This article was downloaded by: [Stony Brook University]

On: 30 October 2014, At: 19:29

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK

Tribology Transactions
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/utrb20

New Test for Mild Lubricated Wear in Rolling-Sliding

Contacts
a a
Jingyun Fan & Hugh Spikes
a
Tribology Group Department of Mechanical Engineering Imperial College London , London,
SW7 2AZ, UK
Published online: 25 Mar 2008.

To cite this article: Jingyun Fan & Hugh Spikes (2007) New Test for Mild Lubricated Wear in Rolling-Sliding Contacts, Tribology
Transactions, 50:2, 145-153, DOI: 10.1080/10402000701255476

To link to this article: http://dx.doi.org/10.1080/10402000701255476

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Tribology Transactions, 50: 145-153, 2007
Copyright
C Society of Tribologists and Lubrication Engineers

ISSN: 1040-2004 print / 1547-357X online

DOI: 10.1080/10402000701255476

New Test for Mild Lubricated Wear

in Rolling-Sliding Contacts
JINGYUN FAN and HUGH SPIKES
Tribology Group Department of Mechanical Engineering
Imperial College London
London SW7 2AZ, UK

A new test method for measuring and studying mild, lu- changes in lubricant formulation, including a general reduction in
antiwear additive concentration, and it is becoming increasingly
Downloaded by [Stony Brook University] at 19:29 30 October 2014

bricated wear is described. The test combines three important

characteristics, (i) a mixed rolling-sliding contact to minimize difficult to find an acceptable compromise between compositional
contact geometry changes during a test; (ii) contra-rotating test constraints, oil longevity, and wear performance.
specimen, to produce slide-roll ratios of greater than two, so as This paper describes a new bench test to study mild, lubricated
wear. The test is based on an existing friction test apparatus, the
to enable high sliding speeds to be obtained while maintaining
minitraction machine (MTM).
low entrainment speeds; and (iii) ICP-AES metal analysis to
This paper will discuss some limitations of existing wear tests
monitor very low levels of wear metals. The new wear test is
and outline the main methods currently used to monitor wear
applied to study the influence of ZDDP concentration on wear. in bench tests. The principles of the new wear test method are
then described, including details of the spectroscopic oil analysis
KEY WORDS method employed to monitor wear. Finally, some wear measure-
Wear Test; Mild Wear; ZDDP; ICP-AES; Slide-Roll Ratio ments are described.

SOME LIMITATIONS OF EXISTING WEAR TESTS

INTRODUCTION
Before outlining the new test method, it is important to exam-
Numerous tests for measuring wear are described in the liter-
ine some of the main limitations of most existing bench tests with
ature. Many have been developed for specific applications, while
respect to the study of mild wear.
others, such as the four-ball (D4172-94 (1)), block on ring (D2714-
94 (2)), and V-block tests (D2670-95 (3)), are standard bench tests Pure Sliding versus Mixed Sliding-Rolling
employed by a large number of laboratories to compare and quan-
The first limitation is that almost all existing tests, such as the
tify the wear properties of lubricants and materials. In 1966, the
four-ball (D4172-94 (1)), block-on-ring (D2714-94 (2)), and re-
ASLE identified over 200 types of wear test in use at the time
ciprocating testers (D6079-04 (6), D5707-98e1 (7)), employ pure
(Benzing, et al. (4)). A useful review of test methods for measur-
sliding contact conditions, with one surface stationary and the
ing wear is also provided by Brown (5).
other moving, rather than mixed rolling-sliding where both sur-
Among all of these tests, there is arguably no satisfactory bench
faces move relative to the contact at different but finite sliding
test to measure the ability of formulated lubricants to control wear
speeds. Pure sliding test conditions are generally easier to set up
in mild, sliding wear conditions, where the Archard wear coeffi-
than mixed rolling-sliding ones, both because it is simpler to con-
cient is less than ca. 10−5 . Most existing bench tests are too severe
trol the speed of one surface than of two and because friction is
to properly represent the behavior of lubricants in such conditions,
often monitored during wear testing, and it is more straightforward
while if the contact conditions are made less severe, too little wear
to measure friction force on a stationary body than a moving one.
occurs within a reasonable test time to be accurately measured.
More fundamentally, however, as will be discussed in the next sec-
This is an unfortunate limitation, since engine and transmission
tion, mild wear is most commonly determined by measuring the ge-
lubricants often provide very mild wear and bench tests are needed
ometry of wear scar produced, and this is more accurate if the wear
to help develop and optimize these lubricants. The need for a
on the surface of interest is localized within a well-defined scar on a
suitable, mild, lubricated wear test is especially important at the
stationary surface. In mixed rolling-sliding conditions, wear is dis-
present time because of the quest to reduce the sulphur, ash, and
tributed along a wear track, producing far less wear depth at any
phosphorus levels of engine lubricants. This is leading to major
one location, which is correspondingly more difficult to measure.
An unfortunate consequence of pure sliding is that wear of
Presented at the STLE Annual Meeting the stationary surface usually results in a significant change of
Las Vegas, Nevada
contact geometry during a test. This, in turn, means that the contact
May 15-19, 2005
Manuscript approved October 5, 2006 pressure and temperature vary within a test and can reach quite
Review led by Cyril Migdal different values from test to test depending on the wear properties

145
 146 J. FAN AND H. SPIKES

of the lubricant being studied. Often, wear in a pure sliding contact (which is half the sliding speed) and thus a thicker hydrodynamic
can produce so much local conformity that the contact regime or elastohydrodynamic film. The latter may actually result in a
changes from boundary lubrication at the start of the test to almost reduction in wear rate at increased speed. As will be described
full film lubrication at its end. below, in the current study this problem is resolved by moving
To overcome this limitation, a mixed rolling-sliding motion is the two rubbing surfaces in opposite directions with respect to the
adopted in the new test so as to distribute the wear on both contact- contact.
ing bodies and thus minimize contact condition variation during
wear.
METHODS OF MEASURING WEAR
Monitoring Wear Ruff (9) has identified four main categories of methods of mea-
Another characteristic of most existing wear tests is that they suring wear: mass loss measurement, linear measurement, area
provide only one measure of wear, at the end of the test. This is measurement, and volume measurement. Measurement of mass
because wear is generally calculated either from the geometry of loss resulting from wear is straightforward in principle but is only
the wear scar or from weight loss, and both of these measurements practicable for quite large wear rates and for small, and thus low-
are difficult to make without removing the test sample from the weight, components. Other drawbacks are that measurement is
wear tester. A few wear tests exist that are able to monitor wear only possible at the end of a test and great care must be taken
Downloaded by [Stony Brook University] at 19:29 30 October 2014

continuously or regularly throughout the test, notably pin-on-disc to clean the test specimen before measurement. For very small
testers where wear is determined from the vertical displacement amounts of wear, the weight of deposits from additives or oxi-
of the pin. Generally, however, these are confined to the study dized oil on the specimen surface can overwhelm the weight lost
of quite high wear rates, such as during abrasive wear, since it is by wear.
very difficult to measure submicron displacements in macroscale Linear measures of wear involve either the measurement of
equipment. displacement of the holder of a test specimen as wear takes place,
Although a single measure of wear at the end of a test is ad- e.g., of the pin in a pin-on-disc rig apparatus, or measurement of
equate for comparing the performance of different lubricant for- the change in diameter of a cylindrical shaft, disc, or bush after
mulations, it is less satisfactory for research purposes since it does wear. As with mass loss, this approach is not suited to the study of
not show whether the wear rate measured is the steady-state value mild wear, where only submicron depths of material are removed.
or whether it encompasses significant variations in wear rate dur- In most existing lubricated wear tests, including the 4-ball, V-
ing the test, due to factors such as running-in or time-dependent block, block-on-ring, and HFRR, wear is determined by measur-
additive reactions. One solution is to conduct repeat wear tests ing the area of the contact scar on the stationary surface at the end
of differing duration, but this is time consuming and expensive. of the test. The volume of material removed by wear can then be
Clearly, a means of monitoring mild wear continuously or at least calculated, if required, from the original geometry of the surface.
intermittently during a wear test is desirable. This approach is able to provide a measure of very small amounts
of wear. Its main disadvantages are (i) because it involves a pure
Accelerated Wear sliding contact, the wear scar on the stationary specimen cannot
A third, general problem in the study of mild wear is that wear easily be monitored throughout a test (since it is buried within
is often so slow as to be very difficult to measure accurately within the contact) and (ii) as wear progresses during a test, the contact
a reasonable test period. Most forms of wear obey, to a first ap- geometry and thus pressure often vary considerably.
proximation, the Archard wear equation, where the volume wear Ruff’s fourth category, volume measurement, involves the use
rate per unit sliding distance, RL, is proportional to the applied of 3-D surface profilometry to assess the volume of material re-
load, W, and inversely proportional to the hardness; i.e., moved from the wear scar. This is able to measure very small
amounts of wear and is potentially a powerful technique for the
W
RL = kA [1] study of mild wear in both pure sliding and rolling-sliding contact
H
conditions. Recently, Gahlin has shown how topographic image
where kA is the non-dimensional Archard wear coefficient subtraction can be used to quantify very small amounts of wear,
(Williams (8)). For constant sliding speed and hardness, this means of less than 50 nm thickness (Gahlin and Jacobson (10)). In princi-
that the wear volume, Q, in a given test is proportional to the speed, ple this approach might even be used to monitor wear of a surface
the applied load, and the test time; i.e., in situ within a test rig, as long as the influence of the supernatant
Q = k1 us Wt [2] lubricant on the measurement could be negated.
In addition to the above four approaches, another method of
where us is the sliding speed, t is the test time, and k1 = kA/H is measuring wear that has gained popularity in recent years is the
the dimensional wear coefficient. For mild wear conditions, the use of thin or surface layer activation (TLA or SLA). In this, a
rate constant, k1 , is very small, so the wear volume is correspond- small region of the surface of a machine component of interest
ingly small. From Eq. [2], it can be seen that to increase the wear is irradiated using high-energy charged particles to generate trace
volume within a given test duration, either the applied load or the quantities of metallic radioisotopes. The wear of this surface region
sliding speed must be increased. However, increasing the load can is then monitored either by measuring the radioactivity of the
lead to a change of wear mechanism, promoting fatigue-type wear lubricant (and thus the amount of radioactive material which has
processes, while increasing the sliding speed in pure sliding condi- been worn from the surface) or from the decrease in radioactivity
tions leads to a proportionate increase in the entrainment speed of the surface, measured using a gamma ray detector positioned
 Mild Lubricated Wear in Rolling-Sliding Contacts 147

close to the activated surface (Conlon (11); Blatchley (12)). Both (ii) small amount of test lubricant (important for wear measure-
provide a sensitive wear monitoring technique. ment)
Since it measures the wear of only the pretreated surface, the (iii) no other metallic rubbing parts in the test chamber (impor-
TLA approach is widely used to discriminate and quantify the tant for wear measurement).
wear of a specific component in a complex lubricated system from
wear taking place elsewhere; for example, of a specific location on Normally the MTM is operated in mixed sliding-rolling, with a
the piston liners in an engine. Because of its expense and probably slide-roll ratio in the range 0.1 to 0.5. This slide-roll ratio (SRR) is
also because the use of radioactive materials, however tiny their defined as the ratio of sliding speed, us = |u D −u B|, to mean rolling
activity, raises health and safety issues, surface activation is rarely, or entrainment speed, U = (u D + u B)/2, where u D and u B are the
if ever, used to monitor wear in bench tests. Another possible surface speeds of the disc and ball, respectively. For slide-roll ratios
disadvantage of the TLA method for fundamental research is that between 0 and 2, the disc and the ball rotate in the same direction
the isotopes formed during activation and subsequently traced with respect to the contact, but at different surface speeds, ranging
are no longer the original iron but other elements such as cobalt, from pure rolling (surfaces have same speed) at SRR = 0 to pure
formed from iron by the gamma irradiation. It is not clear whether, sliding (one surface stationary) at SRR = 2. In this range, the
in very mild, adhesive-corrosive wear conditions, the rate of loss sliding speed can never be more than twice the entrainment speed,
of these elements from the surface will be representative of the and any increase in the sliding speed produces a proportionate
Downloaded by [Stony Brook University] at 19:29 30 October 2014

rate of loss of the bulk ferrous surface. increase in the entrainment speed, U. However, according to the
elastohydrodynamic (EHD) theory, the film thickness in a ball on
New Wear Test flat contact increases with U 0.67 (Dowson and Hamrock (13)), so
any increase in the sliding speed will produce a correspondingly
The wear test described in the current paper seeks to address
thicker, separating oil film and thus reduce the asperity contact.
the main problems outlined in the above two sections; i.e., how to
This means that it is impractical to accelerate the rate of wear by
obtain measurable levels of wear in a rolling-sliding test and how
simply increasing the sliding speed.
to monitor the very small amount of wear levels produced in mild
One way around this problem is to employ very rough surfaces,
wear.
so that the contact operates in mixed lubrication even though the
The test uses a minitraction machine (MTM) (PCS Instru-
entrainment speed is high. However, this tends to lead to the test
ments) with a ball-on-disc configuration, as shown schematically
becoming a running-in rather than a wear test and it is also difficult
in Fig. 1. Both ball and disc are independently driven to obtain any
to produce rough test specimens having the same topography from
desired combination of rolling and sliding and the ball drive shaft
batch to batch.
is tilted so as to minimize spin in the contact. A load cell attached to
In the wear test developed in the current study, this problem is
the bearing housing of the ball shaft is able to measure the friction
overcome by operating the MTM with the ball and the disc rotat-
force. The disc, and thus the contact, is fully immersed in a lu-
ing in opposite directions; i.e., contra-rotation. If the ball and the
bricant, which is contained within a temperature-controlled bath.
disc surface move in opposite directions with respect to the con-
The test apparatus is computer controlled and programmable so
tact, u B and u D have opposite signs, so the sliding speed, |u D − u B|,
that the tests can be carried out in which temperature is raised
becomes large, while the entrainment speed, (u D +u B)/2, becomes
and lowered and motion begun and halted automatically in any
small, to produce a slide-roll ratio that can be much greater than
desired fashion.
2. This means that it is possible to obtain a high sliding speed in
From the point of view of the required wear test, three impor-
combination with a very low entrainment speed. As far as the au-
tant features of the apparatus are
thors can ascertain, this is the first time that this principle has been
(i) independent control of both ball and disc to obtain any mixed employed in wear testing. Blok reports the use of contra-rotation
sliding-rolling combination to decouple entrainment from the sliding speed and thus study the
scuffing properties of lubricants independently of their viscosity
(Blok (14)). More recently, Wedeven has promoted the impor-
tance of decoupling entrainment and the sliding speed to study
in-contact behavior, especially in terms of scuffing and adhesive
wear (Wedeven (15)). However, he achieves this by varying the
angle between the rolling directions of the two bodies in contact
rather than the velocities of the two surfaces. As will be shown later
in this paper, contra-rotation provides a means of obtaining large
sliding distances (and thus high and easily measurable wear) in
very thin-film lubricated contacts within relatively short test times.
Three methods of measuring the wear are applied in the test.
The main method is inductively coupled plasma atomic emission
spectroscopy (ICP-AES). The method is able to measure the con-
centration of most metals, including iron, down to the parts per
billion (ppb) level, independent of the chemical form of the metal
Fig. 1—Schematic diagram of MTM test rig. atoms. In the current study, small lubricant samples were simply
 148 J. FAN AND H. SPIKES

withdrawn from the lubricant bath and analyzed for iron content. TABLE 1—SPEED COMBINATIONS USED
Details of the method employed are described later in this paper. Ball Speed Disc Speed U (u D + u B)/2 us |u D − u B|
The ICP method is very widely used in the lubricants in- u B (m/s) u D (m/s) (m/s) (m/s) SRR
dustry to measure levels of wear in operating machinery, for
−0.075 0.175 0.05 0.25 5.0
troubleshooting and condition monitoring. It appears, however,
to be very rarely used to monitor wear in bench tests. After con-
siderable search, the authors could find no such application re- minute, and negligible cooling of the oil bath took place during
ported, although one was found using atomic absorption spec- this time. At the end of the test, two final samples of the lubricant
troscopy (Eyre and Fitter (16)). This may be because ICP-AES were extracted.
equipment is relatively expensive or because, unlike TLA, it can- In this study, several different speed conditions were explored.
not distinguish between metal originating from different compo- The aim was to identify the highest sliding speed that gave steady,
nents and so has to be applied in a test where there are no other mild wear (i.e., no scuffing) for all lubricants studied while main-
possible sources of the wear metal of interest. Fortunately, in the taining low entrainment speed.
MTM there is only one rubbing metallic contact within the lubri- For the applied load, geometry, the materials, and the lubricants
cant bath, and the lubricant bath has a close-fitting lid that limits used in this study, the theoretical central elastohydrodynamic film
possible external contamination. A second feature of the MTM thickness, h, reduces to:
Downloaded by [Stony Brook University] at 19:29 30 October 2014

that makes it suited to ICP-AES is that the volume of lubricant

h ≈ 1900(Uη)0.67 nm [3]
used is quite small (ca. 50 g). This means that the wear mate-
rial will remain in reasonably concentrated form in the lubricant where U is the entrainment speed in m/s and η is the dynamic vis-
rather than being dispersed in a large volume of fluid. An ad- cosity in Pa·s (assuming a lubricant pressure viscosity coefficient
ditional problem that may have limited the use of ICP-AES in of 15 GPa−1 ). To ensure mixed lubrication conditions, an entrain-
the past is that of inconsistent sampling, since the wear particles ment speed of 0.05 m/s was employed for most of the tests, to
may not evenly distribute in the oil but may sink to the bottom of provide a theoretical initial lambda ratio (ratio of EHD film thick-
the rig. In the MTM it is relatively easy to agitate the oil during ness to composite root mean square roughness) of approximately
sampling. 0.5. The speed conditions used to obtain the results reported in
In addition to ICP-AES, at the end of each wear test, the ball this paper are summarized in Table 1. It can be seen that contra-
and disc were removed, cleaned, and the topography of the wear rotation of the ball and disc enabled a sliding speed of 0.25 m/s to
scars measured using a stylus profilometer. Profiles of the wear scar be combined with the required entrainment speed of 0.05 m/s.
were obtained at several spaced locations around the wear tracks
on both ball and disc, and these were analyzed and averaged to DETAILS OF ICP-AES METHOD
calculate the mean volume of material removed. The ICP-AES is very widely used to measure the concentra-
tion of metals in solution. Its key advantages are (i) it can measure
DETAILS OF WEAR TEST METHOD very low concentrations (down to ppb levels) and (ii) it has a multi-
In the current work, AISI 52100 steel balls (19.0 mm diameter) element capability, being able to measure up to 40 metals simul-
and discs were employed, both having Young’s modulus 210 GPa taneously. In ICP-AES, a small quantity of the metal-containing
and Poisson’s ratio 0.3. The balls had root mean square roughness solution is injected into an argon plasma to reach a temperature of
(Rq ) 10 nm and hardness 800 VPN while the discs had Rq 11 nm ca. 10,000 K. The intensity of light emitted at characteristic wave-
and hardness of 750 VPN. In all tests, the applied load was 30 N, lengths by the metal atoms at this temperature is then used to
which, according to the elastic contact theory, gave an initial Hertz quantify the concentration of the metal present.
contact diameter of 246 µm and a maximum Hertz pressure of 0.95 ICP-AES is routinely used in tribology to measure the metal
GPa. content of lubricants in operating machinery, both for routine con-
The test procedure was as follows. The test apparatus was as- dition monitoring and in post-failure analysis. There are essen-
sembled and a weighed sample of lubricant sufficient to immerse tially two techniques: direct and indirect. In direct methods, the oil
the disc (normally ca. 50 g) was added to the test chamber. Two sample is diluted with an organic solvent such as methyl-iso-butyl-
separate 0.5 g samples of the lubricant were then withdrawn for ketone (MIBK), white spirit, or xylene, and then injected into ICP-
subsequent ICP-AES analysis. The temperature was then raised AES without the other treatment. In indirect methods, the metal
to its set value (80◦ C in all tests in this study) while rotating the ball is transferred from the oil to an aqueous phase and the latter is an-
and the disc without the applied load. Once the temperature had alyzed. There are various ways of accomplishing this transfer, with
stabilized, the load was applied and the ball and the disc driven different levels of severity depending on how strongly the metal is
to produce the set sliding-rolling conditions. The test was then held in the oil phase and its particle size. These include extraction
continued typically for four hours, with the motion being halted into acid, microwave heating, high-pressure bomb digestion, and
periodically to remove the lubricant samples for analysis. To do even dry-ashing of the lubricant followed by a solution of the ash
this, the ball and the disc were halted, nitrogen bubbled into the oil in water.
sample to ensure agitation, and thus dispersal of any particulate A direct measurement from the oil phase is simple and rapid
wear debris, and two 0.5 g samples of oil were removed using a sy- but is generally less sensitive to very low metal levels and also
ringe. A fresh syringe was used for each sample. The nitrogen flow more difficult to calibrate. It is also not suitable for analyzing oils
rate was ca. 2 cm3 /s. These sampling periods lasted typically one containing large wear particles. Transference of metal from the oil
 Mild Lubricated Wear in Rolling-Sliding Contacts 149

TABLE 2—VALIDATION OF ICP DIGESTION AND ANALYSIS METHOD

Different Concentration
Iron Solutions (ppm) ICP Results (ppm) Recovery (%)

0.989 1.08 109

0.989 0.93 94
9.880 10.7 108
9.880 10.2 103

to the aqueous phase is time consuming and carries more risk of

loss of material or contamination during digestion but results in
a product that can be very reliably analyzed with ICP-AES down
to extremely low levels.
In the current work, several indirect methods were investi- Fig. 2—Repeat tests for 0.032% wt. P concentration ZDDP solution at
slide-roll ratio = 5 and 80◦ C.
gated. Initially, samples were simply heated for three hours in
dilute nitric acid at 60◦ C in an ultrasonic bath. This was found to
Downloaded by [Stony Brook University] at 19:29 30 October 2014

be effective in extracting iron from base oil but much less so from each time stage is shown separately. This indicates that most of
solutions containing ZDDP and/or the dispersant. In these cases the variability is between tests rather than between samples in the
the iron was presumably retained in the oil phase by the additive, same test, suggesting that the ferrous wear material was evenly
so a technique was required which destroyed the ability of these dispersed in the test oil. This was seen for all oils tested.
additives to chelate iron. The final technique adopted was to use a Figure 3 compares the ICP-AES measurements for tests on the
microwave digester (Multiwave 3000, Anton Paar, Graz, Austria). base oil at two different temperatures. Both tests are at a slide-roll
Each sample was mixed with 6 ml of Analar 60% HNO3 in a PTFE ratio of 5. The wear rate increases with temperature. This is as ex-
tube and heated with a power of 1600W for 30 minutes. The resul- pected since as the temperature is increased, the lubricant viscos-
tant aqueous solutions were then cooled and the samples injected ity and thus the elastohydrodynamic film thickness will decrease,
directly into the ICP-AES. leading to a lower lambda ratio. According to the EHD theory,
To test the accuracy of the above method, solutions of iron the theoretical values are ≈7 nm and ≈5 nm at 80 and 100◦ C,
organmetallic standard (1.000 mg/g in 75 cSt hydrocarbon oil were respectively. The true lubricant film thickness cannot be reliably
used to prepare 1 and 10 ppm iron using the test base oil and calculated from theory, however, both because of the very high
digested as described. The results are summarized in Table 2. slide-roll ratio and because the contact is in the mixed rather than
the full-film lubrication regime. Both the tests show quite a low
Test Materials wear rate for the first 30 minutes, followed by an increase in wear.
This may reflect the initial protection of an oxide film or the effect
In the work described in this paper, one API Group II base oil
of changes in topography that result in more solid-solid contact
with the viscometric properties listed in Table 3 was employed.
after some time of rubbing.
A single zinc dialkyldithiophosphate (ZDDP) antiwear addi-
Figure 4 compares the wear rate for six ZDDP concentrations
tive was studied (a commercial C3 C6 secondary ZDDP).
ranging from 0.005% P to 0.05% P. There is a clear increase in the
Additionally, two CEC reference oils, one a “high wear” ref-
wear rate as the ZDDP concentration is reduced.
erence oil, RL 142, and the other a “low wear” reference oil, RL
At very low ZDDP concentrations, the rate of wear is actu-
194, were studied.
ally greater than that seen for the mineral oil alone. This was
RESULTS
ICP-AES Results
Figure 2 shows the ICP-AES wear measurements from two
different tests with ZDDP in the base oil at 0.02% P concentra-
tion. Both tests were conducted at a slide-roll ratio of 5 and 80◦ C
and lasted for four hours. It can be seen that there is very good
repeatability, with a wear rate of ≈6 ppm/hour. In this figure, the
iron content of both of the samples withdrawn from the MTM at

TABLE 3—BASE OIL VISCOSITY

Temperature (◦ C) Dynamic Viscosity (cP)

40 16.41
60 8.600
80 5.168
100 3.421 Fig. 3—Effect of temperature on wear behavior of group 2 base oil at
slide-roll ratio = 5.
 150 J. FAN AND H. SPIKES

TABLE 4—INCREASE IN FE CONTENT OF OIL AFTER FOUR-HOUR UN-

LOADED TEST

Fe Content ppm

Base oil 0.06

ZDDP, 0.005% P 0.17
ZDDP, 0.02% P 0.42
ZDDP, 0.05% p 0.56
RL 142 0.54
RL 194 0.34

Stylus Profilometry
After the four hour wear tests, the ball and the disc specimens
Fig. 4—Influence of ZDDP concentration on wear at slide-roll ratio = 5 were cleaned in toluene and the surface topography profiles ob-
and 80◦ C.
tained across the wear tracks, normal to the sliding direction.
Downloaded by [Stony Brook University] at 19:29 30 October 2014

Figures 6 to 10 show measured profiles for the base oil, 0.005%,

0.01% and 0.03% phosphorus ZDDP solutions. The amount of
repeatable and is believed to originate from the ZDDP initially
wear was determined as the area between the original profiles
solubilizing iron oxide film, as suggested by Martin et al. (17), thus
(the thinner solid lines in the figures) and the surface profiles,
resulting in surfaces more prone to adhesion with the counterface.
multiplied by the length of the wear tracks around the ball and
This effect probably occurs at all ZDDP concentrations but for
the disc. The results will be discussed in the section on the analysis
concentrations above 0.01% wt. P is nullified by the formation of
of wear scar profiles.
a protective phosphate glass antiwear coating.
Figure 5 compares the wear of two CEC reference oils, one re- DISCUSSION
garded as a high wear and the other as a low wear oil. The measured
Analysis of ICP-AES Wear Measurements
wear was approximately 25% greater for high-wear reference oil
(RL 142) than the low-wear one (RL 194). In the ICP-AES results described above, the wear is described
One important issue that had to be determined was whether only in terms of Fe content in the lubricant. While this is adequate
any of the measured iron content above was due to corrosion for comparing the wear performance of lubricants, it is of interest
from the overall ball and disc steel surfaces rather than wear from to relate this to the actual dimensional and dimensionless wear
the rubbed tracks. Even if quite small, such corrosion might be a coefficients.
significant proportion of the whole because of the relatively large For most of the lubricants tested, except for the higher ZDDP
surface area involved. To test this, a series of four-hour MTM concentrations, the measured wear increased approximately lin-
tests were carried out at 80◦ C with the ball and the disc surfaces early with rubbing time, indicative of a constant wear rate. From
rotating at the normal test speeds but with zero applied load (and Eq. [2], the dimensional wear coefficient, k1 is given by:
the ball and disc just non-contacting). The results are listed in Q
Table 4. k1 = [4]
us Wt
The amount of corrosion increased with ZDDP concentration
The wear volume is given in terms of the Fe concentration from
but for all fluids tested was negligible compared to the Fe content
the ICP analysis by:
resulting from rubbing.
ppm.m
Q= · 10−6 m3 [5]
0.975ρ

where ppm is the measured Fe content (in ppm), m is the mass of

test oil in the MTM (initially 0.050 kg), and ρ the density of iron

Fig. 5—Comparison of wear behavior of two CEC reference oils at slide- Fig. 6—Surface profiles of wear scar transverse to the sliding direction
roll ratio = 5 and 80◦ C. on disc for base oil at 80 and 100◦ C (slide-roll ratio = 5).
 Mild Lubricated Wear in Rolling-Sliding Contacts 151

Fig. 7—Surface profiles of wear scar transverse to the sliding direction Fig. 9—Surface profiles of wear scar transverse to the sliding direction
on ball for base oil at 80 and 100◦ C (slide-roll ratio = 5). on disc and ball, respectively, for 0.01% wt. P ZDDP solution at
80◦ C (slide-roll ratio = 5).

(7860 kg/m3 ). The divisor 0.975 is present to allow for the fact that
AISI 52100 steel is only ≈97.5% Fe. tact width in the sliding direction of 0.13 mm. The mean pressure
Assuming that k1 has the same value on the ball and the disc, of the contact of length 0.75 mm and width 0.13 mm is then 0.31
this is given by: GPa. This is clearly a very approximate estimate and a numerically
Downloaded by [Stony Brook University] at 19:29 30 October 2014

ppm . m based contact analysis would yield a more precise value.

k1 = .10−6 m3 /J [6]
0.975ρus Wt It is possible to estimate the mean pressure that would have
resulted after the same amount of wear in a pure sliding wear test,
For the test conditions employed (us = 0.25 m/s, W = 30 N) and
with the ball stationary. Assuming that wear effectively removes
after four hours of sliding, t = 14,400 s, this reduces to:
a flat-based cap of radius r from the ball, from solid geometry the
k1 = 6.05 × 10−17 .ppm m3 /J [7] volume of this cap is given by:
The dimensionless Archard wear coefficient is then given by πh 2
Qb = (3r + h2 ) [9]
6
kA = 6.05 × 10−17 .ppm.H [8]
where h is the height of the cap (the maximum wear depth). For
where H is the hardness (average for the two surfaces is 7.75 × small amounts of wear, r  h and we can use the parabolic ap-
109 Pa). proximation h = r 2 /2R, so that Eq. [9] reduces to
This simple analysis assumes that the wear rate and the hard-
π r4
ness are the same for both the ball and the disc. Qb = [10]
4 R
Table 5 lists the calculated average wear coefficients calculated
Taking Qb = 2×10−10 m3 (half the total wear volume for the 80◦ C
from the ICP-AES measurements for all the fluids tested.
base oil test in Table 5), the wear scar radius, r , is 1.2 mm and the
Contact Pressure Variation mean pressure is thus only 6 MPa.
It is of interest to estimate the contact pressure at the end of a Analysis of Wear Scar Profiles
four-hour wear test and compare this to a pure sliding test. From
Table 6 lists the wear volumes from the disc and ball calculated
Fig. 6 it can be seen that the track width and thus the contact
from the wear track profiles. Each value is an average of four
length transverse to the sliding direction for the base oil test at
measured profiles from around the wear track from a single test.
80◦ C was approximately 0.75 mm. To calculate the mean contact
It was found that, for a single test, the profile varied only very
pressure, the contact width in the sliding direction is also needed.
slightly around the track and each location gave practically the
The wear has a negligible effect on the overall radius of the wear
same calculated wear value.
track around the ball and, to a first approximation, the worn con-
For the base oil tested at 80◦ C, the total calculated wear vol-
tact can be treated as the Hertzian contact of a cylinder of radius
ume is reasonably close to the value obtained from ICP-AES, as
R and length 0.75 mm on a flat. Such an analysis produces a con-

Fig. 8—Surface profiles of wear scar transverse to the sliding direction Fig. 10—Surface profiles of wear scar transverse to the sliding direction
on disc and ball, respectively, for 0.005% wt. P ZDDP solution at on disc and ball, respectively, for 0.03% wt. P ZDDP solution at
80◦ C (slide-roll ratio = 5). 80◦ C (slide-roll ratio = 5).
 152 J. FAN AND H. SPIKES

TABLE 5—WEAR COEFFICIENTS USING ICP-AES WEAR MEASUREMENTS CONCLUSIONS

Test ppm Fe A new wear test has been developed based on a combination
Temp. 4-Hour Q m3 × k1 m3 /J × kA × of:
Test Lubricant (◦ C) Test 10−10 10−15 10−5

Base oil 80 62 4.05 3.75 2.85 (i) contra-rotating, rolling-sliding motion, so as to combine a
Base oil 100 80 5.23 4.84 3.68
high sliding speed with a low entrainment speed;
0.005 wt.% P ZDDP 80 73 4.77 4.42 3.35
(ii) ICP-AES oil analysis to monitor wear during a test.
0.01 wt.% P ZDDP 80 77 5.03 4.66 3.54
0.02 wt.% P ZDDP 80 24 1.57 1.46 1.11
0.03 wt.% P ZDDP 80 12 0.78 0.73 0.55 This approach is able to measure wear rates in the mild wear
0.05 wt.% P ZDDP 80 1.5 0.098 0.090 0.069
regime (dimensionless wear coefficients of <10−4 ) with good
RL142 80 21.5 1.41 1.30 0.98
repeatability.
RL194 80 15.5 1.02 0.93 0.71
The results show that for ZDDP-containing oils having less
than 0.05% P, the wear rate is strongly dependent on the additive
shown in Table 5. However, for most other fluids, with the excep- concentration. For P contents of less than 0.02% wt., the wear rate
Downloaded by [Stony Brook University] at 19:29 30 October 2014

tion of 0.005% P, the profile analysis suggests considerably lower is greater than for base oil alone, possibly due to the removal of
wear than does ICP-AES. For the base oil at 100◦ C this may result the iron oxide layer by the ZDDP.
from non-ferrous deposits present in wear tracks. In the 0.02% P The feasibility of monitoring wear in mixed rolling-sliding con-
disc profile in Fig. 10, there is clearly a surface film in the track ditions using stylus profilometry has also been tested. It is found
that results in apparently negative wear. This film is rough and at that this method usually predicts lower wear volumes than ICP–
least 100 nm thick, similar in thickness to ZDDP films previously AES, possibly due to the formation of antiwear films or deposits
studied by optical interferometry (Fujita and Spikes (18)). Simi- in the wear tracks.
lar films were also present on the ball and in tests with higher P
contents, and this will have distorted the wear volumes calculated
ACKNOWLEDGEMENT
using profilometry. In the profiles of the 0.01% P ZDDP test in Fig.
The authors wish to acknowledge the financial support of
9, there appears to have been transfer from the disc to the ball,
RohMax that enabled this work to take place.
which supports the idea that very low concentrations of ZDDP
may promote surface adhesion by removing the iron oxide layer.
Another problem with using surface profiles to estimate wear REFERENCES
can be seen by comparing the 80 and 100◦ C base oil test profiles (1) D4172-94 (2004), “Standard Test Method for Wear Preventive Characteris-
in Figs. 6 and 7. For the 80◦ C test, most wear has occurred on the tics of Lubricating Fluid (Four-ball Method),” ASTME International, West
Conshohocken, PA.
disc, while for the 100◦ C test, most has occurred on the ball. This (2) D2714-94 (2003), “Standard Test Method for Calibration and Operation
variation may result from slight differences in the microstructure of the Falex Block-on-Ring Friction and Wear Testing Machine,” ASTM
or relative hardness of the two surfaces but means that when us- International, West Conshohocken, PA.
(3) D2670-95 (2004), “Standard Test Method for Measuring Wear Properties of
ing profilometry to estimate wear, it is essential to examine both Fluid Lubricants (Falex Pin and Vee Block Method),” ASTM International,
surfaces. West Conshohocken, PA.
From all of the above it appears that the estimation of small (4) Benzing R, Hopkins, V., Petronio, M. and Villforth, Jr., F. (1966), Friction
and Wear Devices – A Survey, ASLE, Park Ridge, IL.
amounts of wear from rolling-sliding tests using profilometry is (5) Brown, R.D. (1969) “Chapter 12. Test Methods,” in Boundary Lubrication,
problematic and certainly requires some means of reliably re- an Appraisal of World Literature, pp 341-292, ed. Ling, F.F., Klaus, E.E. and
moving any reaction films or deposits from the surfaces prior to Fein, R.S., ASME, New York.
(6) D6079-04. (2004), “Standard Test Method for Evaluating Lubricity of
measurement. Diesel Fuels by the High-Frequency Reciprocating Rig (HFRR),” ASTM
International, West Conshohocken, PA.
TABLE 6—ESTIMATED VOLUMES FROM SURFACE PROFILOMETRY (7) D5707-98e1. (2004), “Standard Test Method for Measuring Friction
and Wear Properties of Lubricating Grease using a High-Frequency,
Test Linear-Oscillation (SRV) Test Machine,” ASTM International, West Con-
Temp. Disc, Qd Ball, Qb Total, QT shohocken, PA.
(8) Williams, J.A. (1994), Engineering Tribology, Chapter 5, p167, Oxford,
Test Lubricant (◦ C) m3 × 10−10 m3 × 10−10 m3 × 10−10
OUP.
(9) Ruff, A.W. (1997), “Wear Measurement,” in Friction and Wear Test-
Base oil 80 3.21 0.54 3.75
ing Source Book of Selected References from ASTM Standards and
Base oil 100 0.34 3.29 3.63 ASM Handbooks, pp 22-29, ASTM and ASM, West Conshohocken,
0.005 wt. % P ZDDP 80 0.60 7.43 8.03 PA.
0.01 wt. % P ZDDP 80 0.42 −0.03 0.39 (10) Gahlin, R. and Jacobson, S. (1998), “A Novel Method to Map and Quantify
0.02 wt.% P ZDDP 80 −0.04 0.01 −0.03 Wear on a Micro-Scale,” Wear, 222, pp 93-102.
(11) Conlon, T.W. (1974), “Thin Layer Activation by Accelerated Ions: Ap-
0.03 wt.% P ZDDP 80 −0.01 −0.04 −0.05
plication to Measurement of Industrial Wear,” Wear, 29(1), pp 69-
0.05 wt.% P ZDDP 80 −0.04 0.09 0.05 80.
RL142 80 −0.04 0.05 0.01 (12) Blatchley, C. (1992), “Radionuclide Methods,” in Metals Handbook: Fric-
RL194 80 −0.02 −0.01 −0.03 tion, Lubrication and Wear Technology, Vol. 18, pp 319-330, Peter Blau,
Ed., ASM International.
 Mild Lubricated Wear in Rolling-Sliding Contacts 153

(13) Dowson, D. and Hamrock, B.J. (1981), Ball Bearing Lubrication. The Elas- (16) Eyre, T.S. and Fitter, C. (1983), “Application of Oil Analysis Techniques to
tohydrodynamics of EHD. Elliptical Contacts. John Wiley&Sons, New York. the Study of Cylinder Liner Wear,” Wear, 90, pp 31-37.
(14) Blok, H. (1948), “Gear Wear as Related to Viscosity of Gear Oils,” in Proc. (17) Martin J.J. (1999), “Antiwear Mechanisms of Zinc Dithiophosphate: A
Summer Conference-MIT, Mechanical Wear. ASME, 61(25), pp 199-227. Chemical Hardness Approach.” Trib. Lett., 6, pp 1-8.
(15) Wedeven, L.V., (2005) “Load Capacity Method.” Wedeven Associates Inc. (18) Fujita, H. and Spikes, H.A. (2004), “Formation of Zinc Dithiophosphate
www.wedeven.com. Accessed 2/14/07. Antiwear Films.” Proc. I. Mech. E., J218, pp 265-277.
Downloaded by [Stony Brook University] at 19:29 30 October 2014