Chapter 6

Wear

1 Introduction
Wear is the gradual removal of material obtained at contacting surfaces in relative motion.
While friction results in important energy losses, wear is associated with increased mainte-
nance costs and costly machine downtime.
Wear phenomena are intimately linked to frictional processes. Recall that friction forces
are generally the result of two main physical processes: shearing and ploughing. If solid
surfaces in relative motion are not separated in some way, wear can be expected. Lubricants
are used to separate contacting surfaces in relative motion and thus to reduce wear. Lubri-
cants may completely separate the surfaces, as in fluid film lubrication or allow solid-solid
contact only at a restricted number of locations (mitigated solid contact) as in boundary
lubrication. The tribology of lubricants is discussed in another chapter. Here the focus is on
wear resulting from direct soli-solid contact.
Wear phenomena are heavily influenced by the fact that most engineering surfaces are
rough (and hence surfaces come in contact at single asperities and the real area of contact
is usually much smaller that the nominal contact area). Furthermore, wear behavior is also
influenced by the presence of adsorbed species and/or surface layers.
Many different wear mechanisms have been identified. A first classification of mechanisms
is based on their relative importance in engineering practice. According to this, the following
types of wear are often encountered:
Adhesive Wear (plus Fretting Wear). Encountered in 23-45 % of cases.

Abrasive Wear (plus Erosive Wear). Encountered in 36-58 % of cases

Surface Fatigue Wear. Encountered in 14-15 % of cases.

Corrosive Wear. Encountered in 4-5 % of cases.

Wear phenomena can also be classified based on the underlying physics. According to
this, mechanistic-based classification, the following types of wear have been identified:
Adhesion and Transfer. Wear takes place by gradual removal of adhered fragments of
material picked up by the contacting surfaces during frictional interaction.

1
 Corrosion Film Wear. Wear is associated with the removal of fragments of protective
corrosion/passivating layers from the surface of the worn material.

Cutting. Wear is the result of intermittent or continuous chip formation in the soft
material due to cutting action by a harder tool.

Plastic Deformation. Wear being associated with the removal of sheared layers resulting
from excessive plastic deformation.

Surface Jetting. Wear resulting from interfacial instabilities associated with localized
softening at the contact interface.

Surface Fracture. Wear resulting from breakage of atomic bonds in embrittled surface
layers.

Surface Fatigue. Wear resulting from subsurface cracking and fracture induced by
cyclic loading.

Surface Reactions. Wear associated with the removal of reaction products from the
surface which in turn were produced by frictional flash heating.

Melting. Wear associated with the melting transition.

Electrochemical. Gradual wear associated with anodic reactions on the worn surface.

Details about several of the above mentioned wear mechanisms are discussed in the sequel.
However, first we describe general aspects of wear processes, including the measurement of
wear as well as some details about adhesive wear.

2 Measurement of Wear
Wear rate measurements are routinely performed using standard or customized friction and
wear testing equipment. The same configurations used for friction measurements are also
used in wear research. While the occurrence of an initial running-in or breaking-in stage
is often detected when measuring wear, experiments tend to focus on the also commonly
encountered subsequent stage of the process: steady state wear.
The amount of material worn in a tribological system is measured in relation to the
duration and extent of contact. Specifically, a common measure of wear is the volume of
material removed per unit sliding distance. Consider a tribological system where a volume
Vw of the softer component of the couple is removed by wear in the same amount of time
it takes for the sliding distance to become L. The volume of worn material per unit sliding
distance is then w = Vw /L. The wear rate thus expressed has units of area and it is often
useful to compare the observed values of wear rate against the other most important area

2
measure encountered in friction studies, namely, the real area of contact of the friction couple
Ar which, for plastically deforming asperities, is simply given by
Fn
Ar =
H
Here Fn is the applied load and H is the hardness of the softer component in the friction
couple.
The ratio of the wear in units of volume removed per unit sliding distance to the real
interfacial area of contact is a meaningful dimensionless quantity useful in wear studies and
is called the dimensionless wear coefficient, or simply the wear coefficient, K. From the
above, this is defined as
Vw /L w
K= =
Fn /H Ar
A couple of useful physical interpretations can be given to the coefficient K. First, since,
K is the ratio of two areas, the worn area Aw and the real contact area Ar . One can
thus interpret K as the fraction of the real contact area Ar removed by wear. Consider a
tribological system consisting of a hard flat surface and a rough surface of a softer material
with of identical asperities. Assume the surfaces make contact at N aspertities. Hence, the
wear coefficient represents the number n of those asperities were conditions are such the
conditions that material is torn forming wear debris, i.e.
Aw n
K= =
Ar N
This interpretation, together with the commonly found low values of K found in practice
indicates that while all asperity contacts contribute to friction only a very small fraction of
contacts result in wear. Therefore, while friction is dominated by the contact events that
occur without damage, wear is dominated by the (small) chance that a particular contact
event results in rupture.
Note also that the quantity (Fn /H)L has dimensions of volume. This represents the total
volume of the plastically deformed zone underneath the surface of a worn area, Vp following
sliding by a distance L. Hence
Vw
K=
Vp
Therefore, K also represents the proportion of the plastically deformed volume that is re-
moved by the wear process.

3 Adhesive Wear
An adhesive wear model assumes that wear is the result of adhesion between asperities
followed by fracture. A first simple model of adhesive wear states that the volume of worn

3
material removed Vw as a result of a tribological interaction is directly proportional to the
load Fn and also to the total sliding distance L. However, since less wear is observed when the
hardness of the softer member of the tribological couple increases, Vw can also be regarded
as inversely proportional to the hardness H of the material being worn away. In symbols,
the above is represented as
Fn L
Vw = K
H
or expressed in terms of the worn volume per unit sliding distance w
Vw Fn
w= =K
L H
where K is the previously introduced wear coefficient. This expression is sometimes called
Archards law. Although an equation of this form was first proposed by Holm from his
studies on electric contacts, Archard first obtained it using a simplified model of the contact
interaction described below.
Note that if the sliding distance is the result of sliding at constant velocity U, it is then
given by L = Ut where t is the sliding time. If Archards law is divided by the nominal
(apparent) contact area Aa , substitutes the sliding distance in terms of sliding velocity and
time and solves for the time one gets
H 1
t=d
K pm U

where d = Vw /Aa is the worn depth and pm = Fn /Aa is the mean or nominal pressure. This
is an indication of the life of a wearing component in terms of the admissible worn depth
and the material and process parameter H, K, pm and U.
It is often the case, that measured worn volumes vary in direct proportion with the
total sliding distance. In contrast, while worn volumes often vary in proportion with the
applied load over certain load ranges, abrupt changes in wear rates (wear transitions) are
observed at specific critical loads. Such changes are the result on the complex interplay
between the softening and chemically reacting behaviors of the material induced by high
flash temperatures. Abrupt increases in wear rates are commonly found at high loads and
these are often associated with welding and seizure. However, in some cases these high wear
rates may revert to low values at even higher loads .
Measured values of K are frequently small and range from 108 for incompatible metals
rubbing against each other with good lubrication to 103 for clean unlubricated surfaces of
like metals. The smallness of these values together with the original interpretation of K given
above suggest a probabilistic interpretation of the wear coefficient, namely, it represents the
fraction of the actual contact surface which is actually removed by the wear process. It could
also represent the probability that any given individual friction contact event culminates with
the breakage and removal of a wear particle.

4
3.1 Some Experimental data on Adhesive Wear
Following is a short compilation of specific wear data obtained using a pin and disk system
with a load of 0.4 kgf and a speed of 1.8 m/s (adapted from Archards article in the Wear
Control Handbook).

Wearing Surface Counter Surface w (mm3/m) H (kg/mm2 ) K

Mild Steel Mild Steel 1.57 102 186 0.62 7 103
60-40 Brass Tool Steel 2.40 103 95 0.24 6 104
PTFE Tool Steel 2.00 104 5 0.18 2.4 105
Stellite Tool Steel 32 106 690 0.60 5.5 105
Ferritic SS Tool Steel 27 106 250 0.53 1.7 105
Polyethylene Tool Steel 3 106 17 0.53 1.3 107
WC WC 0.2 106 1300 0.35 1 106

Adhesive wear coefficient correlate well with friction coefficients, however, unlike friction
coefficient values, wear rates vary over a very large range.
Other experimental techniques useful in wear research include:

Microscopic examination of wear debris.

Metallographic examination of worn surfaces.

Radioactive isotope methods.

Besides the sharp differences in wear rates that exist between mild and severe wear, the
wear debris produced in both cases also exhibits significant differences. Specifically, wear
debris resulting from mild wear frequently consists of fine particles (0.01-1 micrometer),
many of them oxides. In contrast, the debris resulting from severe wear consists of much
larger (20-200 micrometer) plate-like particles consisting of an assembly of smaller particles
stuck together.
To gain insight into adhesive wear mechanisms it is useful to examine in some detail the
results of selected wear experiments carried out in well characterized systems. The advantage
of this approach is that the resulting information has been reproduced and the underlying
processes are generally well understood. The systems examined below include:

Wear of plain carbon steel rubbing against itself in air.

Wear of leaded brass (59Cu-39Zn-2Pb) rubbing against stellite in air.

Wear of 60-40 brass and ferritic stainless steel against tool steel in air.

5
Details about the observed wear mechanisms in these two systems are now described.
In the first case, wear observations are performed on a plain carbon steel pin rubbing
against a disk of the same material in air as a function of load under dynamic but otherwise
steady-state conditions.
When the applied normal load is relatively small, fine metal particles are removed from
the rubbing surfaces and at least some of them are rapidly oxidized while dispersed in the gap
between the contacting surfaces. The oxide particles then act as abrasive medium producing
further wear. However, the observed wear rates are relatively small, and this is called mild
wear.
As the load increases, larger metal particles are torn from the rubbing surfaces. The
oxidation rate induced by flash temperature is not large enough to fully oxidize the wear
particles and many particles are ejected from the gap and discarded as wear debris. Because
the wear rates involved are larger, this is called severe wear.
For the highest loads, the resulting high flash temperatures lead to rapid formation of
oxides and a hard, tempered layer on the contact surfaces. Fine metal particles are formed,
oxidized and removed from the gap between the rubbing surfaces as wear debris. Since the
contacting surfaces harden rapidly, the resulting wear rate is relatively small. This is known
as the high temperature regime of mild wear.
When one uses instead a leaded brass pin and rubs it against a disk made of stellite in
air the following scenario results.
At the smallest loads (< 0.1N) and low temperatures. the contacting surfaces meet
at fairly isolated asperity contact points. Small, discrete particles of brass are transferred
to asperities on the stellite counterface and some oxidize rapidly. On further rubbing, the
accumulated particles on counterface are smeared over and diffuse into the stellite surface and
form a continuous thin layer of brass containing oxidized particles mixed with counterface
material. Later on, material from the smeared layer at the counterface transfers back to the
pin forming a microcomposite hard layer on the pin surface while small composite particles
break away from the pin and the disk. forming loose wear debris. As a result of the above,
the wear rate is relatively low (mild wear).
As the applied load - and the resulting temperature - increase, the following distinct
sequence of events is obtained. The real area of contact increases and electrical contact
resistance at the interface reaches a minimum. Particles of brass are transferred to asperities
at the counterface and at least some oxidize rapidly. The accumulation of discrete particles
on the counterface results in aggregation into larger clusters on the surface of the disk. These
clusters subsequently break away from the disk forming loose wear debris and the resulting
wear rate is high (severe wear).
At the highest loads, the contacting surfaces meet at over a much larger contact area.
The resulting high flash temperatures lead to rapid surface oxidation. Particles of brass
are transferred to asperities at the counterface and oxidize rapidly. The oxide particles
are smeared over the contacting surfaces and effectively form an intervening layer between
the rubbing surfaces, while small particles continuously break away from the pin and the
disk. forming loose wear debris. The observed wear rate in this case is relatively low (high

6
temperature mild wear).
When the same pin and disk experiment is carried out using a stainless steel pin rubbing
against a tool steel disk one finds that Stainless steel exhibits mild wear for loads below 30
N and there is a transition to severe wear mode for larger loads.
When the pin and disk experiments are performed in a high temperature environment,
a somewhat different behavior is observed. Specifically, for the case of a brass pin at low
sliding speeds one finds that the imposed high temperatures produce some softening of the
material while the low flash temperatures resulting from the low sliding speed lead to minor
softening and oxidation. Small discrete particles of brass are transferred to asperities at
the counterface and oxidize. The brass particles are smeared over the counterface forming
a continuous thin layer of microcomposite material consisting of a brass matrix containing
oxidized particles and tool steel. The material from the layer continuously transfers back to
the pin forming a corresponding microcomposite hard layer on the pin surface. while small
composite particles break away from the pin and the disk. forming loose wear debris. The
resulting wear rate is low (mild wear).
For the same system, as the sliding speed increases the increased flash temperatures
lead to further material softening and oxidation, while the prevailing high temperatures
produce additional softening of the material. Small discrete particles of brass are transferred
to asperities at the counterface and accumulate there resulting in larger aggregates on the
surface of the disk. Such aggregates break away periodically from the disk forming loose
wear debris and the resulting wear rate is high (severe wear).
If beside increasing the sliding speed, one also increases the temperature of the pin and
disk system, one observes additional softening and significant oxidation (due to both the
high flash temperatures as well as the experiment temperature). Small particles of brass
are transferred to asperities at the counterface, oxidize rapidly while they accumulate and
smear over the counterface forming a continuous thin and microcomposite layer containing
large amounts of oxidized particles and counterface material that make it relatively hard.
From time to time, small composite particles break away from the pin and the disk surfaces
forming loose wear debris. As a result, the wear rate is relatively low (mild wear).
Many studies similar to those described above have been performed and a couple of
generic observations are the following. Wear occurs in both materials of a tribological system
regardless of the differences in hardness of the materials involved. However, in most systems,
the softer material wears out at a faster rate, as expected. Transitions from mild wear
to severe wear are the result of the competing processes of surface oxidation and surface
softening. At high oxidation rates, intervening oxide layers form readily reducing both
friction and wear. When the softening rate predominates large material removal rates result.

3.2 Adhesive Wear Modeling

In the adhesion theory of friction, asperities of the contacting surfaces come in contact against
each other briefly forming an adhesive junction. Assuming, correctly, that plastic behavior

7
prevails at the asperities, the friction coefficient becomes
Ft
= =
Fn H
where is the shear strength of the junction and H the hardness of the softer material. The
adhesive junctions formed during frictional processes play a key role in the development of
the adhesive wear mechanism and in the production of wear debris.
The following wear model was first proposed by Archard. Consider two asperities in a
tribological couple. The asperities first approach each other, then establish adhesive contact
and finally move past each other. As a result of the interaction, a wear particle may be
created. To simplify, consider asperities of hemispherical shape and radius a and assume the
volume of the wear particle Vw is approximately given by
2
Vw = a3
3
Moreover, let the normal load Fn supported by the (plastically deforming) asperity be given
by
Fn = a2H
where H is the hardness of the softer component in the contact couple. Finally, the sliding
distance for this asperity interaction is L = 2a.
Only a small fraction of contact interactions result in worn particles, therefore, let the
proportion of interactions resulting in wear be so that the volume of material removed per
unit sliding distance w is given by
Vw a2
w = =
L 3
The total amount of wear (per unit sliding distance) w is the sum over all interactions, while
the total load is the sum of all the loads supported by individual asperities. The wear rate
is then given by
Fn Fn
w= =K
3H H
where K = /3 is the wear coefficient.
Archards model yields an expression for the wear rate as a function solely of the macro-
scopic quantities Fn and H. It is well known that other factors significantly influence adhesive
wear rates, the most important include
Material Compatibility.
Crystal Structure.
Microstructure.
These effects are accounted for only implicitly in Archards model though the wear coefficient
K.

8
4 Other Forms of Wear
Material can be removed from a surface in a tribological couple by mechanisms other than
adhesive/sliding wear. Particularly important are abrasive wear mechanisms involving either
hard asperities or third bodies, erosive mechanisms involving high velocity swarms of particles
and wear assisted by chemical reactions.

4.1 Abrasive Wear

Abrasive wear takes place when hard asperities or third phase particles rub under load
against a relatively softer surface. If the wear process involves only two materials it is known
as two-body abrasive wear. If extraneous abrasive particles are used one has three-body
wear. In general, wear rates for two body wear are higher than for three body since loose
abrasive particles tend to roll over the soft surface while abrading asperities firmly attached
to the abrading surface do not.
Depending on the attack angle and the interfacial shear strength three modes of abrasive
wear are usually encountered in ductile metals:

Ploughing: Ridges form along the sides of wear track.

Wedging: A short wedge forms in front of the abrading asperity.

Cutting: A long ribbon-like chip forms in front of the abrading asperity.

In all cases wear particles form when ridges, wedges or chips fracture and detach from
the underlying surface, perhaps by a process of delamination.
The Archard wear equation derived for adhesive wear situations has also been found
useful in the representation of abrasive wear. Consider for instance the case of an abrasive
surface consisiting of conical asperities of included angle that ploughs through the surface
of a softer material. The volume of material removed from the ploughed groove per unit
sliding distance is
Vw Fn
cot Fn
w= = K ( ) = Kabr
L H H
where cot is the average value of cot for all asperities. Therefore, the abrasive wear coef-
ficient Kabr depends on the geometry of the abrading asperity and typically ranges between
106 to 101 .
In cutting wear, material removal proceeds as in machining by forming a chip. Specifically,
for cutting by a single, hard wedge asperity of width b perpendicular to the direction of
motion and rake angle under orthogonal cutting conditions the chip thickness has been
given by Merchant as
cos( )
= f
sin

9
where is the shear plane angle formed by the thin, primary deformation zone, extending
from the tip of the asperity to the free surface of the workpiece and the direction of motion
and f is the undeformed chip thickness (called the feed in machining application). While
the feed is a process parameter set by the machinist, the feed in this case can be expected
to be directly proportional to the applied (normal) load and inversely proportional to the
hardness of the material, i.e.
Fn
f = K
H
The volume of material removed as a chip by the wedge over a cutting time t is simply given
by
Fn L
Vw = Ufbt = Lfb = Kcut
H
Moreover, the thickness of the resulting worn particles should be of the same order as the
chip thickness .

4.2 Wear by Brittle Fracture

Brittle solids often fail under abrasive wear by subsurface lateral fracture crack propagation.
The volume of worn material per unit sliding distance as a result of N asperities rubbing
on a surface of modulus E, hardness H and fracture toughness Kc under load Fn has been
found to be given by

Vw (E/H)Fn9/8
w= = cN 1/2
L Kc H 5/8
where c is a material constant

4.3 Fatigue Wear

Subsurface cracks may also form in more ductile materials by fatigue processes associated
with cyclic loading. While little wear may be observed prior to fatigue events extensive
pitting may be observed once fatigue sets in.
Consider again a tribological system formed by a large number of microscopic contacts.
Each individual asperity experiences multiple contact interactions with asperities of the
counterface. When fatigue failure takes place at a particular asperity contact, the volume
of material removed can be expected to be proportional to a3 where a is the contact ra-
dius. Moreover, the sliding distance over a particular contact is 2a. An expression for the
worn volume per unit sliding distance for the system can then be obtained by adding the
contributions from all contacting asperities where fatigue has occurred, i.e.
Vw Fn
w= = Kf atigue Ar = Kf atigue
L H

10
Since K can be interpreted as the proportion of contacts that contribute to wear debris by
failing, the number of stress cycles to failure Nf is simply
1
Nf =
Kf atigue

Fatigue wear may also occur in rolling contacts even if they function under a fluid film
lubrication regime. The reason for this are the large cyclical stresses transmitted through
the lubricating film. The resistance of bearings against rolling contact fatigue is estimated
statistically through their lives. Specifically, the Weibull distribution with density function
(for t > 0) given by

t 1 t
f(t) = ( ) exp(( ) )

where and are the parameters of the distribution is used as a foundation of the Weibull
life model. The life (in millions of revolutions) for 90 percent of the bearings in a given
population is given by

L10 = (C/Fn )3

where C is the bearings basic load capacity. The above equation is adjusted by the use of
multiplicative factors to incorporate variations in geometry, material properties and lubrica-
tion conditions.

4.4 Impact Wear

Impact wear is the gradual wastage of material due to repeated impact by particulate streams
(erosive wear) or by continued hammering with a hard object (percussive wear).
In solid particle erosion, the kinetic energy of impacting particles is transferred to the
worn surface where is transformed into work of plastic deformation. Worn particles may
then by removed by ploughing, wedging or cutting (ductile materials) or by fracture (brittle
materials).
By equating the kinetic energy of impinging particles to the work of plastic deformation
and recognizing that only a fraction of impacts results in worn particles the following
expression has been obtained for the erosion ratio E (mass of worn particles divided by mass
of erosive particles) due to a total mass m of erosive particles impinging the surface with
velocity U,

U 2
E=
2H
where is the density of the worn material. Note that is again similar to Archards equation
but with the load Fn replaced by U 2 . Wear coefficients for erosive wear range between 105
and 101 .

11
 An expression identical to Archards law can be obtained by recognizing that the term
Fn L represents a sliding energy Es . Consider n identical particles of mass m impacting a
surface with the same velocity v. Their individual kinetic energy is 12 mv 2 Of the total kinetic
energy of the particles only a fraction results in material removal. Therefore the impacting
energy resulting in wear is given by Ei = Ek = 21 mv 2n. Therefore, using this impact
energy instead of the sliding energy into Archards equation yields
Ek
Vw = Kimpact
H
The wear coefficient in this case can then be interpreted as the fraction of all impact events
that results in material removal.
Liquid impingement erosion may result when liquid droplet sprays hit a surface at high
speeds due to the high pressures developed when the kinetic energy of the fluid particle is
transferred to the surface.
Erosion can also result from imploding bubbles near the surface since implosion results
in a fluid microjet directed at high speed against the surface.
When a surface is subjected to repetitive hard body impact, multiple wear mechanisms
may become simultaneously active. The wear volume is estimated per impact cycle. Specif-
ically for the impact of a slug with mass m of duration ti and peak force F0, onto a surface
moving tangentially with speed u, the worn volume is given b y
uti F0S
Vw = k
2H
where S = mu/F0ti is the slip factor and is the friction coefficient between the slug and
the surface.

4.5 Wear associated with Chemical Reactions

Tribological systems embedded in a corrosive environment often experience chemical attack.
Also, the localized heating produced by flash temperatures stimulates chemical reactions
which may also result in wear
Fretting Wear, the wastage of materials subjected to oscillatory motions of small am-
plitude is often associated with chemical reactions. As asperities on one surface slide past
the surface of the counterface, the latter may be rapidly oxidized. The oxide layer may be
intermittent or continuous and may stay in place or be removed by subsequent sliding. De-
pending of the fate of the oxide layer, the surface clearance may increase leading to sloppiness
or undesired seizure may occur.
A model of corrosive wear can be developed by considering that the contact interaction
takes places in a reactive environment that leads to the formation of a tribofilm on the
wearing surface. Wear takes place once the film reaches a critical thickness that becomes
easily removed by rubbing. At a single contacting asperity covered with a tribofilm of critical
thickness the worn volume is a2. Since the sliding distance is 2a the volume removed

12
per unit sliding distance during a single asperity interaction is 12 a. The result due to n
such identical asperity interactions is
Vw Fn
w= = na2 = Ar = Kch
L 2a 2a H
because Ar = Fn /H. Note that Kch is directly proportional to the tribofilm thickness.
Hence, increasing the chemical aggressiveness of the environment and/or the temperature
should increase the wear rate since both effects will make the film grow faster.

5 Examination of Wear Debris

Insight about wear mechanisms may be obtained by careful examination of wear debris.
Microscopic (optical and electron), Coulter analysis and Ferrography have all been used.
Particular interest exist in the deployment of wear debris examination methods for on-line
condition monitoring of tribological systems since abrupt changes in wear debris character-
istics are indicative of changes in wear mechanisms.

6 Wear Mechanism Maps

The various wear regimes and transitions observed in practice can often be usefully sum-
marized in the form wear mechanism maps. A map is usually a two dimensional graph of
normalized applied pressure against normalized relative velocity. Depending on the applica-
tion, other combinations of coordinates may be selected. Regions within the map indicate
process regions where a specific wear mechanism is dominant. Many different types of wear
mechanism maps are now available for a variety of tribological systems. These are useful
since one can identify at a glance specific values of tribosystem parameters that are most
likely to yield desired results.

13