñ ⫽ Nc / tc (8) mechanism during the generation of a wear debris particle by chemi-
cally activated reactions between reactive species, A, and the bonds,
The presence of a corrosive solution can accelerate B, at the crack tip.
wear in two ways. First, corrosion can increase the num-
 J. Jiang et al. / Tribology International 35 (2002) 669–679 673

3. Modelling the corrosion wear—mathematical contact is N0 and the size distribution of such sites fol-
representations lows the normal (Gaussian) statistical distribution with
an average size of a0

冋 册
On the basis of the model described in the last section,
formal mathematical representations for corrosion wear 1 (a⫺a0)2
fN(a) ⫽ exp ⫺
冑2πs
(12)
will be developed as follows. a
2s2a

3.1. Corrosion component in corrosion wear Under given loading conditions, only those cracks
which are larger than some critical size can propagate.
In Eq. (5b), the first term is equivalent to Mischler et Suppose this critical size is equal to ac, then the number
al.’s model in Eq. (3), i.e. of micro-cracks, Nc, that can propagate to generate wear
debris particles is obtained by
1

冕
/f

冕
⬁
Arir ⫽ Kwlf(W / H) 1/2
idt (9) N0 ac⫺a0
Nc ⫽ N0 fN(a)da ⫽ erfc( ) (13)
0
ac
2 冑2sa
According to experimental observations [38], the
damping of the transient current, i, as a function of time, where, erfc(z) is the second type error function.
t, after the metal surface is depassivated can be approxi- Congleton et al. [39] experimentally measured the
mated by growth of micro-cracks on 50D low carbon steel surface
(in 3.5% NaCl) under both stress-free and constant-load
i ⫽ (i0⫺is)exp(⫺t / t0) ⫹ is (10) conditions. The variations of maximum pit depth, hp, as
The current density, is, at the steady state of corrosion a function of immersion time, t, followed the following
is the same as ia in Eq. (5b), i.e. is ⫽ ia. The character- relationships: (a) for the stress-free test, hp ⫽ 8.68t0.62;
istic passivation time, t0, reflects how easily the metal and (b) for the constantly-loaded (400 MPa) test, hp ⫽
surface passivates. The lower the value of t0, the easier 15.25t0.80.
the metal tends to passivate. For non-passive materials, The presence of stresses significantly increased the
i0 ⫽ is, and the anodic dissolution current, i ⫽ is, is con- growth rate of pits.
stant. In the case of wear, the size scale of micro-cracks is
By substituting Eq. (10) into Eq. (9) and applying the much smaller and it is almost impossible to make any
integration result to Eqs. (5a) and (5b), Vcc is obtained as direct observation. However, as a first-order estimation,
it can be assumed that the growth of the existing active
Vcc ⫽ Vc ⫹ ⌬Vc ⫽ Ia,wtAatom / (zeFr) sites, a0c, also follows a power law as a function of
⫽ [Aatoms / (zeFrv)]{Aais ⫹ Kal(W / H)1/2(i0 (11a) time, t,
⫺is)ft0[1⫺exp(⫺1 / ft0)]} a0c(t) ⫽ ar ⫹ k’atm ⫽ ar ⫹ kaf⫺m ⫽ ar ⫹ ⌬ac (14a)
where, where, the increase in the average size, ⌬ac, of potential
Vc ⫽ AaisAatoms / (zeFrv) (11b) micro-crack initiation sites due to the presence of cor-
rosion environment is equal to
and
⌬ac ⫽ kaf⫺m (14b)
⌬Vc ⫽ [Aatoms / (zeFrv)]Kal(W / H) (i0⫺is)ft0[1
1/2
(11c)
The number of micro-cracks, Nc, larger than the criti-
⫺exp(⫺1 / ft0)] cal crack size, ac, under corrosion wear conditions can be
obtained by replacing a0 in Eq. (13) with a0c in eq. (14a):
3.2. Wear component in corrosion wear
N0 ac⫺⌬ac⫺ar N0 ac⫺ar
Nc ⫽ erfc( ) ⫽ erfc(
冑2sa 冑2sa
)
The two major effects of corrosion on wear loss— 2 2
increased number of micro-crack initiation sites and
(ac⫺ar) / 冑2sa
accelerated micro-crack propagation—will be discussed
separately in this section.
⫹
N0
2 冕 exp(⫺l2)dl ⫽ Nr (15a)
(ac⫺⌬a⫺ar) / 冑2sa
3.2.1. Corrosion-induced micro-crack initiation
On a wear surface, it can be envisaged that there are
many sites that are potential micro-crack initiation pos- ⫹ ⌬Nc
itions for the generation of wear debris particles. Sup-
pose the number of such sites on a unit apparent area of where,
674 J. Jiang et al. / Tribology International 35 (2002) 669–679

Nr ⬅
N0
2
erfc冉 冊
ac⫺ar
冑2sa
(15b)
ti is a characteristic reaction time for the chemical reac-
tion controlling the corrosion fatigue process.
Although the above equation was derived by assuming
and a particular reaction step, it can be expected that in other
environments and for other metals, some other con-
(ac⫺ar) / 冑2sa trolling step of chemical or electrochemical reactions

⌬Nc ⬅
N0
2 冕 exp(⫺l2)dl (15c)
causing weakening of materials at the crack tip will
occur and ti will be the characteristic time for the corre-
(ac⫺⌬a⫺ar) / 冑2sa
sponding reaction step. The same concept can be applied
to the case of corrosion wear.
represent respectively the number of critical micro- By substituting (16) into Eq. (7) and applying (17),
cracks in a reference environment and the increment of the time, tc, to generate a wear debris particle with a
this number due to corrosion. diameter D can be obtained from

冕 冕
ftc ftc

3.2.2. Corrosion-induced micro-crack propagation

According to studies on corrosion fatigue [40–42], the D ⫽ (da / dn)rdn ⫹ (da / dn)cfdn
rate of crack growth under a given amplitude of stress 0 0
intensity in a deleterious environment is composed of

冕
ftc
three components:
⫹ (da / dn)sdn ⫽ {gr ⫹ gcfs[1⫺exp(⫺1 / fti)] (18a)
(da / dn)e ⫽ (da / dn)r ⫹ (da / dn)cf ⫹ (da / dn)s (16)
0

where, (da / dn)e represents the crack growth rate in the ⫹ gs}ftc
presence of deleterious environment, (da / dn)r is the rate
of crack growth in a reference environment where the i.e.
‘pure’ mechanical fatigue occurs, (da / dn)cf represents
the crack growth rate due to the interaction of cyclic tc ⫽ D / {gr ⫹ gcfs[1⫺exp(⫺1 / fti)] ⫹ gs}f (18b)
mechanical loading (fatigue) and attack of environment,
and (da / dn)s is the contribution by sustained-load crack
3.2.3. Wear loss in corrosion wear
growth (that is by ‘stress corrosion cracking’) at stress
By substituting Eqs. (15a) and (18b) into (8), the total
intensity levels above KISCC [43].
generation rate of wear debris particles, Ñ, during cor-
Under corrosion wear conditions, the process of wear
rosion wear is obtained
debris particle generation is analogous to a corrosion
fatigue process except that the degree of deformation Ñ ⫽ (Nr ⫹ ⌬Nc){gr ⫹ gcfs[1⫺exp(⫺1 / fti)] (19)
under wear conditions may be higher than that under a
normal corrosion fatigue condition. Therefore, (16) can ⫹ gs}fD ⫺1

be used to describe the micro-crack propagation during

corrosion wear processes. (da / dn)s essentially represents The total material loss, Vwc, due to wear in the pres-
a steady micro-crack growth due to residual stresses at ence of corrosion is thus obtained by substituting (19)
the crack tip when the region is not in contact with the into Eq. (6b),
opposite rubbing surface; it is therefore dependent of πD2Aas
time and surface deformation/load. It is also temperature Vwc ⫽ [Nrgr ⫹ ⌬Ncgr ⫹ (Nr (20a)
dependent [43] as corrosion reactions are involved in 6l
the process. ⫹ ⌬Nc){gcfs[1⫺exp(⫺1 / fti)] ⫹ gs}]
For the term (da / dn)cf in (16), Wei and Shim [42]
derived an expression for corrosion fatigue of steels in where, the first term of the right hand side of Eq.(20a)
water by assuming that the chemical reaction step of
water molecules, with the steel surface to generate πD2Aas
Vw ⫽ Nrgr (20b)
hydrogen was the control step in promoting embrittle- 6l
ment at the crack tip:
is equal to the pure wear in a reference environment and
(da / dn)cf ⫽ gcfs[1⫺exp(⫺1 / fti)] (17) the rest is the corrosion-induced wear,
First-order reaction kinetics were assumed in deriving πD2Aas
this equation. The term gcfs ⬅ (da / dn)cfs denotes a ⌬Vw ⫽ [⌬Ncgr ⫹ (Nr ⫹ ⌬Nc){gcfs[1 (20c)
6l
maximum attainable crack growth rate under given con-
ditions and is independent of temperature and frequency. ⫺exp(⫺1 / fti)] ⫹ gs}]
 J. Jiang et al. / Tribology International 35 (2002) 669–679 675

3.3. The total corrosion wear and the synergism (a) reduced time for the reaction at or near the crack tip
that weakens the bonding and hence less corrosion
Finally, substituting Eqs. (11a) and (20a) into (4a), fatigue crack growth in each cycle; and (b) decreased
the following expression for the total material loss in number of corrosion-induced micro-crack initiation sites
corrosion wear, Vcw, is obtained: (Eqs. (14b), (15b) and (15c)), ⌬Nc. Simply stating, the
time for interaction between corrosion and wear is
πD2Aas
Vcw ⫽ Nrgr ⫹ AaisAatoms / (zeFrv) reduced. On the other hand, increase in the contact fre-
6l quency, f, leads to more frequent depassivation of the
πD2Aas metal surface, resulting in more anodic dissolution of
⫹ [⌬Nc{gr ⫹ gcfs[1⫺exp(⫺1 / fti)] ⫹ gs} metal. Therefore, with increase in contact frequency, the
6l
increase in mass loss due to wear-induced corrosion
⫹ Ncr{gcfs[1⫺exp(⫺1 / fti)] ⫹ gs}] (21) becomes more important while the contribution due to
⫹ [Aatoms / (zeFrv)]Kal(W / H)1/2(i0⫺is)ft0[1⫺exp( corrosion-induced wear decreases.
Iwabuchi and his group found that the enhancement
⫺1 / ft0)] ⬅ Vw ⫹ Vc ⫹ ⌬Vw ⫹ ⌬Vc of mass loss for sliding wear of stainless steel [44,45]
and SKD61 die steel [33] against an alumina ball in sea
water and in Na2SO4 solution was mainly due to wear-
4. Discussion induced corrosion. However, others [19–23] have
observed that corrosion-induced wear for various stain-
In deriving (21), the concept of micro-cracking and less steels was more dominant. The major difference in
fatigue has been invoked for the generation of wear experimental conditions in these cases was that Iwabuchi
debris particles. Given the complexity of the interactions et al. [33,44,45] used a ball-on-disk reciprocating wear
between corrosion and cracking/fatigue and the fact that rig (6.4 mm alumina ball, 10 N, 1.5 mm amplitude recip-
these are still in the formative stages of understanding, rocating at 8.3 Hz) while the Li’s group [19–23] used
data for several of the parameters in (21), such as the a pin-on-ring continuous sliding wear machine (6 mm
crack growth rates under the various conditions, gr, gcfs, alumina ball, 38 mm diameter ring, 0.025–0.19 m s-1, 5–
gs, are difficult to estimate either from the literature or 35 N). The time interval between two successive contact
from experiments. It is therefore difficult at this stage to between the ball and the given point on the specimen
make a generalised quantitative assessment of the model. surface was 60.2 ms in Iwabuchi et al.’s experiments
Some issues, such as the regeneration and removal of which is much shorter than the repassivation time, t0, of
oxides or compact layers that may develop and may the alloys (typically approximately 200 ms [33]), while
affect the wear processes, are not considered in this this time interval for the Li group’s rig was 628–4773
model. Nevertheless, the model covers most of the ms which is far beyond the typical repassivation time of
important issues involved in corrosion wear and provides typical 200 ms for most of the materials studied. Thus,
a useful basis for further development. Some qualitative the metal surface was always at the active dissolution
or semi-quantitative discussion based on (21) would be state during sliding in Iwabuchi et al.’s tests, leading to
instructive and would offer some insights into the pro- a large contribution of corrosion to the total corrosion
cess of corrosion wear. wear loss. On the other hand, the active dissolution and
Although (21) has been derived for a sliding wear sys- the repassivation process in the pin-on-ring rig used by
tem, similar expressions would be obtained for erosion Li’s group took only a small fraction of the total sliding
and abrasion with corresponding substitutions for the time and processes damaging the near- and sub-surface
sliding distance, s, and the contact/interaction frequency, layer of the material such as corrosion, stress corrosion,
f, between the encountering bodies. Mischler et al.’s [28] corrosion fatigue, hydrogen embrittlement and other pro-
analysis on the wear-induced corrosion (eq. (3)) is also cesses had enough time to take place. These processes
applicable in erosion and abrasion. Broadly speaking, the can greatly accelerate the initiation and propagation of
contact frequency, f, should be interpreted as the fre- micro-cracks in the generation of wear debris, signifi-
quency at which a given micro-cracking site is loaded cantly increasing the synergistic effect due to corrosion-
by the counter body which is an asperity in sliding and induced wear.
is a hard particle in abrasion and erosion. Yahagi and Mizutani [46] studied the effect of off-
time (the time interval at which the load was not applied)
4.1. The effect of contact frequency, f, on the relative during a cyclic loading (for 60 s)—unloading test on
importance of wear-induced corrosion and corrosion- corrosion wear of grey cast iron in sulphuric acid (65%
induced wear H2SO4) on a pin-on-disk (continuous sliding) wear rig.
In their experiments, it was found that the corrosion wear
In (21), the contact frequency, f, appears in several rate of the iron initially increased considerably with
terms. For corrosion-induced wear, increase in f means increase in the off-time and, after reaching a maximum
676 J. Jiang et al. / Tribology International 35 (2002) 669–679

wear rate at an off-time of approximately 10 s, decreased 4.2. The effect of susceptibility of the metal to
with further increase in the off-time above 100 s to attain corrosion
a wear rate similar to the rate in a continuous sliding
test. This phenomenon was explained by the authors as From Eqs. (15a) and (15b), factors that can cause
a result of the development and removal of oxide films increase in localised corrosion will promote the initiation
on the specimen surfaces. At short off-times, the oxide of micro-cracks for the generation of wear debris par-
films became thicker and more fragile with increase in ticles, ⌬Nc, and hence significantly increase the cor-
off-time, which were easily removed by subsequent slid- rosion-induced wear (Eq. (20c)). Chloride ions are well
ing action. With further increase in the off-time, the known to have such effects in corrosion fatigue and
thickness of the oxide films increased to such a thickness stress corrosion cracking. Therefore, their presence con-
that could not be removed by the sliding action and acted siderably accelerates corrosion wear, especially enhances
as a protective layer, reducing the wear rate at longer the corrosion-induced wear component, as has been
off-times. However, according to the experimental con- observed in many experiments [11,15,21,27,32,50].
ditions used (wear track diameter 34 mm, 3 mm diameter As in stress corrosion cracking (SCC), the contri-
pin, 0.13 m s⫺1), the time interval for the pin to make bution of crack growth due to sustained-load (gs)
contact with a given point on the disk specimen surface depends strongly on the combination of material and
was 819.6 ms, which was already sufficient for the environment. Under suitable conditions, corrosion-
development of a passive film on the specimen surface. induced wear can be expected to increase by the SCC
Also, the maximum wear rate corresponding to the off- mechanism due to the sustained-load on the wear sur-
time of 10 s was 0.08 mg m⫺1 which was more than face. In a slurry erosion study on a low carbon steel,
double the wear rate, 0.03 mg m⫺1, at the non-stop slid- Zhou et al. [51] measured the variation of erosion wear
ing. If it is assumed that wear occurred uniformly over and erosion-corrosion synergism as a function of polaris-
ation potential in a NaCO3 solution. In their experimental
the whole apparent area of contact, then this wear rate
set-up, the interaction frequency between the eroding
increase, 0.05 mg m-1, requires the formation and
particles and a given cracking site was high, constantly
removal of an extra oxide layer of 138 nm per revol-
depassivating the steel surface. As a result, it was
ution. The total thickness of the oxide layer would be
observed that wear-induced corrosion was the main
220.5 nm per revolution for a wear rate of 0.08 mg m-
1
source of synergism under most of the conditions. How-
. If it is considered that actual wear normally occurs at ever, considerable corrosion-induced wear was observed
the real area of contact which is only a very small frac- in the active to passive transition potential region. SCC
tion of the apparent area of contact, the total thickness is well known to occur under these conditions. It can
of the oxide layer to be formed and removed in each thus be reasonably suggested that sustained stresses
revolution would be much thicker than the before esti- played a dominant role in accelerating micro-crack
mated values. Such a thick oxide film is obviously propagation (gs) and erosion.
impossible to develop. Typical thickness of passive films
on iron [47] and stainless steels [48,49] is generally far 4.3. The relationship between synergism and pure
below 10 nm. Apparently, according to the present wear rate
model, during the unloading period (the off-time), some
activation process promoting the initiation and propa- In summarising experimental data published in the
gation of micro-cracks within the highly strained wear literature on erosion wear, mainly on cavitation erosion,
surface might have taken place, which would greatly in a wide range of conditions, Wood and Hutton [52]
facilitate the generation of more and larger wear debris found that except for one or two data points, the data
particles in the following sliding action, enhancing the fell well into two groups, the medium and the high syn-
wear loss. On the other hand, during the prolonged off- ergistic systems. The erosion-corrosion enhancement
time immersion in the electrolyte, dissolution of the wear ⌬Vs, was found to be closely related to the pure erosion
surface was able to blunt the tips of micro-cracks and component, Vw, as expressed by the following
reduce the number of active micro-cracks within the equations:(a) for the medium synergistic effect group,
wear surface, decreasing the wear damage during the fol-
lowing sliding. Eventually, after long enough off-time, ⌬Vs / Vc ⫽ exp[1.277ln(Vw / Vc)⫺1.9125] (22a)
the status of the wear surface became similar to that in ⫽ 0.1477(Vw / Vc) 1.277

the continuous sliding, leading to a similar total wear

loss to the continuous sliding because the corrosion only and (b) for the high synergism group,
contributed less than 3.5% to the total material loss even ⌬Vs / Vc ⫽ exp[0.755ln(Vw / Vc) ⫹ 1.222] (22b)
in the prolonged off-time immersion.
⫽ 3.3940(Vw / Vc) 0.755
 J. Jiang et al. / Tribology International 35 (2002) 669–679 677

In the above equations, Vc represents material loss due the present model to be proportional to W⫺1/2, very close
to corrosion in the absence of erosion. However, the to the observed value of ⫺0.458 for the power index.
rationale behind such relationships remains to be under-
stood. 4.5. Wear map construction
According to the model presented in this paper, (21),
if it is assumed that the presence of corrosive agent In the past decade or so, significant progress has been
serves mainly to accelerate the initiation of critical made in the development of methodologies for the con-
micro-cracks, i.e. to increase ⌬Nc, then the main syner- struction of wear maps under various conditions [53–
gistic effect due to corrosion-induced wear, ⌬Vw, is 64]. In a wear map, dominant wear mechanisms are
approximated by mapped over a parameter space. Such an approach pro-
πD2Aas vides a powerful tool for analysis of mechanisms of
⌬Vw ⫽ [⌬Nc{gr ⫹ gcfs[1⫺exp(⫺1 / fti)] material damage, optimisation of process parameters and
6l
selection of materials. Recently, wear maps for slurry
⫹ gs} ⫹ Ncr{gcfs[1⫺exp(⫺1 / fti)] (23) erosion-corrosion of metals showing regime transitions
πD2Aas have been developed as a function of parameter groups
⫹ gs}]⬇ ⌬Ncgr ⫽ (⌬Nc / Nr)Vw obtained from dimensional analysis [65]. This method-
6l ology is very useful because more parameters can be
and incorporated in one map and it is possible to make gener-
alised applications of the map in physically similar situ-
⌬Vw / Vc⬀Vw / Vc (24) ations. The dimensionless parameter groups used in the
The synergism is thus related to the pure wear, Vw, in analysis of Stack et al. [65] were derived based on sev-
the reference environment with a power index of 1 which eral discrete basic models/expressions related to erosion
lies between the power indices of 0.755 and 1.277 from and corrosion. Similar work has not been available in
the empirical expressions in Eqs. (22a) and (22b). other types of corrosion wear processes.
Based on the present model, the dependence of syner- Simplified wear maps can also be constructed. As an
gism on the pure erosion described by Eqs. (22a) and example, in order to assess the relative importance of
(22b) can be interpreted as follows. In cavitation erosion, corrosion-induced wear, ⌬Vw, and wear-induced cor-
the micro-crack initiation stage is important in the wear rosion, ⌬Vc, in causing the synergistic effect of corrosion
process because the specimen surface is relatively wear, the ratio, ⌬Vw / ⌬Vc, can be used as a criterion. If
smooth compared with that in sliding wear and abrasion. this ratio is less than some critical value, the material
At the same time, the interaction frequency, f, between loss can be said to be dominated by wear-induced cor-
the collapsing cavities and the wear surface is expected rosion. Otherwise, the process is dominated by cor-
to be high, making the contribution of sustained-load rosion-induced wear. The expression for ⌬Vw in (20c) is
crack growth and the cyclic fatigue growth of micro- complicated and contains parameters for crack growth
cracks unimportant. As a result, the corrosion-promoted rates which are closely related to the load conditions and
crack initiation dominates the corrosion-induced wear to mechanical properties of the material but which can-
and the synergism. not be explicitly expressed. If it is assumed that ⌬Vw is
proportional to load, W, and is inversely proportional to
4.4. The effect of load on synergism the material hardness, H, as has been commonly
observed in many of the corrosion wear studies
[5,19,22,23,32], and the effect of the other parameters
In Ref. [5], an empirical relationship between the con-
such as contact frequency, f, on ⌬Vw is ignored, then
tribution of corrosion, y (%), to total material loss and
normal load, W, was presented for abrasion of a low ⌬Vw / ⌬Vc⬀(W / H) / {[Aatoms / (zeFrv)]Kal
carbon steel in flowing mining water containing chlor- (26)
(W / H)1/2(i0⫺is)ft0[1⫺exp(⫺1 / ft0)]}
ides and sulphates under various conditions,
The boundary representing transitions from wear-
y ⫽ AW⫺0.458 (25) induced corrosion dominated regime to corrosion-
From (21), the wear-induced corrosion, ⌬Vc, is pro- induced wear dominated regime can thus be determ-
portional to W1/2, while the expression for the effect of ined by
load on the corrosion-induced wear is very complicated. W / H ⫽ kcw{ft0[1⫺exp(⫺1 / ft0)]}2 (27)
Wear (including sliding wear, abrasion and erosion) is
frequently observed to be proportional to load, W. Such Using an arbitrary value 1 for the proportionality con-
a linear relationship has also been observed in corrosive stant kcw, (27) is plotted in Fig. 2 where normalised para-
wear [5,12,15,16,25]. Thus, the contribution of wear- meters have been used for axes. W/H represents the
induced corrosion to the corrosive wear is predicted by severity of the loading condition and (ft0)⫺1 describes
678 J. Jiang et al. / Tribology International 35 (2002) 669–679

ergistic effect have been derived which incorporate most

of the influencing factors. Further studies are required to
verify and refine the concepts developed in the model.

References

[1] Kajdas C, Harvey SSK, Wilusz E. In: Encyclopaedia of tribology.

Tribology Series 15. Amsterdam: Elsevier; 1990. p. 7-6.
[2] Wei J, Xue Q. Effects of synthetic additives on the friction and
wear properties of a Cr2O3 coating. Wear 1994;176:213–6.
[3] Dunn DJ. Metal removal mechanisms comprising wear in mineral
processing. In: Ludema KC, editor. Wear of materials. New
Fig. 2. A wear map based on the present model showing the effect of York: ASME; 1985. p. 501–8.
loading intensity (W/H), interaction frequency, f, between encountering [4] Barker KC, Ball A. Synergistic abrasive corrosion wear of chro-
bodies and passivation characteristic time, t0, of the metal on dominant mium-containing steels. Brit Corros J 1989;24:222–8.
mechanisms for the synergistic effect. [5] Noël REJ, Ball A. On the synergistic effects of abrasion and cor-
rosion during wear. Wear 1983;87:351–61.
[6] Batchelor AW, Stachowiak GE. Predicting synergism between
the non-contact time between successive loading at a corrosion and abrasive wear. Wear 1988;123:281–91.
[7] Kotlyar DC, Pitt CH, Wadsworth ME. Simultaneous corrosion
given site on the wear surface, f⫺1, scaled to the charac- and abrasion measurements under grinding conditions. Corrosion
teristic depassivation time of the metal, t⫺1 0 . From this 1988;44:221–8.
simple map, the corrosion wear tends to be dominated [8] Yang J, Swisher JH. Erosion-corrosion behaviour and cathodic
by wear-induced corrosion at low loads and at high fre- protection of alloys in seawater–sand slurries. J Mater Eng Per-
quency of interactions between the counter object and form 1993;2:843–50.
[9] Madsen BW. Measurement of erosion-corrosion synergism with
the metal surface where the metal surface is kept at a a slurry wear test apparatus. Wear 1988;123:127–42.
high rate of dissolution. On the other hand, metals with [10] Neville A, Hodgkiess T, Dallas JT. A study of the erosion-cor-
high passivation capabilities (low t0) tends to be domi- rosion behaviour of engineering steels for marine pumping appli-
nated by corrosion-induced wear, especially when the cations. Wear 1995;186-187:497–507.
applied load is high and the sliding contact frequency is [11] Neville A, Hodgkiess T. A study of the effect of liquid corrosivity
in liquid-solid impingement of a cast iron and austenitic stainless
relatively low. steel. Br Corros J 1997;32(3):197–205.
It should be pointed out that, although reasonably [12] Neville A, Hodgkiess T, Xu H. An electrochemical and micro-
good agreement with published results in the literature structural assessment of erosion-corrosion of cast iron for marine
has been demonstrated, the present model is still a quali- applications. Wear 1999;235:596–607.
tative and descriptive one. Much more effort is required [13] Neville A, Hodgkiess T. Electrochemical and mechanical ero-
sion-corrosion of high-velocity oxy-fuel (HVOF) coatins. Wear
to justify the concepts presented here, to refine the model 1999;235:623–34.
and to extend the application of the model to particular [14] Neville A, Hodgkiess T. Characterisation of high-grade alloy
corrosion wear systems. behaviour in severe erosion-corrosion conditions. Wear
1999;235:523–34.
[15] Neville A, Reyes M, Hodgkiess T, Gledhill A. Mechanisms of
wear on a Co-based alloy in liquid-solid slurries. Wear
5. Summary 2000;238:138–50.
[16] Perry J, Neville A, Hodgkiess T, Wilson VA. Erosion-corrosion
A model for corrosive sliding wear of metals and the of WC-Co thermal sprayed cermet coatings in saline environ-
synergistic effect has been presented. According to this ments. Surf Coat Technol 2001;137(1):43–51.
model, the effect of wear on corrosion, or the wear- [17] Hu X, Neville A. An investigation of erosion-corrosion behaviour
of high-alloyed stainless steels under liquid–solid impingement
induced corrosion, is caused by the removal of passive conditions. Mater Corros 2001;52:1–9 [Werkstoffe Und Cor-
surface films by the rubbing action of the counter body, rosion].
leading to accelerated dissolution corrosion. The gener- [18] Stott FH, Breakell JE, Newman RC. The corrosion wear of cast-
ation of wear debris particles is regarded as a micro- iron under potentiostatically-controlled conditions in sulphuric-
crack initiation and propagation process analogous to a acid-solutions. Corros Sci 1990;30:813–30.
[19] Jiang XX, Li SZ, Tao DD, Yang JX. Accelerative effect of wear
low cycle fatigue process. The presence of corrosive on corrosion of high-alloy stainless steel. Corrosion
environment contributes to increased wear for two major 1993;49:836–41.
reasons: (a) localised corrosion attacks increase the [20] Zhang TC, Jiang XX, Li SZ, Lu XC. A quantitative estimation
micro-crack initiation/nucleation rate; and (b) of the synergy between corrosion and abrasion. Corros Sci
chemical/electrochemical interactions between active 1994;36:1953–62.
[21] Zhang TC, Jiang XX, Li SZ. Acceleration of corrosive wear of
species from the environment and bonding at the crack duplex stainless steel by chloride in 69% H3PO4 solution. Wear
tips accelerate the micro-crack propagation. Parametric 1996;199:253–9.
expressions for the total corrosion wear and for the syn- [22] Zhang TC, Jiang XX, Li SZ. Hydrogen-induced embrittlement
 J. Jiang et al. / Tribology International 35 (2002) 669–679 679

wear of a high-strength, low-alloy steel in an acidic environment. Crooker TW, Leis BN, editors. Corrosion fatigue: mechanics,
Corrosion 1997;53:200–5. metallurgy, electrochemistry, and engineering. ASTM STP 801.
[23] Jiang XX, Li SZ, Duan CT, Li M. A study of the corrosion wear ASTM; 1983. p. 5–25.
of Ti–6Al–4V in acidic medium. Wear 1989;129:293–301. [43] Wei LJD. Correlation between sustained-load and fatigue crack
[24] Watson SW, Madsen BW, Cramer SD. Wear-corrosion study of growth in high-strength steels. Mater Res Stand 1969;9(7):25–7.
white cast irons. Wear 1995;181-183:469–75. [44] Iwabuchi A, Sasaki T, Hori K, Tatsuyanage Y. Tribological
[25] Kim KY, Bhattacharyya S, Agarwala V. An electrochemical properties of SUS304 steel in seawater-(electrochemical approach
polarisation technique for evaluation of wear-corrosion in moving to the wear behaviour). JSME Int J Series I 1992;35:117–22.
components under stress. In: Ludema KC, editor. Wear of [45] Iwabuchi A, Tsukamoto T, Shimizu T, Yashiro H. The mech-
materials. New York: ASME; 1981. p. 772–8. anism of corrosion wear of an austenitic stainless steel in an aque-
[26] Hong MH, Pyun SI. Corrosion wear behaviour of 304-L stainless- ous electrolyte solution. Tribol Trans 1998;41:96–102.
steel in 1-N H2SO4 solution Part 1: effect of applied potential. [46] Yahage Y, Mizutani Y. Corrosion wear of cast-iron in sulphuric-
Wear 1991;147:59–67. acid. J Tribol-Trans, ASME 1987;F109:238–42.
[27] Hong MH, Pyun SI. Corrosion wear behaviour of 304-L stainless- [47] Sato N, Kudo K, Noda T. The anodic oxide film on iron in neutral
steel in 1-N H2SO4 solution Part 2: effect of chloride-ion concen- solution. Electrochim Acta 1971;16:1909–21.
tration. Wear 1991;147:69–78. [48] Tomashov ND, Chernova GP. In: Passivity and protection of met-
[28] Mischler S, Debaud S, Landolt D. Wear-accelerated corrosion als against corrosion. New York: Plenum Press; 1967. p. 3-.
of passive metals in tribocorrosion systems. J Electrochem Soc [49] Goswami KN, Staehle RW. Growth kinetics of passive films on
1998;145:750–8. Fe, Fe–Ni, Fe–Cr, Fe–Cr–Ni alloys. Electrochim Acta
[29] Archard JF. Contact and rubbing of flat surfaces. J Appl Phys 1971;16:1895–907.
1953;24:981–8. [50] Abd-El-Kader H, El-Raghy SM. Wear-corrosion mechanism of
[30] Archard JF, Hirst W. The wear of metals under unlubricated con- stainless steel in chloride media. Corros Sci 1986;26:647–53.
ditions. Proc R Soc Lond 1956;A236:397–410. [51] Zhou S, Stack MM, Newman RC. Characterisation of synergistic
[31] Adler TA, Walters RP. Corrosion and Wear of 304 stainless steel effects between erosion and corrosion in an aqueous environment
using a scratch test. Corros Sci 1992;33:1855–76. using electrochemical techniques. Corrosion 1996;52:934–46.
[32] Zhang TC, Jiang XX, Lu XC, Li SZ, Shi CX. Corrosion wear [52] Wood RJK, Hutton SP. The synergistic effect of erosion and cor-
of duplex stainless-steels in sulphuric-acidic solution containing rosion: trends in published results. Wear 1990;140:387–94.
chloride. Corrosion 1994;50:339–44. [53] Stack MM, Bray L. Interpretation of wastage mechanisms of
[33] Iwabuchi A, Tsukamoto T, Tatsuyanagi Y, Kawahara N, Nonaka materials exposed to elevated-temperature erosion-corrosion
T. Electrochemical approach to corrosion wear of SKD61 die using erosion-corrosion maps and computer-graphics. Wear
steel in Na2SO4 solution. Wear 1992;156:301–13. 1995;186:273–83.
[34] Jiang JIAREN, Arnell RD. The dependence of the fraction of [54] Stack MM, Corlett N, Zhou S. A methodology for the construc-
tion of the erosion-corrosion map in aqueous environments. Wear
material removed on the degree of penetration in single particle
1997;203:474–88.
abrasion of ductile materials. J Phys D: Appl Phys
[55] Akagaki T, Kato K. Wear mode diagram in lubricated sliding
1998;31:1163–7.
friction of carbon steel. Wear 1989;129:303–17.
[35] Cook RH, Skelton RP. Environment-dependence of the mechan-
[56] Stack MM, Corlett N, Zhou S. Construction of erosion-corrosion
ical properties of metals at high temperature. Int Metall Rev
maps for erosion in aqueous slurries. Mater Sci Technol
1974;19:199–222.
1996;12:662–72.
[36] Westwood ARC. Tewksbury lecture: control and application of
[57] Stack MM, Corlett N, Zhou S. Some thoughts on the effect of
environment-sensitive fracture processes. J Mater Sci
elastic rebounds on the boundaries of the aqueous erosion-cor-
1974;9:1871–95.
rosion map. Wear 1998;214:175–85.
[37] Lawn BR, Wilshaw TR. In: Fracture of brittle solids. London: [58] Lim SC, Ashby MF, Brunton JH. Wear-rate transitions and their
Cambridge University Press; 1975. p. 91–108. relationship to wear mechanisms. Acta Metall 1987;35:1343–8.
[38] Bronson A, Nelson J, Kang C. Analysis of the scribing technique [59] Ashby MF, Lim SC. Wear-mechanism maps. Scripta Metall
for determining the corrosion wear of an Fe-21wt% Ni alloy. Mater 1990;24:805–10.
Wear 1992;154:387–401. [60] Hokkirigawa K. Wear maps for metals and ceramics. In: Kozma
[39] Congleton J, Craig IH, Olieh RA, Parkins RN. Some electro- M, editor. Proceedings of sixth International Congress Tribology,
chemical and microstructural aspects of corrosion fatigue. In: August 30–September 2, Budapest—Hungary, vol. 5. 1993. p.
Crooker TW, Leis BN, editors. Mechanics, metallurgy, electro- 32–7.
chemistry, and engineering. ASTM STP 801. ASTM; 1983. p. [61] Hutchings IM. Ductile-brittle transitions and wear maps for the
367–89. erosion and abrasion of brittle materials. J Phys D: Appl Phys
[40] Leis BN, Rungta R, Mayfield ME, Beavers JA. Corrosion-fatigue 1992;25:A212–A21.
crack initiation in an iron-caustic system. In: Crooker TW, Leis [62] Childs THC. The sliding wear mechanisms of metals, mainly
BN, editors. Corrosion fatigue: mechanics, metallurgy, electro- steels. Tribol Int 1980;13:285–93.
chemistry, and engineering. ASTM STP 801. ASTM; 1983. p. [63] Childs THC. The mapping of metallic sliding wear. Proc Inst
197–228. Mech Eng 1988;C202:379–95.
[41] Santner JS, Kumar M. Corrosion-fatigue crack propagation rates [64] Lim SC, Ashby MF. Wear mechanism maps. Acta Metall
in commercial 7075 and P/M X7091 aluminium alloys. In: 1987;35:1–24.
Crooker TW, Leis BN, editors. Corrosion fatigue: mechanics, [65] Stack MM, Corlett N, Turgoose S. Some recent advances in the
metallurgy, electrochemistry, and engineering. ASTM STP 801. development of theoretical approaches for the construction of ero-
ASTM; 1983. p. 229–55. sion-corrosion maps in aqueous conditions. Wear 1999;233-
[42] Wei RP, Shim G. Fracture mechanics and corrosion fatigue. In: 235:535–41.