Home Search Collections Journals About Contact us My IOPscience

A review of micro-scale abrasion testing

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2011 J. Phys. D: Appl. Phys. 44 073001

(http://iopscience.iop.org/0022-3727/44/7/073001)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 93.180.53.211
This content was downloaded on 22/11/2013 at 07:20

Please note that terms and conditions apply.

IOP PUBLISHING JOURNAL OF PHYSICS D: APPLIED PHYSICS
J. Phys. D: Appl. Phys. 44 (2011) 073001 (15pp) doi:10.1088/0022-3727/44/7/073001

TOPICAL REVIEW

A review of micro-scale abrasion testing

A J Gant and M G Gee
Electronics and Modelling Group, Division of Industry and Innovation, Module 9, Space G9-A5,
Queens Road, Teddington, Middlesex ,TW11 0LW, UK

Received 18 March 2010, in final form 23 June 2010

Published 28 January 2011
Online at stacks.iop.org/JPhysD/44/073001

Abstract
Micro-scale abrasion (commonly referred to as ‘ball cratering’) is a small-scale tribological
test method which can be operated on a desktop. It offers the possibility of providing a quick,
cheap, localized abrasion test that can be used with small samples. In principle its operation is
simple, but in practice there are issues with wear scar measurement, wear mode and its
applicability to a wide variety of monolithic materials and coatings.
(Some figures in this article are in colour only in the electronic version)

1. Principle of the test

The basic principle of the test is to press a rotating ball into the
surface of a test sample in the presence of a slurry suspension of
abrasive that is drip fed onto the interface between the sample
and the ball (figure 1). A spherical depression (crater) in the
sample is produced, providing a measure of the wear that has
taken place to the sample. By comparing the size of the crater
produced under the same conditions for different materials, a
comparison in the abrasion performance for different materials
can be obtained. The size of the crater can be measured by a
number of methods, most often through optical microscopy
but also through various types of profilometry.

2. Origins of the test

Figure 1. Schematic diagram of micro-scale abrasion test.
The basic test concept derives from the need to determine
composition-depth profiles in Auger electron spectroscopy [1]; amongst these early authors; namely either the introduction
the provision of accurate composition-depth profiles to depths of a fine (1 µm) metallographic paste into the ball/cylinder–
>1 µm is important in the characterization and investigation sample interface [3], or a fine spherical abrasive suspended in a
of thick films, surface coatings, surface treatments and medium of light oil [1, 2]. Samples prepared by these methods
many other technologically important applications. Formerly needed shallow taper angles, which required large diameter
samples were prepared using ion bombardment as a surface cylinders or balls (up to 50 mm). Conformal spherical cap
layer removal method on the micrometre scale [1–4]. Ion geometries thus proved useful for analysis by virtue of their
bombardment was found to be long and tedious when the sole predetermined well-defined taper angles. Since crater-depth
intention was to remove a depth in the order of micrometres; profiles are known accurately, interface widths can be obtained
the new method offered the possibility of being quicker and directly using line-scan analysis.
simpler to apply. Essentially a cylinder [1, 2] or a ball [3, 4] was Walls et al [4] was perhaps the first to highlight the
allowed to rotate freely against a sample in conjunction with an materials limitation of the preparation method (and hence
abrasive medium. The nature of the abrasive medium differed for the future application of its modern-day tribological

0022-3727/11/073001+15$33.00 1 © 2011 IOP Publishing Ltd Printed in the UK & the USA
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

Drive shaft Ball Pivot Lever arm

Pivot

Sample

Figure 2. Schematic diagram of free ball test. Abrasive feed

descendent); namely that it cannot be readily used on brittle

surfaces such as thick/loose oxides (which fracture rather than
abrade in a controlled manner) and neither for soft materials Weight
which smear.

3. Principal variants of the tests Test ball

Sample
There are two principal variants of the ball crater test which
have gained widespread acceptance amongst the tribological Figure 3. Schematic diagram of fixed ball test.
community; the ‘free ball’ and the ‘fixed ball’ systems. The
notation refers to the means by which the ball is rotated with accurate reporting of the number of revolutions completed and
respect to the sample and is explained below. hence distance slid. The sample is pressed into the rotating
ball from the side by test loads placed on the weight hanger.
3.1. Free ball This design also allows for the application of loads exceeding
the mass of the ball. Test systems of this type are now also
In free ball testing (figure 2) the ball is rotated against the marketed within Europe [7]. One drawback of the test is that
sample by a notched rotating shaft, with the ball being driven the ball is constrained to rotate in a fixed orientation which
by friction between the sample and the shaft. gives rise to the possibility of the ball developing a wear track
The ball is located by placing it in the notch on the shaft. which is flat in profile, producing non-spherical craters on the
Because the ball is not directly driven there is the possibility
test specimen.
of slip occurring between the shaft and the ball, giving rise
to uncertainty in terms of ball rotational speed and number
of revolutions made. This can, however, be mitigated by the 4. Test metrology: what the test can tell you
use of rubber driving elements on the shaft [5]. The normal
load is effectively produced by dead weight loading of the As the size of the contact between sample and test ball is
ball on the test specimen. The normal load on the specimen relatively small (of the order of a millimetre), the micro-
can be adjusted by changing the inclination of the sample scale abrasion test is ideally suited to testing small samples
holding plate. There are, however, practical limitations in of material, or the variation in wear performance from one
increasing the load by reducing the sample inclination; as position on a large sample to another. It is also ideally suited
there is an increasing tendency for the ball to slide up the to studying the wear of coatings due to the localization of
sample (thereby producing non-spherical craters [5]) as the wear. Normally micro-abrasion tests, as the name suggests,
inclination is reduced. Normal loads are calculated by simple are carried out with an abrasive slurry, but the test system can
resolution of forces on an inclined plane and do not take into also be used to carry out dry or lubricated sliding tests where
account friction between ball and plane; this can be a source friction can be measured through the addition of a friction
of error as the reaction of the ball on the plane is reduced. In a measurement device.
commercially available version of this test, this potential error For coatings there are two district variants of the test.
is mitigated by the use of a load cell under the specimen holding Where the test has a short duration or the coating is relatively
plate to measure the actual reaction force [6]. It is clear that the thick, testing can be carried out without perforating the coating.
applied load using this test format is limited both by the weight This variant of the test can also be used for monolithic
of the test ball and the range of inclinations which can be used materials. When the test is carried out on thin coatings,
to produce sufficiently spherical craters; in practical terms this perforation of the coating into the substrate can occur and wear
limits the applied loads to a maximum of approximately 0.4 N. rates for both the coating and the substrate can be obtained.

3.2. Fixed ball 4.1. Tests without perforation

In this test (figure 3) the ball is driven directly by clamping the For monolithic materials, or for tests where perforation of the
ball against two co-axial drive shafts which allows balls to be coating does not occur the volume V of a crater is given by
removed and replaced easily.
A rubber ‘O’ ring in the driven co-axial shaft is π b4
V = , (1)
incorporated to avoid slippage between ball and shaft, ensuring 16R

2
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

1.4 Wear vs. metal content

W-C:H and Ti-C:H coatings
1.3 14

Specific Wear Rate, 10-5mm3N-1m-1

Outer Crater Diameter, mm

1.2 12
10
1.1
8 W-C:H
1.0
6 Ti-C:H
0.9 4
0.8 2
TiN 0
0.7 TiN Perforated 0 5 10 15 20 25 30 35
MoS2
0.6 Metal content [at%]
DLC-A
DLC-A Perforated
0.5
DLC-B
Figure 5. Use of non-perforating test to determine wear rate of
metal containing DLC coatings as a function of metal content.
0.4
0 100 200 300 400 500 Free ball-single shaft system [5].
(a) Number of Revolutions
6.0x10
-6
Equation (1) is used to calculate the wear rates shown in
figure 4(b). Only the values for crater diameter for the craters
TiN
5.0x10
-6
MoS2
where perforation does not occur are used. It can be seen that
-1

DLC-A reasonably consistent (independent of sliding distance) wear

Wear Rate, mm N mm

DLC-B
-6 rates are obtained for the four different coatings.
-1

4.0x10
Another example is shown in figure 5 which shows the
3

3.0x10
-6 variation in wear rate with metal content for W–C : H and
Ti–C : H coatings [5].
2.0x10
-6 This technique has been shown to be effective for coatings
as thin as 1 µm [8].
-6
1.0x10
4.2. Tests with perforation
0.0
0 50 100 150 200 250 When perforation of the coating occurs, the procedure that was
(b) Number of Revolutions originally developed by Rutherford and Hutchings [9–12] can
be used.
Figure 4. Tests performed on a range of coated samples, A series of tests are performed with different durations.
(a) increase in overall crater diameter with increased test duration,
(b) calculated wear rates from non-perforated craters.
The inner crater diameters (a in figure 1) and overall crater
diameter (b) are both measured. There is a steady increase in
the size of both diameters. Perforation of the coating occurs at
with b the crater diameter and R the ball radius (for b
R), a relatively early stage of the testing.
as per nomenclature illustrated in figure 1. The analysis assumes that progress of the wear in a
This relationship assumes that the shape of the crater is cratering test follows Archard’s law, but this needs to be
conformal to the shape of the ball, which is true in most cases. modified to take account of the perforation of the coating after
The Archard wear law states that the volume of wear a short while. This is done by using an extended Archard’s
V = KSN, where K is a constant (the wear rate), S is the law for a coating–substrate system which can then be used:
sliding distance and N is the applied load. This has often been V c Vs
found to be true, but in some cases a strong dependence of SN = + , (3)
Kc K s
wear rate on the number of revolutions and hence the sliding
distance was found [7]. where Kc and Ks are the wear coefficients of the coating
However, if the Archard wear law is followed, the wear and substrate, respectively, Vc and Vs are the measured wear
rate K is described by volumes and SN the sliding distance multiplied by the applied
load.
πb4 1 There are different ways of implementing equation (3).
K= · . (2) It could be solved using a multiple least squares approach.
16R SN
The wear coefficients Kc and Ks are then considered both as
By performing tests without perforation of the coating wear unknowns. However, the equation is rather unstable, because
rates can be measured for different coatings. This process of the relationship between Vc and Vs .
is illustrated in figure 4 where a number of experiments A better methodology is to rewrite equation (3) as a linear
have been performed on a range of coatings. In figure 4(a) function. Conventionally two forms are used. The first one is
measurements of the overall crater diameter are given. As the

test duration (number of revolutions) increases the size of the SN 1 1 Vc 1
= − + , (4)
crater increases. Perforation occurs for some of the coatings. Vt Kc K s Vt K s

3
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

where Vt is the total worn volume. The formula is more figure 6(c) where the true crater edge (marked by arrows)
stable when (Vc /Vt ) < 0.5. A second way of rewriting is surrounded by an area of scuffing. This scuffing can be
equation (4) is visualized as part of the crater leading to an overestimate in

 the crater diameter.
SN 1 1 Vs 1 The definition of the crater edges gets more difficult
= − + . (5)
Vt Ks K c Vt K c to define correctly as the abrasive slurry becomes coarser.
This can be mitigated by the use of profilometry
For further details on the analysis of results see [13].
(however, profilometry cannot of course differentiate between
substrate/coating crater boundaries that show the true shape
5. Repeatability and reproducibility of the crater). In particular 3D optical microscopy has
been shown [20] to provide excellent measurements of crater
Gee et al [14, 15] participated in an EU-funded project volumes. Figure 7 shows a crater produced on a TiN-
(‘CRATER’) which aimed at standardizing the test, and in coated tool steel sample visualized by a 3D focus variation
so doing, establishing best practice that was consequently microscope with (a) high resolution image optimized to show
published as ISO standard 26424:2008 [16]. An both the substrate and the coating clearly, (b) 3D height
interlaboratory exercise in which ten partners participated map, a simulacra and (c) the image superimposed on the
produced data on reproducibility and repeatability. The height map and (d) profile taken across the crater showing
reproducibility and repeatability of the results were analysed rounding of the edge of the crater. Because the height map
following the procedure laid down in ISO 5725 Part 2 and the image data are closely integrated it is possible to
[17]. The results of this analysis are given in analyse this data to produce an accurate measurement of the
table 1. substrate crater size and the total crater size for direct use in
For the non-perforating tests for the coating wear rate equation (4) [20].
Kc the repeatability standard deviation (measure of expected The vast majority of work reported on micro-abrasion has
variation when a test is repeated in a single laboratory on involved the use of fine slurries, with 4 µm silicon carbide
the same machine with the same operator) Sr is 24% and being a popular abrasive to use. With this abrasive medium,
the reproducibility standard deviation (measure of expected good crater edge definition can be produced on a wide variety
variation when a test is carried out within different laboratories of monolithic materials and coatings in the three-body wear
by different operators) sR is 26%. In the perforating tests, for regime. However, Stachowiak and Stachowiak [21] used silica
the substrate wear rate Ks Sr was 7% and sR was 11%. For the sand and other abradants, using particle sizes commonly found
non-perforating test coating wear rate Sr was 8% and sR was in macroscopic abrasion tests; namely 250–300 µm. Two main
17%. For the measurement of crater diameter Sr was 2% and drawbacks were found using this particle size. Firstly, crater
sR was 3%. edge definition was generally inferior to that using micro-
These values of reproducibility and repeatability are abrasives. Secondly, control over the prevailing wear regime
considered to be acceptable for a wear test and are comparable appeared difficult; mostly co-existing regions of two-body and
to other wear tests. Thus, the ASTM G99 pin-on-disc sliding three-body abrasion were found.
wear test quotes a standard deviation of wear scar diameter However, 3D optical measurement instruments are time
of 0.27 for a value of 2.11, or 12.8% [18] compared with consuming to use and are not yet readily available. The crater
the reproducibility calculated here for b of 2.9%, and the project found that an alternative accurate method of assessing
ASTM G133 reciprocating test quotes a standard deviation coating wear was to avoid the issues of the measurement of
of wear volume for silicon nitride of 0.189 on a mean value total crater diameter by measuring the thickness of the coating
of 0.543 mm3 , or 34.8%, compared with the reproducibility instead; this was found to be an equivalent and more accurate
calculated here, for example for Kc in the perforating tests of method of measurement method for the assessment of the
26% [19]. coating wear rate. The coating thickness can be measured
by the cap grinding method (i.e. using a much finer abrasive
6. Improving test accuracy slurry) [15]. It is also possible to use other independent
thickness measurements to give a mean value that can be used
It has been recognized for some time that silicon carbide in the analysis. Equation (5) is then used to complete the
abrasive slurries normally used in these experiments can analysis.
produce ill-defined crater edges. If crater edges are not
clearly defined this obviously has an impact on the accuracy
7. Material applications of the test
of thickness measurement. This issue is illustrated in figure 6.
Figure 6(a) shows a comparison between two craters that have 7.1. Polymers
been placed close together and have been imaged optically.
Figure 6(b) shows the same two craters, but here the image The micro-scale abrasion of filled and unfilled polyamide-
is formed from height information obtained by a white light 11 coatings was reported by Bello and Wood [22]; yielding
interferometer. It can be seen that the gap between the two a surprising result that the fillers (TiO2 ) did not act as
craters is apparently much narrower in figure 6(a) than in load-bearing elements within the coating matrix; however, it
figure 6(b). The reason for this discrepancy is shown in may be the case that the volume fraction addition of fillers

4
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

Table 1. Results of analysis of reproducibility and repeatability to ISO 5725 (figures in brackets are percentage of means).
Perforating tests Non-perforating tests
Number of
participants 13 13 12 4
−13 3 −1 −1 −13 3 −1 −1 −13 3 −1 −1
Kc , 10 m N m Ks , 10 m N m Kc , 10 m N m b, mm
Mean 8.00 8.34 5.35 0.561
Sr 1.94 (24) 0.57 (7) 0.41 (8) 0.011 (2)
sR 2.09 (26) 0.92 (11) 0.94 (17) 0.016 (3)

(14–16.5%) was not sufficient to have any significant effect 7.2. Cermets, ceramics, tool steels, hardmetals and hard
on load portioning between the polyamide matrix and filler. coatings
Nevertheless, these materials did show a good correlation
In recent years a number of papers have been published
between specific wear rate (κ) and the product of strain at
reporting the micro-scale abrasion of hard materials both in
break multiplied by stress at break (εb σb ); a modified form
the form of coatings and monolithic materials. One important
of the Ratner–Lancaster correlation [23]. Micro-ploughing
aspect of the behaviour of these materials in micro-scale
was found to be the dominant wear mechanism characterized
abrasion is that not only can rolling and grooving wear as well
by the formation of grooves, typical of two-body abrasion,
as ridging occur, but also wear by fracture. However, wear
with extruded filaments of polyamide at the edges of the
by fracture cannot be entirely divorced from rolling, grooving
grooves. Figure 8 shows the Bello and Wood relationship
or ridging since it is merely the response of a (brittle) test
between specific wear rate and the modified Ratner–Lancaster
material to the passage of abrasive particles occurring under
parameter. The authors also show comparative results from
a given loading regime. Micro-scale abrasion of WC-based
Shipway and Ngao [24] on other common bulk polymers
hardmetals (commonly referred to as ‘cemented carbides’) has
such as HDPE (high density polythene), PMMA (Poly(methyl been studied by a number of workers in the field [26–30] both as
methacrylate)), PS (polystyrene) and polyvinyl chloride monolithic materials and as coatings. In (macroscale) abrasion
(PVC) from micro-abrasion tests at 0.5 N, 17.2% SiC slurry, testing WC hardmetals exhibit varying degrees of plasticity
and 0.112 m s−1 ball speed, which generated predominantly (binder extrusion, slip lines in WC grains), fracture (of WC
grooving type abrasion. It can be seen that the monolithic grains) and pluckout of WC grains depending on the nature
PA11 coatings show better wear resistance than the polymers of the contact conditions and whether the test is conducted
tested by Shipway and Ngao [24], assuming the differences in dry or with a carrier medium [31]. As a general maxim,
test conditions do not greatly influence the results. It is also increased severity of contact conditions (high load per abrasive
evident that the PA11 samples have a different dependence on particle, rigid counterbody and/or dry conditions) promotes
1/(σb εb ) values. wear by fracture of WC grains whereas more benign wear
Buchanan and Shipway [25] proposed a fracture- regimes promote binder extrusion and undermining (pluckout)
dominated wear mechanism in ultra-high molecular weight of WC. Despite the competing mechanisms of volume loss in
polyethylene (UHMPWPE) after partial disentanglement of macroscale abrasion, wear of this type of material appears to
the polymer chains. The abrasion rate was found to increase follow a modified form of the Archard equation, in which wear
with hardness and yield strength of the polymer; counter- varies with respect to hardness in accordance with an inverse
intuitive if plasticity-dominated wear occurs. However, power law.
if wear occurs by brittle fracture this would explain Micro-scale abrasion of WC hardmetals appears to be
wear being a fracture-dominated process (following chain fundamentally different in at least one important respect; the
disentanglement). Shipway also used balls of differing contact and/or abradant motion conditions are such that they
materials. Nylon balls produced high wear and low scatter are unable to fracture WC grains. Pluckout of small WC
in results; postulated as being due to nylon allowing slurry grains occurs, leaving the large WC grains exposed [26, 27, 30]
embedment being low in modulus. Bearing steel (AISI 52100) (figure 9). Binder removal was postulated by Chen et al [30]
balls produced less wear but higher scatter; this was thought to be a prerequisite for WC grain pluckout; abrasion of WC-
due to them promoting embedment of abrasive into the polymer based hardmetals using more conventional abrasives such as
specimen. Conversely, particle embedment does not occur silica sand or alumina of the order 200–300 µm in size is
with nylon balls (and the specimen) as both are similar in commonly found to obey a Hall–Petch type relationship [32];
terms of hardness and modulus; thus promoting the passage whereby wear volume loss is inversely proportional to the
of particles through the contact zone. Repeatability was found square root of the WC grain linear intercept [26]. The fact
to be better with the nylon ball; a reflection of the lack of that in micro-scale abrasion the smaller grains effectively wear
embedment in the sample; choice of ball material with respect more quickly than the larger ones would seem to suggest that
to specimen material is important; basically to ensure abrasive Hall–Petch/wear relationship would actually be inverted; this
entrainment occurs through careful consideration should be was indeed found to be the case by Gant et al [26] when
given to the relative hardness/modulus of the ball/specimen reporting the comparative micro-scale abrasion behaviour of a
couple. number of WC-based hardmetals. Chen et al [30], working on

5
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

(d)

Figure 7. Examination of crater with 3D focus variation

microscope, (a) high resolution image, (b) height contrast map,
(c) image superimposed on height information to give 3D simulacra
of surface, (d) profile across crater showing rounding of crater edge.

Figure 9 shows a sequence of micrographs of a specific

Figure 6. Craters produced in TiN-coated tool steel samples, (a) area on a 6% wt. Co hardmetal (hardness 1576 HV30;
pair of craters viewed by low power optical microscopy, (b) the WC arithmetic mean linear intercept 1.15 µm). In the early
same pair of craters as in (a) visualized by white light interferometer
and (c) edge of another crater showing grooving inside true crater stages of wear the removal of binder phase can be seen
boundary and scuffing outside (approximate position of true crater clearly, followed by progressive removal of WC grains as wear
boundary marked with arrows). proceeds.
Stack and Mathew [33] have also reported results on
high velocity oxy-fuel (HVOF) coatings noted that nanoscale WC/Co coatings and plotted wastage maps of applied load
WC/Co coatings wore appreciably faster than those with a versus sliding distance. It was found that greater wear rates
conventional microstructure due to the fact that the depth of were produced at higher sliding distances; it could be the case
binder removal necessary for pluckout of a nanoscale WC grain that at long test durations, larger WC grains are plucked out
was much smaller than that for a WC grain typically present and/or fractured, but micrographs of wear scars produced after
in a conventional scale WC/Co HVOF coating. long test durations were not published. In a later paper [29]

6
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

that the micro-scale abrasion test was capable of discriminating

clearly between different coatings and could yield a statistically
consistent wear coefficient (κs ) for the substrates. The results
indicated that the test could be used to measure independent
wear coefficients for both coating and substrate from a single
test. The test was also shown to produce coating wear
coefficients independent of coating adhesion; unlike erosion
testing [38].
Batista et al [39] investigated the micro-scale abrasion
behaviour of PVD TiN coatings on untreated and plasma
nitrided AISI H13 steel. They found that the coating wear
coefficient, κc was not independent of the surface condition of
the (H13) substrate (unlike Rutherford and Hutchings [11])
Figure 8. Bello and Wood [22]: specific wear rate plotted against a prior to being PVD coated. It was reported that plasma
modified Ratner–Lancaster parameter and comparative data from
nitriding approximately halved the value of κc . However,
Shipway and Ngao [24].
the substrate wear coefficients (κs ) were largely unaffected by
surface treatment. One possible explanation for the change in
the same authors showed that in HVOF WC/Co coatings it κc values could be that nitriding the substrate may change the
was possible at high applied loads to not only induce ridge coating’s net elastic response to external loading.
formation but also oxide formation due to slurry starvation and Batista et al [40] also used ball cratering to study Cr–N
hence direct contact between ball and flat specimen, producing coatings, again on H13 steel substrates, but found that Cr–N
frictional heating. coatings on H13 had a higher wear coefficient than uncoated
Jones [34] has reported on micro-scale abrasion of silicon H13. The Cr–N was relatively soft with respect to the hardness
nitride and sialon–TiB2 composites. The paper features a of the SiC in the abrasive slurry, but more importantly, the SiC
number of micrographs of wear surfaces that are of note; firstly particles were larger than the coating depth; hence they were
in the case of silicon nitride it appears that both grooving able to plough through the depth of the coating, leading to
wear and microfracture can occur (instanced by the free ball delamination of the coating rather than abrasion.
test) simultaneously. Unlike metallic materials of moderate
Although strictly speaking it is not ‘ball cratering’, the
hardness, grooving wear does seem to produce a well-defined
dimple grinder method can be used for micro-scale abrasion.
crater periphery in this ceramic. The behaviour of the sialon
The test format was used by Hedenqvist et al [41] to study
cermet was akin to that of WC hardmetals; the second phase
PVD TiN coatings deposited at low temperature. The test
is preferentially removed, leaving the sialon grains standing
produced ‘microcutting’; a form of two-body wear, but also
proud, being worn down at a relatively slow rate by a process
adhesive and cohesive failure in TiN-coated ASP30 high
consonant with a metallographic polish. Jones also reported
speed steel.
the use of a tool steel (D2) in the test and showed ridging
occurring at a load of 2.5 N. However, the ridging appears
to be quite different to that found commonly in the test on 7.4. Other coatings and surface treatments
other materials when using high applied loads; reported in
detail first by Shipway [35] for soda-lime glass. The tool Although the test was originally devised for the measurement
steel ridging is quite distinct in three principal details; firstly of the wear rates of wear resistant thin coatings typically
it covers almost two thirds of the wear scar (it is not a central produced by PVD, it can be readily applied to many other
diametral ‘ribbon’), secondly it does not bear the hallmarks of coatings and surface treatments. Figure 10 shows some
slurry starvation found in other materials (reported in a later examples of application.
section concerned specifically with ridging); there appears to Figure 10(a) shows the micro-scale abrasion of a
be no direct physical contact between ball and ridge crests and multilayer coating. Because of the multilayer nature of this
thirdly the characteristic wear mechanism appears to be 3-body coating the straightforward analysis described earlier cannot
rolling. The ridge formation found by Jones does appear to be be used, although the analysis could be extended in principle.
similar in nature to that found by Dube and Hutchings (in mild The figure shows how micro-scale abrasion can be used to
steel) [36] and by Gant and Gee (in high speed steel and WC determine the structure of multilayer coatings in a similar
hardmetals) [37] under high-stress abrasion conditions and was manner to using taper sections. However, because the micro-
attributed to a preferential ‘channelling’ of the abradant. scale abrasion test produces craters that are conformal with the
test ball, the geometry of the crater is well defined so that the
thickness of the different layers can be determined.
7.3. PVD coatings
Figure 10(b) shows the application of the test method to
Rutherford et al [10] were among the first in the field to a traditional hard chrome coating, showing a well defined and
fulfil the potential of the ball crater test in terms of wear easy to measure crater. Anodized aluminium, laser cladding,
characterization of coated materials. Working on PVD TiN, codeposited chromium carbide with cobalt and enamel can also
TiCN, TiAlN and CrN coated M2 high speed steel, they showed be effectively evaluated with the test (figures 10(c)–(f )).

7
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

Figure 9. Sequence of micrographs from same area on sample A2a produced with load of 0.25 N with 20% SiC, (a) 2 cycles, (b) 10 cycles,
(c) 20 cycles, (d) 30 cycles and (e) 50 cycles.

7.5. Glass shown that both modes can operate concurrently, effectively
producing a ‘mixed mode’ abrasion regime. Williams
Microabrasion of glass has been extensively studied by
and Hynica [43] sought to further the understanding of
Shipway [35, 42]. The principal aim of Shipway’s studies has
abrasive wear mechanisms in lubricated contacts by modelling,
been to investigate the formation of ribbon/ridge formation in
showing from force equilibrium considerations (effectively a
ball cratering scars, but at the same time scanning electron
couple is produced when a model rhombic-shaped particle
microscopy was undertaken which showed that where slurry
is trapped between two opposing surfaces), that a change in
starvation and hence ridge formation had not occurred, the
wear mechanism ought to occur at a critical ratio D/h; where
principal wear mechanism was wear by (micro-) fracture. The
‘D’ is the largest particle dimension and ‘h’ the smallest film
wear mode of the glass was also found to be influenced by
choice of ball material [35]; stiff metallic and ceramic balls thickness; see figure 11. They showed that the critical ratio was
produced wear by chipping and fracture in the glass, whereas approximately 2. Williams and Hynica did, however, conduct
the use of compliant polymeric balls resulted in ductile cutting their experiments using a foil bearing in which separation of
of the glass by a two-body wear mechanism. the surfaces (effectively ‘h’) was determined by hydrodynamic
pressure; in the case of micro-abrasion it is determined by
load and the number of particles in the contact zone; in effect,
8. Operating regimes of the test load borne per particle [44]. Trezona et al [44] undertook
an extensive study to investigate the principal determinants of
8.1. Two- and three-body wear
wear mode (two- versus three-body) in micro-scale abrasion.
There is general consensus amongst authors that the test can Basically the higher load per particle, the greater the propensity
operate in 2-body, 3-body wear regimes, or indeed it has been for two-body (grooving) to occur. Higher load per particle is

8
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

Figure 10. Examples of micro-scale abrasion applied to engineered surfaces, (a) TiCN/TiC/Al2 O3 produced with 0.25 µm diamond
abrasive, (b) hard chrome after 500 revolutions under a load of 0.2 N with alumina abrasive, (c) anodized alumina with SiC abrasive,
(d) crater produced on Ni/WC laser cladding revealing WC particle structure, (e) codeposited electroplate and (f ) enamel.

favoured by of particles when there are very few of them in contact.

• low slurry concentrations, Wear rate was found to vary with abrasive volume fraction
• high applied loads. in a non-linear way, with the most severe wear occurring at
intermediate concentrations. At low concentrations, the wear
In mixed mode regimes it was found that two-body wear
rate is influenced predominantly by slurry concentration; the
occurred at the centre of the contact patch and a three-body
regime prevailed towards the contact zone periphery. normal (applied) load is a much weaker factor than at high
The authors also showed that the three-body regime abrasive volume fractions.
obeyed Archard’s law [45] (wear was proportional to distance To further the Williams and Hynica [43] criterion and
slid and load; but for two-body, volume loss was proportional adapt it specifically for micro-abrasion, Adachi and Hutchings
to a power of load less than unity. At very low abrasive [46] showed that the critical surface separation could be
volume fractions, the volume abraded was found to be virtually described in terms of the ‘severity of contact’ (‘S’); where
independent of normal load; explained by the total embedment S is a dimensionless group W /AvH . The H  parameter is

9
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

Figure 11. Williams’ notation and schematic of ball, trapped

abrasive and substrate—the idealized two-dimensional model [43].

effectively a system aggregate hardness, given by Figure 12. Ridge formed at centre of crater on TiN-coated tool steel

 sample under high load dry sliding conditions [13].
1 1 1
= + , (6)
H Hb Hs abrasion to produce 3-body (rolling) wear, 2-body (grooving)
where Hb is the ball hardness and Hs is the hardness of wear and ridging in that order, transitions would be seen.
the counterface, shedding further light on Williams’ [43] However, it has been shown experimentally by Jones [34]
prediction of 2-body grooving being promoted by a relatively whilst investigating the micro-scale abrasion of silicon nitride
soft substrate and a hard ball, for a given applied load and and sialon–titanium diboride cermets that 2-body wear is not a
slurry concentration. Adachi and Hutchings’ contact severity necessary precursor to ridging on increasing the applied load.
hypothesis was verified by an extensive series of wear mode In this context it would be pertinent at this juncture to highlight
mapping experimental fits with silicon carbide, diamond and that ridge formation (see figure 13) seems critically dependent
alumina abrasives. on the initial contact/surface conditions [44]; formation being
a function of the ease with which slurry is entrained in the
ball/specimen contact.
8.2. Ridge formation The ridging seen by Jones is akin to that seen by Gant
Ridging is a phenomenon which can affect results of micro- in high-stress (macro-) abrasion of tungsten carbide-based
scale abrasion in an often unpredictable and unforeseen way, hardmetals, albeit on a smaller scale [37]; see figure 14.
but has been shown to be possible in a wide variety of materials
subjected to micro-scale abrasion [13, 35, 42, 47]. Allsopp 9. Use of different carrier media
et al [47], working with aluminium, and Shipway [35, 42],
working with soda-lime glass (as mentioned in an earlier The greater part of the canon of papers dealing with micro-
section), showed that the ridging phenomena was principally scale abrasion have to date used water as a carrier medium for
one of slurry starvation and verified this by profilometry; the fine abrasives used. However, some work has been done
showing that ridge height corresponded to the upper end of the using other media. Stack and Mathew [28] reported the use of
slurry particle size spectrum. Shipway also showed that ribbon bio-oils as an abrasive carrier medium in ball cratering. Mild
formation was promoted by increasing load (in the range 0.25 steel, aluminium and copper specimens were used, the balls
to 5 N) and also by decreasing slurry concentration. Shipway being 440 grade stainless steel and the applied loads being in
attributed the slurry concentration effect as being one of the the range 1–5 N. Straight vegetable oil (i.e. no added abrasive)
changes in viscosity. and vegetable oil slurries were used; in the former case wear
To date no systematic studies have been undertaken to increased with load, whereas for the latter the converse was
investigate this; one possible avenue to explore might be to true. Although not reported at the time, it could be the case that
perform a series of controlled experiments at fixed slurry for the slurries they may have been excluded from the contact
concentrations with carrier media of different viscosities. The zone, giving rise to ridging, and hence a drop in wear rate. The
question arises as to whether slurry viscosity is the intrinsic test did, however, discriminate between two bio-oils (vegetable
reason behind the tendency to ridge at a given load, or is oil and sunflower oil) and mineral gear oil; though the latter
it more fundamentally the load borne per abrasive particle? would contain ‘extreme pressure’ (EP) additives, modifying
The latter could explain in a concurrent fashion both ridging its behaviour at high contact stresses.
in terms of applied load and slurry concentration. It may
also be the case that surface tension effects also have a role 10. Ball surface condition
to play in terms of slurry concentration; however, this has
to date never been investigated. It might be expected that Among other factors, the surface characteristics of balls used
given the tendency for increasing applied loads in micro-scale in micro-scale abrasion has been found to play a critical role

10
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

an existing wear track on the ball. Without prior running-in or

pitting, results were found to be variable and were explained
in terms of the entrainment of slurry being impeded; this
was confirmed by displacement measurements; roughening
the ball by whatever means was found to aid repeatability and
give sensible wear coefficients due to its promotion of slurry
entrainment between ball and specimen. Flattening of the ball
track was also found to have no effect on abrasive entrainment.
Ball topography changes were found to affect results to
a greater extent with a soft substrate; it was suggested that
it is easier for abrasive to become embedded in the softer
specimen when loaded by the ball. The duration of the
necessary running-in period has been shown to be influenced
by the nature of the ball material; Shipway [35] studying micro-
scale abrasion of soda-lime glass found that running-in periods
for nylon and polypropylene balls were shorter than those for
their steel or alumina counterparts. The explanation given was
that for the softer polymeric balls, the abrasive particles were
more easily transported into the wear zone; the particles easily
indenting the low moduli polymers, giving a mechanical key to
aid their transport. Gee et al [14], after conducting an extensive
interlaboratory exercise, also concluded that the predominant
factor in determining reproducibility in micro-scale abrasion is
ball surface condition, suggesting that a run-in ball with no flat
was optimum (unlike Allsopp’s earlier study [47] the authors
did not explore ball surface modification by pitting etc).

11. Application to tribo-corrosion: electrochemical

micro-abrasion of surgical alloys, stainless steels and
tungsten carbide-based materials

Several authors have reported the application of the ball crater

test to tribo-corrosion; some studies applying electrochemistry,
Figure 13. Ridge formation reported by Jones [34] on D2 tool steel whilst others have used free-corrosion conditions. The test
with load = 2.5 N. (a) Low magnification of the wear scar and (b) lends itself quite readily for the study of prosthetic materials;
the ridges showing the three-body rolling character of their surface.
the size of the abrasives which are applicable to the test can be
of the same order as bone debris or bone cement fragments [48].
Sinnett-Jones et al [48] working on micro-abrasion–corrosion
of surgical grade F75 CoCrMo alloy found that there were
strong synergistic effects between wear and corrosion ranging
from negative to positive (i.e. beneficial to accelerated volume
loss rates). The extent of synergy appeared to be dependent on
the integrity of the passive films and the repassivation kinetics.
The national Centre for Advanced Tribology at Southampton
(nCATS) under Professor Robert Wood have since done much
to adapt micro-scale abrasion for joint prosthesis materials and
to optimize in situ electrochemistry in conjunction with the
Figure 14. Ridging in dry high-stress abrasion of a WC-25%Co micro-scale abrasion test [49–51]. Although tribo-corrosion
hardmetal [37]; surface illuminated by grazing incidence light. has been widely reported in the context of micro-scale
Relative wheel motion is from left to right; abradant exits at extreme abrasion, little has been done to simulate in vivo conditions
right of scar. seen in joint prostheses. Wood and co-workers have adapted
the basic test in several important ways. Firstly, the recognition
in influencing abradant motion (and hence wear mode) and that as commonly used as an engineering material wear test,
repeatability of test results, Allsopp et al [47]. Allsopp found both abrasive volume fractions and abrasive particle sizes are
that if balls were roughened prior to testing, reproducibility in excess of what would be realistic for in vivo conditions.
could be improved and anomalously low wear coefficients The aforementioned surgical grade F75 CoCrMo alloy has
avoided. Ball surfaces were roughened prior to testing both received further attention since Sinnett-Jones et al [48]; later
by surface pitting produced by aqueous corrosion and using work has used 1.6 µm alumina as the abrasive medium;

11
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

chosen for similar hardness, size and morphology to carbides 11.1. Application to WC-based alloys
which can be plucked out of the alloy and induce micro-
Due to their manifold applications in harsh environments,
abrasion due to being trapped in the interface. The exact
alloys based on tungsten carbide have been the subject of
details of post-operative in vivo lubrication regimes are not
tribochemical micro-abrasion using a commercially available
precisely known, but a mixed lubrication regime is thought
Plint/Phoenix Tribology TE66 (essentially as shown in
to prevail [52]; in this scenario, the hardness, angularity and
figure 3) modified to accommodate a three-electrode cell
size of abrasive are important in determining the severity (platinized titanium counter electrode, modified Ag/AgCl
of contact. Wood and co-workers [48–51] have conducted reference electrode, with the working electrode being the
tests using a carrier medium consisting of 25% bovine specimen exposed to the abrasive slurry). The alloys fall into
serum. The medium contains proteins which denatured during two basic groups; sprayed coatings (usually as high velocity
testing and modified the behaviour in the interface in several oxy-fuel (HVOF)) and sintered or sinter-sinterHIPed materials.
key ways: The main difference between the two generic material types
is that the former commonly have a markedly heterogeneous
• Adsorbed protein (possibly a few tens of nanometres) microstructure. HVOF materials have relatively a large pool
acting as tribo-layers under boundary lubrication. of cobalt or cobalt-rich binder in the microstructure. Thakare
• The denatured proteins mixed with metal debris forming et al [55] found that this had a marked effect on repassivation
conglomerates with micrometre and sub-micrometre kinetics once a sample stopped being abraded; they were
dimensions, which roll between the contacts and far more pronounced than in an equivalent sintered material;
separate/polish the surfaces. binder rich areas repassivated due to Co(OH)2 film formation in
• The adsorption of proteins onto the abrasives may the absence of abrasion. In sintered materials localized binder
influence the contact mechanics/entrainment, and reduces corrosion occurred in crevices between WC grains and was
the abrasivity of the particles. thought to depend on local pH. Earlier papers [55, 56] showed
that changing bulk pH of the abrasive slurry had a marked
In terms of contact conditions the (denatured) protein effect on wear–corrosion synergy; lower wear was found with
had several important ramifications: it was found that at a pH11 (NaOH) based slurry than a neutral one; thought to be
high abrasive volume fractions, the protein enhanced the wear due to inhibition of binder dissolution.
loss due to enhanced particle entrainment, whereas at much
lower abrasive volume fractions, adsorbed and/or denatured 11.2. Stainless steels
proteins reduced the wear loss by acting as boundary lubricant
and/or rolling elements which help to separate the surface, Bello et al [57, 58] subjected three stainless steels (UNS
lubricate the contact and reduced the abrasive damage from the S30403. S31603 and S32760; the principal difference being
particles. in chromium, nickel and molybdenum contents) to micro-
abrasion electrochemical corrosion and found that the ranking
Once fitted, the surface oxide film on CoCrMo hip joints
of the tendency of the three steels agreed with other, more
may be ruptured (depassivation) under mechanical loads and/or
conventional, tests. Moreover, Bello reported one of the few
articulation due to abrasion and/or sliding wear. This will
comparative series of tests to ascertain the relative magnitude
potentially expose the bare metal to the corrosive body fluids, of micro-abrasion corrosion in the three-body regime and that
and result in accelerated metal–ion dissolution. Conversely, caused by two-body abrasion; a 14 fold increase was seen with
the repassivation process will take place immediately the former compared with the latter.
after damage due to the surface oxide (primarily Cr2 O3 )
reformation, which will suppress the corrosion process.
Repassivation kinetics are therefore of prime interest and
12. Dry ball/sliding crater testing
Wood and co-workers [53, 54] have looked into the influence Relatively little has been reported on the use of the ball
(using control studies) of serum (i.e. denatured proteins) cratering test without the use of a carrier medium for the
in electrochemical micro-abrasion. However, repassivation abrasive (typically water). Allsopp and Hutchings [59] studied
kinetics were not found to be significantly influenced by the various PVD coatings on tool steel substrates; TiN, TiCN and
proteins; the main role of the denatured proteins was found to TiAlN. The paper did bring to light several pitfalls which can
be mechanical rather than electrochemical; namely negative arise with the use of dry abrasives in the test; namely scuffing of
tribochemical synergy especially at low abrasive volume the crater edge in tool steel and the occurrence of non-spherical
fractions by forming protein/metal debris conglomerates craters in PVD coatings; the latter making volume loss
that lubricated/separated the surfaces. Despite the lack calculations from depth or diametral measurements inaccurate.
of influence on repassivation kinetics, some influence on Allsopp also found that elevated temperatures (350 ◦ C) could
electrochemistry has been found due to the inclusion of the influence the predominant wear mode, with two-body wear
denatured proteins in the carrier medium; namely at low being favoured at ambient temperature and three-body wear
slurry volume fractions (<0.03 vol), less corrosion induced occurring at 350 ◦ C; reasons for this transition were not given.
current was found, thought to be associated with the denatured Gee and Wicks [60] reported use of ball cratering entirely
proteins or inorgano-metallic conglomerates on the wear without the use of a slurry; i.e. with dry sliding contact between
surfaces [49]. ball and flat specimen. This approach was proven to be a quick

12
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

0.24
16
sp106 0.37 N, 20% SiC
sp106 0.37 N, 20% SiC 14

0.22 12
Friction Coefficient

Displacement, µm
10

8
0.20
6

0.18 2

-2
0.16
0 100 200 300 400
0 100 200 300 400
(a) Cycles (b)
Cycles
16
0.28 -1 -1
sp106 20% SiC, 0.37 N, 0.06 ms 14 sp106 20% SiC, 0.37 N, 0.06 ms

0.26 12
Friction Coefficient

Displacement, µm
10
0.24
8

0.22 6

4
0.20
2

0.18 0

-2
0.16
0 50 100 150 200 250 300 0 50 100 150 200 250 300
(c) Cycles (d) Cycles

Figure 15. Results of measurements of wear displacement and friction during micro-scale abrasion tests on ASP23 tool steel, (a) test A
friction coefficient, (b) test wear displacement, (c) test B friction coefficient and (d) test B wear displacement.

and simple means of friction measurement of coatings. As particles between the surfaces separating the surfaces with a
Stack and Mathew [61] had found when using high loads in simultaneous increase in friction. The friction results shown
conjunction with slurries, Gee and Wicks [60] similarly found in figure 15 are very similar to those reported by Kusano and
that at high contact loads under dry conditions that oxide Hutchings [62].
formation resulted, causing high friction coefficients due to
abrasion with oxide debris. However, extremely high contact 14. Application to curved surfaces
loads served to exclude the oxide debris from the contact area,
thereby lowering wear rates. The vast majority of work reported in the field has concerned
planar specimens, with few authors reporting the application of
13. Friction and displacement measurements the micro-scale abrasion principle to curved surfaces; notable
exceptions being the contribution of Rutherford and Hutchings
Micro-scale abrasion tests can be fitted with measurement [63], with later addenda published by Allsopp et al [64]. The
devices to enable measurements of the wear displacement premise of their methodology is that in the case of a non-planar
(the movement of the steel ball into the sample) and friction specimen, the wear scar volume is that common to a sphere
during the test. Normal behaviour is shown by sample A and the curved specimen surface. In the earlier paper [63],
with a friction coefficient that declines slowly over the an expression was derived for the common volume, for a
period of the test, and a slowly increasing wear displacement cylindrical specimen, by considering area elements parallel to
(figures 15(a) and (b)). Unusually, transient behaviour is the cylinder axis and normal to the line between the cylinder
sometimes observed where the friction coefficient showed axis and the centre of the sphere. The common volume was
periodic increases in friction that corresponded to peaks in the then obtained by integrating the area elements to give the
wear displacement (figures 15(c) and (d)). The magnitude of following expression:
the wear displacement increases (about 4 µm), is the same size
1/2

as the SiC abrasive used in these tests, supporting the concept π(2l)4 r1 (R + r2 ) 1 1
V = + , (7)
that these transients are due to entrapment of additional SiC 64 r2 (R + r1 ) R r2

13
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

where R is the sphere radius, r1 and r2 are the principal radii [4] Walls J M, Hall D D and Sykes D E 1979 Composition-depth
of curvature and 2l is the crater length. profiling and interface analysis of surface coatings using
ball-cratering and the scanning Auger microprobe Surf.
A key assumption [63] was that the area elements were Interface Anal. 1 204–10
assumed to be elliptical, whereas it is a rectangle capped with [5] Bromark M, Gahlin R, Hedenqvist P, Hogmark S,
two circular sections. However, practical work undertaken Hakansson G and Hansson G 1995 Influence of recoating
by Allsopp and co-workers showed the errors involved using on the mechanical and tribological performance of
the original expression (7) were for all practical purposes, TiN-coated HSS Surf. Coat. Technol. 77 481–6
[6] Bromark M, Larsson M, Hedenqvist P and Hogmark S 1997
negligibly small (<0.5% in the practical results obtained). Wear of PVD Ti/TiN multilayer coatings Surf. Coat.
Technol. 90 217–23
[7] Diserens M, Patscheider J and Levy F 1998 Improving the
15. Conclusions properties of titanium nitride by incorporation of silicon
Surf. Coat. Technol. 109 241–6
(1) Within certain parameters, micro-scale abrasion has [8] Allsopp D N 1999 Abrasive wear of bulk materials and hard
shown itself to be a convenient method for assessing wear coatings PhD Thesis University of Cambridge
performance of monolithic materials and of coatings. [9] Rutherford K L and Hutchings I M 1996 A micro-abrasive
wear test, with particular application to coated systems Surf.
(2) The test offers certain advantages over more well- Coat. Technol. 79 231–9
established tribological test methods; namely that it is [10] Rutherford K L, Bull S J, Doyle E D and Hutchings I M 1996
relatively cheap, it requires only small sample areas Laboratory characterisation of the wear behaviour of
and is quick. Since the test produces craters which PVD-coated tool steels and correlation with cutting tool
performance Surf. Coat. Technol. 80 176–80
are of the order of micrometres deep, it is particularly
[11] Rutherford K L and Hutchings I M 1996 Micro-scale abrasive
suited to assessment of thin coatings; in terms of wear testing of PVD coatings on curved substrates Tribol.
abrasion resistance, thickness measurement, profiling and Lett. 2 1–11
revealing internal structure (taper sectioning). [12] Rutherford K L and Hutchings I M 1997 Theory and
(3) Although the principle of a conformal spherical contact is application of a micro-scale abrasive wear test J. Test. Eval.
25 250–60
simple, the measurement of craters thus produced can be [13] Gee M G, Gant A, Hutchings I, Bethke R, Schiffmann K,
non-trivial. A variety of techniques can be used, but it is Van Acker K, Poulat S, Gachon Y and von Stebut J 2001
recommended that the resolution of a particular technique Review of ball cratering or micro-abrasion wear testing of
is verified prior to its application to micro-scale abrasion, coatings NPL Report MATC(A)62, National Physical
Laboratory
particularly since crater diameter measurement errors are
[14] Gee M G et al 2003 Progress towards standardisation of ball
subject to error to the fourth power when calculating cratering Wear 255 1–13
volume loss. [15] Gee M G et al 2005 Results from an interlaboratory exercise to
(4) The test shows the greatest utility when it produces three- validate the micro-scale abrasion test Wear 259 27–35
body wear; the craters thus produced are better defined [16] ISO 26424:2008 Fine ceramics (advanced ceramics, advanced
technical ceramics)—Determination of the abrasion
and there is less scatter in terms of volume loss results resistance of coatings by a micro-scale abrasion test
than with two-body wear. [17] ISO 5725—Part 2: 1994 Accuracy (trueness and precision) of
(5) Certain steps can be taken to promote three-body (i.e. measurement methods and results
rolling) wear; high slurry concentrations, low applied [18] ASTM G99-95a, 2003 ASTM Annual Book of Standards,
Volume 03.02
loads and careful consideration of ball/sample hardnessses
[19] ASTM G133-02, 2003 ASTM Annual Book of Standards,
to prevent slurry embedment and/or two-body wear. Volume 03.02
Ridging and/or slurry starvation are undesirable as the [20] Danzl R, Helmli R and Gee M 2007 Wear crater analysis with
results will invariably not obey Archard’s law. an optical measurement device based on a colour focus
(6) It should be borne in mind that some materials will by their sensor Int. Nanotechnol. Conf. (Vienna, Austria, 2007)
[21] Stachowiak G B and Stachowiak G W 2004 Wear mechanisms
vary nature produce an atypical response to micro-scale in ball-cratering tests with large abrasive particles Wear
abrasion; e.g. glass, which can undergo microfracture, 256 600–7
and tungsten carbide-based hardmetals, in which wear is [22] Bello J O and Wood R J K 2005 Micro-abrasion of filled and
non-uniform in terms of tungsten carbide grain size. unfilled polyamide 11 coatings Wear 258 294–302
[23] Lancaster J K 1969 Abrasive wear of polymers Wear
14 223–39
References [24] Shipway P H and Ngao N K 2003 Microscale abrasive wear of
polymeric materials Wear 255 742–50
[1] Happ W W and Shockley W 1956 Diffusion depths in silicon [25] Buchanan F J and Shipway P H 2002 Microabrasion—a
measured using cylindrical grooves Bull. Am. Phys. Soc. 1 simple method to assess surface degradation of
382 UHMWPE following sterilisation and ageing Biomaterials
[2] McDonald B and Goetsberger A 1962 Measurement of the 23 93–100
depth of diffused layers in silicon by the grooving method [26] Gant A J, Gee M G and May A T 2004 Microabrasion
J. Electrochem. Soc. 109 141–4 of WC–Co hardmetals in corrosive media Wear
[3] Thompson V, Hintermann H E and Chollet L 1979 The 256 954–62
determination of composition depth profiles using spherical [27] Shipway P H and Howell L 2005 Microscale
erosion and scanning Auger electron spectroscopy Surf. abrasion–corrosion behaviour of WC–Co hardmetals and
Technol. 8 421–8 HVOF sprayed coatings Wear 258 303–12

14
J. Phys. D: Appl. Phys. 44 (2011) 073001 Topical Review

[28] Stack M M and Mathew M 2004 Some comments on [48] Sinnett-Jones P E, Wharton J A and Wood R J K 2005
micro-abrasion interactions for pure metals in bio-oils Micro-abrasion–corrosion of a CoCrMo alloy in
J. Synth. Lubr. 212 105–18 simulated artificial hip joint environments Wear
[29] Stack M M and Mathew M T 2005 Transitions in 259 898–909
microabrasion mechanisms for WC–Co (HVOF) coated [49] Sun D, Wharton J A and Wood R J K 2009
steel Proc. IMechE Part J: J. Eng. Tribol. 219 49–57 Micro-abrasion–corrosion of cast CoCrMo—effects of
[30] Chen H, Xu C, Zhou Q, Hutchings I M, Shipway P H and micron and sub-micron sized abrasives Wear 267 52–60
Liu J 2005 Micro-scale abrasive wear behaviour of HVOF [50] Sun D, Wharton J A and Wood R J K 2009 Micro-abrasion
sprayed and laser-remelted conventional and nanostructured mechanisms of cast CoCrMo in simulated body fluids Wear
WC–Co coatings Wear 258 333–8 267 1845–55
[31] Gant A J, Gee M G and Roebuck R 2005 Rotating wheel [51] Thakare M R, Wharton J A, Wood R J K and Menger C 2009
abrasion of WC/Co hardmetals Wear 258 178–88 Investigation of micro-scale abrasion–corrosion of
[32] Hansen N and Ralph B 1982 The strain and grain size WC-based sintered hardmetal and sprayed coating using
dependence of the flow stress of copper Acta Metall. in situ electrochemical current-noise measurements Wear
30 411–7 267 1967–77
[33] Stack M M and Mathew M T 2004 Mapping the [52] Wang A, Essner A, Polineni V K, Stark C and Dumbleton J H
micro-abrasion resistance of WC/Co based coatings in 1998 Lubrication and wear of ultra-high molecular
aqueous conditions Surf. Coat. Technol. 183 337–46 weight polyethylene in total joint replacement Tribol. Int.
[34] Jones A H 2005 Micro-scale abrasive wear of silicon nitride, 31 17–33
sialon–TiB2 composites and D2 tool steel using a multiple [53] Sun D, Wharton J A, Wood R J K, Ma L and Rainforth W M
load method Wear 258 942–52 2009 Microabrasion–corrosion of cast CoCrMo alloy in
[35] Shipway P H 1999 The role of test conditions on the simulated body fluids Tribol. Int. 42 99–110
micro-abrasive wear behaviour of soda-lime glass Wear [54] Sun D, Wharton J A and Wood R J K 2009 Abrasive size and
233–235 191–9 concentration effects on the tribo-corrosion of cast CoCrMo
[36] Dube N B and Hutchings I M 1999 Influence of particle alloy in simulated body fluids Tribol. Int. 42 1595–604
fracture in the high-stress and low-stress abrasive wear of [55] Thakare M R, Wharton J A, Wood R J K and Menger C 2007
steel Wear 233–235 246–56 Exposure effects of alkaline drilling fluid on the microscale
[37] Gant A J and Gee M G 2001 Wear of tungsten carbide–cobalt abrasion–corrosion of WC-based hardmetals Wear
hardmetals and hot isostatically pressed high speed steels 263 125–36
under dry abrasive conditions Wear 251 908–15 [56] Thakare M R, Wharton J A, Wood R J K and Menger C 2008
[38] Preece C M (ed) 1979 Treatise on Materials Science and Exposure effects of strong alkaline conditions on the
Technology: Erosion (New: York: Academic) microscale abrasion–corrosion of D-gun sprayed
[39] Batista J C A, Joseph M C, Godoy C and Matthews A 2002 WC–10Co–4Cr coating Tribol. Int. 41 629–39
Micro-abrasion wear testing of PVD TiN coatings on [57] Bello J O, Wood R J K and Wharton J A 2005
untreated and plasma nitrided AISI H13 steel Wear 249 Micro-abrasion–corrosion of UNSS31603L stainless steel
971–9 Proc. IRG-OECD on Wear of Engineering Materials
[40] Batista J C A, Godoy C and Matthews A 2002 Micro-scale (Uppsala, Sweden)
abrasive wear testing of duplex and non-duplex [58] Bello J O, Wood R J K and Wharton J A 2007 Synergistic
(single-layered) PVD (Ti,Al)N, TiN and Cr–N coatings effects of micro-abrasion–corrosion of UNS S30403,
Tribol. Int. 35 363–72 S31603 and S32760 stainless steels Wear 263 149–59
[41] Hedenqvist P, Bromark M, Olsson M, Hogmark S and [59] Allsopp D N and Hutchings I M 2001 Micro-scale abrasion
Bergmann E 1994 Mechanical and tribological and scratch response of PVD coatings at elevated
characterisation of low temperature deposited PVD TiN temperatures Wear 251 1308–14
coatings Surf. Coat. Technol. 63 115–22 [60] Gee M G and Wicks M J 2000 Ball crater testing for the
[42] Shipway P H and Hodge C J B 2000 Microabrasion measurement of the unlubricated sliding wear of
of glass—the critical role of ridge formation Wear 237 90–7 wear-resistant coatings Surf. Coat. Technol.
[43] Williams J A and Hyncica A M 1992 Mechanisms of abrasive 133–134 376–82
wear in lubricated contacts Wear 152 57–74 [61] Stack M M and Mathew M 2003 Micro-abrasion transitions of
[44] Trezona R I, Allsopp D N and Hutchings I M 1999 Transitions metallic materials Wear 255 14–22
between two-body and three-body abrasive wear: influence [62] Kusano Y and Hutchings I M 2005 Sources of variability in
of test conditions in the microscale abrasive wear test Wear the free-ball micro-scale abrasion test Wear 258 313–317
225–229 205–14 [63] Rutherford K L and Hutchings I M 1996 Micro-scale abrasive
[45] Hutchings I M 1992 Tribology: Wear and Friction of wear resistance of PVD coatings on curved substrates
Engineering Materials (London: Edward Arnold) Tribol. Lett. 2 1–11
[46] Adachi K and Hutchings I M 2003 Wear-mode mapping for [64] Allsopp D N, Hutchings I M and Trezona R I 1999 Comments
the micro-scale abrasion test Wear 255 23–9 on Micro-scale abrasive wear testing of PVD coatings on
[47] Allsopp D N, Trezona R I and Hutchings I M 1998 The effects curved substrates [by K L Rutherford and I M Hutchings]
of ball surface condition in the micro-scale abrasive wear Tribol. Lett. 7 229–31
test Tribol. Lett. 5 259–64

15