You are on page 1of 8

Tribol Lett (2010) 38:25–32

DOI 10.1007/s11249-009-9565-9

ORIGINAL PAPER

Influence of Steel Type on the Propensity for Tribochemical Wear


in Boundary Lubrication with a Wind Turbine Gear Oil
R. D. Evans • G. L. Doll • C. H. Hager •

J. Y. Howe

Received: 22 September 2009 / Accepted: 23 December 2009 / Published online: 23 January 2010
Ó Springer Science+Business Media, LLC 2010

Abstract Tribochemical wear may occur at the interface macroscopic surface-initiated damage such as micropitting
between a surface and a lubricant as a result of chemical in bearings and gears.
and mechanical interactions in a tribological contact.
Understanding the onset of tribochemical wear damage on Keywords Antiwear additives  Extreme pressure
component surfaces requires the use of high resolution additives  Additive interaction  Additive decomposition 
techniques such as transmission electron microscopy Power generation  Boundary lubrication wear 
(TEM). In this study, two steel types, case carburized AISI Rolling element bearings: general  Ferrous alloys: steel 
3310 and through-hardened AISI 52100, were wear tested Tapered roller bearings  TEM  Wear mechanisms
using a ball-on-disk rolling/sliding contact tribometer in
fully formulated commercial wind turbine gearbox oil
under boundary lubrication conditions with 10% slip. With 1 Introduction
the exception of steel type, all other test conditions were
held constant. Conventional tribofilm analysis in the wear Tribochemical wear occurs at the interface between a
tracks was performed using X-ray photoelectron spectros- surface and a lubricant as a result of chemical and
copy, and no significant composition differences were mechanical interactions in a tribological contact. Evidence
detected in the tribofilms for the different steel disk types. of this type of wear is found by analyzing residual surface
However, TEM analysis revealed significant tribochemical tribofilms and the near-surface material exposed to the
wear differences between the two steel types at multiple tribological contact (depth \ 500 nm). Damage on this
length scales, from the near-surface material microstructure length scale is not usually detectable in its early stages
(depth \ 500 nm) to the tribofilm nanostructure. Nano- using standard observation techniques, such as optical
meter-scale interfacial cracking and surface particle microscopy or surface analytical techniques that do not
detachment was observed for the AISI 52100 case, whereas have nanometer-scale location resolution. Cross-sectional
the tribofilm/substrate interface was abrupt and undamaged transmission electron microscopy (TEM) is well-suited for
for the AISI 3310 case. Differences in tribofilm structure, probing the microstructure of materials and surfaces on the
including the location and orientation of MoS2 single sheet nanometer scale. For steel surfaces, the authors have
inclusions, were observed as a function of steel type as observed that tribochemical wear can be manifested as
well. It is suggested that the tribochemical wear modes residual material strain, nanometer scale shallow cracks,
observed in these experiments may be origins of and surface particle lift-off into the tribofilm. Although
other mechanisms must be considered, it is believed that
tribochemical wear damage on the nanometer scale
R. D. Evans (&)  G. L. Doll  C. H. Hager undoubtedly has the potential of initiating macroscopic
Timken Technology Center, North Canton, OH, USA surface-related damage modes such as micropitting, peel-
e-mail: ryan.evans@timken.com
ing, and point-surface origin fatigue spalling.
J. Y. Howe Enabled by advances in focused ion beam technology
Oak Ridge National Laboratory, Oak Ridge, TN, USA (FIB) to produce TEM specimens during the last decade,

123
26 Tribol Lett (2010) 38:25–32

tribochemical wear studies from real hardware surfaces rig tested identically in a commercially available wind
have become possible. Rainforth et al. [1] were able to turbine gearbox oil lubricant. Results indicated that the
characterize friction-induced oxide layers generated in rig manifestation of tribochemical wear was strongly influ-
sliding tests from a variety of material types using classical enced by surface material composition and structure.
TEM sample preparation techniques, but their ability to
perform site-specific analysis on their wear tracks was
admittedly limited. On the other hand, Evans et al. utilized 2 Experimental
FIB sample preparation and were the first to evaluate
tribofilm formation and near-surface material damage Friction and wear tests were performed using the WAM6
trends in tapered roller bearings as a function of lubricant ball-on-disk tribometer (Wedeven Associates, Inc., Edg-
type and dimensionless film thickness k [2, 3]. In those mont, PA), which allows independent control of ball and
studies, shallow cracks were observed on case carburized disk spindles under load while monitoring traction (fric-
AISI 8119 steel raceways only for the most severe tion) coefficient for the contact. Test balls were made of
boundary lubrication conditions in an oil without sulfur– AISI 52100 steel with an outer diameter of 20.64 mm and
phosphorus extreme pressure/antiwear additives [2]. No finished to AFBMA grade 25. Surface finish was measured
cracks were observed under comparable test conditions for each ball using an interferometric surface-mapping
when commercially available sulfur-phosphorus additives microscope. The root mean square roughness Sq was cal-
were used [3]. In a separate but similar study, shallow culated for each measurement using three-dimensional
cracks were observed in a tested tapered roller bearing datasets. Ten measurements were made per ball using 209
raceway surface made of case carburized AISI 8119 steel magnification and the median roughness for each ball fell
only when an intentionally aggressive experimental sulfur– within the range Sq * 23–33 nm. Medians are reported
phosphorus additive package was used [4]. Reichelt et al. because they better represented the modes of the non-
used FIB/TEM to evaluate tribochemical wear of through- Gaussian datasets than means. A different ball wear track
hardened AISI 52100 bearing surfaces after testing with a was used for each test on the tribometer. The disks were
variety of lubricants, showing that differences in lubricant either through-hardened AISI 52100 or case carburized
additive formulation can be responsible for stark differ- AISI 3310 steel with an outer diameter of 101.6 mm. The
ences in tribofilm thickness and composition [5]. Other nominal composition of AISI 52100 steel is 1.04 wt% C,
recent studies have applied FIB/TEM to evaluate tribofilms 0.35 wt% Mn, 0.25 wt% Si, and 1.45 wt% Cr with the
formed in a sliding contact using a lubricant containing balance Fe. The nominal composition of AISI 3310 steel is
zinc dialkyl dithiophosphate (ZDDP) [6] and in a rolling/ 0.11 wt% C, 0.53 wt% Mn, 1.58 wt% Cr, and 3.5 wt% Ni
sliding contact using ZDDP and molybdenum dialkyl with the balance Fe. Both disk types had a circumferen-
dithiocarbamate (MoDTC) [7]. The later study by Laine tially ground finish with median Sq * 471 nm for the AISI
et al. sought to characterize the additive effects on micro- 52100 disk and median Sq * 435 nm for the AISI 3310
pitting, and shows an example of tribochemical wear disk. The disks were measured with an interferometric
damage (detached steel particles suspended in the tribo- surface-mapping microscope in the same manner as the
film) on a case carburized AISI 5115 (equiv.) steel surface, balls, but with fifty measurements per disk at 109 mag-
albeit for the sample tested in ZDDP ? MoDTC with the nification. The hardness of the test balls and disks was
best micropitting resistance. It is assumed that the intent of within the range 58-62 HRC. Post-test optical metallo-
that study was not to do a full tribochemical wear com- graphic evaluation confirmed the expected microstructures
parison because TEM data were not included for surfaces for the two steel types. Disk surface finish and hardness
tested in the lubricant most susceptible to micropitting levels were representative of those expected for large
(ZDDP with no MoDTC). rolling element bearings typically found in heavy industrial
In the previous studies, emphasis was placed on char- and power generation machinery.
acterizing the effects of lubricant additive formulation and/ Commercially available wind turbine gear oil with ISO
or contact conditions on tribofilm formation and tribo- 320 viscosity grade and an extreme pressure and anti-wear
chemical wear. To the hardware designer, material selec- (EP/AW) additive package was used as the test lubricant.
tion for tribological surfaces is of equal concern, especially The base stock for the oil was polyalphaolefin with specific
in applications such as wind turbine transmissions where gravity *0.87 at 15.6 °C as reported by the manufacturer.
new components must work well in a pre-determined Measured kinematic viscosities were 332 cSt and 33 cSt at
lubrication environment. The objective of this work was to 40 and 100 °C, respectively, giving a viscosity index of
characterize tribochemical wear with FIB/TEM for two 140. X-ray fluorescence and optical emission spectroscopy
different types of steel surfaces, one case carburized (AISI measurements indicated significant quantities of S, Mg, P,
3310) and one through-hardened (AISI 52100), that were Zn, and Mo in the oil, which are constituents of the

123
Tribol Lett (2010) 38:25–32 27

otherwise unknown fully formulated additive package in performed on the wear track surface. A PHI-VersaProbe
the lubricant. XPS Scanning Microprobe was used with an Ar? gun for
Both Stribeck-type tests and 24-h wear tests were per- depth profiling. Cross-sectional wear track surface samples
formed in this study. Stribeck tests were performed to were prepared using focused ion beam (FIB) milling for
evaluate traction coefficients in the lubricated contacts as a evaluation with TEM. Prior to FIB milling, the samples
function of dimensionless lubricant film thickness k, which were ultrasonically cleaned in isopropyl alcohol for 5 min,
is the calculated minimum film thickness divided by the tungsten metal was sputtered onto the wear tracks, and the
composite root mean square surface roughness for the samples were sectioned to a workable size using a diamond
contact ((S2q,1 ? S2q,2)1/2). Elastohydrodynamic minimum saw. A Hitachi FB-2000 FIB system with 30 kV Ga? ions
film thicknesses were calculated using the Hamrock– and a micro-sampling apparatus was used for preparing the
Dowson equation for fully flooded elliptical contacts [8] lift-out TEM specimens, which were then examined with a
using pressure-viscosity coefficient information that was Hitachi HF-3300 FEG TEM/STEM at 300 kV. The lift-out
provided by the lubricant manufacturer. The applied nor- sample orientation was such that the ball rolled across the
mal load for Stribeck tests was 100 N giving a calculated horizon of all images, that is, the major axis of the lift out
Hertzian contact stress of *1.32 GPa. Entrainment sample was parallel with the wear track circumference.
velocity was ramped from 11 to 0 m/s over 150 s, slide-to- Energy dispersive spectroscopy (EDS) for chemical anal-
roll ratio or slip was set at 10%, and the disk temperature ysis was performed in the HF-3300 TEM using a Noran
was held at 100 °C throughout the tests. Neglecting the Six X-ray Special Acquisition system, which uses a
impact of slip on lubricant film thickness and assuming Nano Trace EDS detector with NORVAR light elements
initial ball and disk roughness levels for k calculations, the window.
Stribeck tests were estimated to scan from k * 1.2 ? 0.
As a result of these assumptions, the Stribeck plot k scaling
is approximate. 3 Results and Discussion
The test conditions for the 24-h wear tests were different
than those for the Stribeck tests. Their purpose was to Traction performance was compared for experiments per-
generate conditions amenable to tribofilm formation in a formed on case carburized AISI 3310 and through-hard-
short period of time. To do this, the contact stress was ened AISI 52100 disks using Stribeck and 24-h wear tests
increased and the entrainment velocity was decreased to as shown in Fig. 1. Initial Stribeck tests on the disks with
produce boundary lubrication conditions. Applied normal the wind turbine gear oil did not show a significant traction
load was 800 N resulting in a calculated Hertzian contact difference between the two steel types as seen in Fig. 1a.
stress of *2.65 GPa. Entrainment velocity was 0.6 m/s This was expected since the disks had comparable rough-
with slip held at 10%. The test duration was 24 h, and the ness and the test duration was presumably too short to
disk temperature was held at 100 °C throughout the entire permit full run-in and tribofilm formation. Traction coef-
test. These conditions provided for k * 0.1, based on the ficient trends during the 24-h tests (two tests per disk type)
initial surface roughness. Lubricant was continuously in Fig. 1b indicate lower traction for the AISI 3310 disk
administered to the center of the rotating disk where it was than for the AISI 52100 disk. The wear tracks experienced
spun out across the disk surface in all tests providing fully run-in and tribofilms formed in these longer duration tests,
flooded conditions. In addition, the ball spindle angle was so the traction coefficient trends reflect the results of those
set according to the disk track diameter to minimize/ dynamics. It is noteworthy that no significant wear loss or
eliminate spin in the contact prior to each Stribeck and plastic deformation was observed in either of the disk wear
24-h test. tracks using an interferometric surface-mapping micro-
In addition to the above tests, a combined protocol of scope after the 24-h tests. The wear tracks were distin-
24-h tests followed immediately by Stribeck tests were guishable from the original disk surface by a slight
performed on the same wear track without dismounting the reduction in roughness, on the order of Sq * 50–100 nm
specimens using the test conditions described above for lower depending on measurement location. When a Stri-
each test type. beck test was performed immediately after a 24-h wear test
Two Stribeck tests, two 24-h wear tests, and two com- on the same track, the traction coefficients across k were
bined 24-h wear tests plus subsequent Stribeck tests were decreased for both disk types as shown in Fig. 1a. This
performed on each of the AISI 52100 and AISI 3310 disks. trend is in agreement with the traction coefficients in the
Examinations of the tribofilms were performed on the 24-h wear tests, and indicates the presence of run-in
second 24-h wear test track for each disk. To confirm the roughness reduction and/or the presence of low friction
presence of a tribofilm and characterize its chemical tribofilms. AISI 52100 appeared to run at a slightly higher
composition, X-ray photoelectron spectroscopy (XPS) was traction coefficient in the boundary lubrication regime than

123
28 Tribol Lett (2010) 38:25–32

(a) 0.20 Fe and O. Chemical shift information indicated the pres-


0.18 52100 - initial ence of mixed phosphates, sulfides, zinc oxides/sulfides,
3310 - initial
0.16 52100 - after 24-hr test and molybdenum disulfide (MoS2) in both tribofilms.
Traction Coefficient
3310 - after 24-hr test
0.14 Excluding film coverage appearance in SXI imaging, the
0.12 XPS chemical data suggested indistinguishable tribofilms
0.10 on the AISI 52100 and AISI 3310 wear tracks.
0.08 To further characterize the differences in tribochemical
0.06 wear damage as a function of steel type, TEM evaluations
0.04 were performed on FIB samples taken from both inside and
0.02 outside of the disk wear tracks. A low magnification cross-
0.00 sectional view of the AISI 52100 sample ground surface
0.01 0.1 1 outside of the wear track is shown in Fig. 2. Selected area
λ (approx.) micro-diffraction (micro-SAD) patterns are inset for the
near-surface material (depth \ 500 nm) and the larger
0.12
(b) subsurface grains. The diffraction patterns were consistent
52100-1
0.10 52100-2 with mixed Fe phases (e.g., ferrite/martensite). Clear spot
3310-1
Traction Coefficient

3310-2
patterns such as the one shown for the deeper grains
0.08 indicated large grains with respect to the diffracting vol-
ume. On the other hand, large quantities of small grains
0.06 within the same diffracting volume produced many over-
lapping diffraction patterns, giving the appearance of rings.
0.04
Qualitatively by these trends, the patterns inset in Fig. 2
0.02 indicated smaller grains near the surface of the sample and
larger grains in the subsurface. The formation of small
0.00 near-surface grains observed in this specimen taken from
4 4 4 4
0 2x10 4x10 6x10 8x10 outside the wear track was attributed to the grinding pro-
Time (s) cess during disk fabrication. Large spheroidal carbides
(size * 1 lm) also were observed in the AISI 52100
Fig. 1 a Stribeck plots before and after 24-h wear tests. Two tests
were performed per disk type to confirm repeatability, but one curve sample shown in Fig. 2. Diffraction and the presence of
from each condition is shown for clarity. Initial disk roughness was bend contours in the image indicated that the carbides were
used to calculate k in the plots. b Traction coefficient as a function of crystalline. EDS revealed that the carbides contained Fe,
test time for the 24-h wear tests. Two tests were performed per disk Cr, and C, whereas minimal Cr was detected outside of the
type
carbides in the Fe matrix. Dark shadows/streaks visible in

AISI 3310, consistent with the 24-h test results in Fig. 1b,
but additional tests would be required to confirm that
observation.
The 24-h wear tests and subsequent Stribeck tests sug-
gested the presence of tribofilms on the surface because
traction coefficients decreased from their initial levels with
increasing contact cycles and test duration. However,
detailed wear track analysis on the disks was required to
explore tribochemical interaction and wear differences.
The first step was to determine whether films were present
on the 24-h wear test tracks using X-ray photoelectron
spectroscopy (XPS). Scanning X-ray imaging (SXI) indi-
cated that tribofilms were present, but their appearance
differed depending on steel type. The tribofilm on the AISI
52100 appeared to be thin and covered the track in thin
streaks. On the other hand, the tribofilm on the AISI 3310
Fig. 2 Cross-sectional TEM image of the AISI 52100 surface outside
appeared thick in blotchy patterns within the wear track
the wear track. Shadows/streaks observed below the spheroidal
area. Depth profiling of these films indicated the presence carbides are attribued to FIB milling. Micro-SAD patterns were
of Mg, P, S, Zn, and Mo in both tribofilms, in addition to collected at the approximate depths indicated by arrows

123
Tribol Lett (2010) 38:25–32 29

C
1500

Counts
75 Zn W
Mg
P
Mo, S
50 Zn
W
Cu
25 Fe Zn Mo
Fig. 3 Cross-sectional TEM image of the AISI 52100 wear track Si
surface. The ball rolled across the horizon of the image in the tribo-
0
test. Two images were spliced together in this figure, and white
0 2 4 6 8 10 16 18 20
arrows indicate the location of the overlap
Energy (keV)

Fig. 2 below the carbide particles were attributed to the Fig. 4 EDS data indicating the overall composition of the tribofilm
on the AISI 52100 wear track sample. The Cu signal is from the TEM
different etch rate of the carbide as compared to the matrix grid
during ion milling from a direction normal to the original
surface.
Tribochemical wear damage was apparent in the wear views. Measured fringe spacing was *0.6 nm, consistent
track of the AISI 52100 sample as shown in Fig. 3. The with the c-axis basal plane spacing of MoS2. In addition to
tribofilm appears bright in the TEM image because it has a large MoS2 sheets, suspended Fe particles were detected in
smaller interaction cross-section than the steel substrate. the tribofilm with EDS that presumably detached from the
The surface experienced shallow cracking and penetration original substrate as the tribofilm was forming during tri-
of the tribofilm into the surface. This damage was limited to bological contact.
the near-surface material and did not penetrate deeper than Significant differences were observed in the near-sur-
200 nm into the surface. In the center of the spliced image face material and tribofilm structures found for the AISI
in Fig. 3, the tribofilm completely surrounds a portion of the 3310 disk as compared to the AISI 52100 disk. Figure 6
original surface which presumably would have become a shows a cross section of the AISI 3310 disk at a location
debris particle by mechanical detachment with continued outside of the wear track. Similar to the AISI 52100 case,
tribological contact. The tribofilm was discontinuous across fine-grained material was found at the near-surface region
the wear track sample, with thick (*100 nm) and thin (depth\500 nm) and larger grains were observed deeper in
(*10 nm) regions, and there were areas with no tribofilm the material. The inset micro-SAD pattern of the near-
present. surface material in Fig. 6 shows distinct rings consistent
At conventional TEM magnification, the AISI 52100 with Fe phases that indicate a finer grain structure than that
wear track tribofilm appeared to be mostly homogeneous in observed for the AISI 52100 sample at the same depth. In
structure and composition. EDS data for the tribofilm addition, significantly smaller spheroidal carbide inclusions
shown in Fig. 4 indicated the presence of Mg, P, S, Zn, and (size * 100 nm, an order of magnitude smaller than car-
Mo, as well as C, O, and Fe, consistent with the XPS bides in AISI 52100) were uniformly distributed through-
observations. Cu and W in Fig. 4 are attributable to the out the sample. EDS indicated that Cr was segregated to
TEM grid and FIB protective cap layer, respectively. the carbide phase, and that Ni was present in the Fe matrix
However, closer inspection of the AISI 52100 tribofilm in as expected.
high-resolution TEM mode (HRTEM) as shown in Fig. 5 Tribochemical wear damage was not observed within
revealed atomic-scale crystalline features. Figure 5a shows the AISI 3310 wear track. In fact, the AISI 3310 tribofilm
a thick region of the tribofilm (*100 nm) in which long differed from the AISI 52100 tribofilm in thickness and
lattice fringe sheets are visible (assuming three-dimen- structure considerably. As shown in Fig. 7, the tribofilm on
sional structure). The large sheets were found predomi- AISI 3310 was mostly smooth, uniform, and continuous
nantly near the bottom of the tribofilm at the interface with across the sample. Its thickness was *200 nm, and a sharp
the steel and near the top surface of the tribofilm. As shown interface was observed between the tribofilm and the steel
in a thinner region of tribofilm (*30 nm) in Fig. 5b, the surface. No cracks, tribofilm penetration, or suspended
sheets could consist of 2–10 lattice fringes and generally debris particles were observed. Closer inspection of
were oriented parallel to the surface in these cross-sectional the tribofilm structure in Fig. 8 revealed a mostly

123
30 Tribol Lett (2010) 38:25–32

Fig. 6 Cross-sectional TEM image of the AISI 3310 surface outside


the wear track. Gray spheroidal inclusions in the matrix are carbides.
Micro-SAD patterns were collected at the approximate depths
indicated by arrows

Fig. 5 Cross-sectional HRTEM image of the AISI 52100 wear track


tribofilm at a thick and b thin locations. Arrows indicate example
locations of lattice fringes with a measured spacing of *0.6 nm

homogeneous structure through the thickness of the film.


However, many *10 nm long curved sheets (again
assuming three-dimensional structure) filled the volume of
the film as indicated by lattice fringes in Fig. 8b. The
predominant orientation of these sheets was parallel to the
surface, but many examples of sheets that curl in random
directions were found. Although most of the sheets did not
appear to have long-range order in the c-axis direction, Fig. 7 Cross-sectional TEM image of the AISI 3310 surface inside
short stacks of sheets (2–5 sheets thick) were found with the wear track. The ball rolled across the horizon of the image in the
tribo-test
spacing of *0.6 nm, again consistent with MoS2 as was
observed for the larger sheets in the AISI 52100 tribofilm.
Micro-SAD was performed on the thick AISI 3310 tribo- A tribofilm structure containing dispersed small MoS2
film and rings indexable to MoS2 (molybdenite-2H) as sheets was reported previously by Grossiord et al. [9]. The
shown in Fig. 9a were observed. EDS as shown in Fig. 9b sheets were generated in reciprocating sliding tests in an
confirmed the expected presence of P, S, Zn, and Mo, as AISI 52100 steel contact with base oil containing the
well as C, O, and Fe in the tribofilm. MoDTC additive only. Wear debris were collected on a

123
Tribol Lett (2010) 38:25–32 31

(a)

2500
(b) C

2000
Mo, S
1500
Counts

1000

O
Cu
500 Cu P Zn Mo
W Fe
Cu Mo
0
0 2 4 6 8 16 18 2
20
Energy (keV)

Fig. 9 a Micro-SAD pattern from the bulk tribofilm on the AISI 3310
wear track surface indexed to MoS2 (molybdenite-2H) and b EDS
spectrum also collected from the AISI 3310 tribofilm
Fig. 8 Higher magnification cross-sectional TEM images of the AISI
3310 tribofilm structure. a A view of the tribofilm attached to the
substrate and b a higher magnification view of the film structure tribofilm nano-structure was observed in the AISI 3310
revealing curved lattice fringes with a spacing of *0.6 nm in case only in the present rolling/sliding contact study. This
locations where they are stacked observation suggests that base material alone likely does
not solely determine tribofilm structure, but that many
holey carbon grid after testing and examined in TEM. other tribological factors must influence tribofilm structure
Their tribofilm particles contained almost identical small formation on a nanometer scale.
dispersed MoS2 sheets as were observed in the intact AISI
3310 tribofilm in the present study (Fig. 8). Consistent with
the present observations, Grossiord et al. did not detect 4 Conclusions
MoS2 crystals with long-range order or stacked sheets (i.e.,
no (002) rings in diffraction patterns collected from the Tribochemical wear differences between the AISI 52100
tribofilm particles). Because commercial wind turbine gear and 3310 steel types were apparent in the TEM on multiple
oil was tested in the present experiments, it is not possible length scales. Whereas the traction coefficient results
to conclude whether MoDTC was present in this formu- during rig tests were similar and the tribofilm compositions
lation, but the agreement between the present and the as measured by XPS were approximately the same, the
Grossiord et al. results are striking. It is also interesting that near-surface material structures and tribofilms were dif-
the Grossiord et al. tribofilms were produced in an AISI ferent at the micro- and nanometer levels in the samples
52100 reciprocating sliding contact whereas a similar studied. The AISI 52100 near-surface material structure

123
32 Tribol Lett (2010) 38:25–32

consisted of a Fe matrix with a relatively coarse grain damage, so it is recommended that tribochemical wear and
structure containing large (Fe, Cr) spheroidal carbides. In lubricant compatibility be evaluated during the material
contrast, the AISI 3310 near-surface material structure was selection phase of new product development programs
Fe containing Ni as expected and a relatively finer grain whenever possible.
structure with smaller (Fe, Cr) spheroidal carbide inclu-
sions. Tribochemical wear was detected for the AISI 52100 Acknowledgments The Timken Company is acknowledged for
support of this work and permission to publish. D. W. Coffey from
wear track only. It had a tribofilm with non-uniform Oak Ridge National Laboratory is acknowledged for FIB sample
thickness, shallow cracks, and penetration of the tribofilm preparation. G. A. Richter from The Timken Company is thanked for
into the surface. On the other hand, the AISI 3310 wear tribology test lab assistance and for performing the rig tests. P. J.
track had a uniformly thick tribofilm with no cracks or Shiller and R. L. Aubrey from Timken are also acknowledged for
lubricant characterization. W. Jennings from Case Western Reserve
penetration into the surface and a sharp, smooth interface University is thanked for X-ray photoelectron spectroscopy assis-
with the original substrate. The tribofilm nanostructure for tance. Microscopy research at Oak Ridge National Laboratory’s
the AISI 52100 case contained large MoS2 sheets that SHaRE User Facility was supported by the Scientific User Facilities
generally ran parallel to the substrate and were segregated Division, Office of Basic Energy Sciences, U.S. Department of
Energy.
to the surface and interface, as well as suspended Fe par-
ticles in the film. Smaller MoS2 sheets were dispersed
uniformly and oriented somewhat randomly throughout the References
AISI 3310 tribofilm in sufficient quantity for identification
by micro-SAD. 1. Rainforth, W.M., Leonard, A.J., Perrin, C., Bedolla-Jacuinde, A.,
This study introduced many new technical questions. Wang, Y., Jones, H., Luo, Q.: High resolution observations of
What characteristics of the AISI 52100 surface made it friction-induced oxide and its interaction with the worn surface.
Tribol. Int. 35, 731–748 (2002)
more susceptible to tribochemical wear (e.g., cracking, 2. Evans, R.D., More, K.L., Darragh, C.V., Nixon, H.P.: Trans-
tribofilm penetration, surface particle detachment) under mission electron microscopy of boundary-lubricated bearing
the present test conditions than the AISI 3310 surface? surfaces. Part 1: mineral oil lubricant. Tribol. Trans. 47, 430–439
Why was MoS2 incorporated in the AISI 52100 tribofilm as (2004)
3. Evans, R.D., More, K.L., Darragh, C.V., Nixon, H.P.: Trans-
large sheets rather than smaller uniformly dispersed sheets mission electron microscopy of boundary-lubricated bearing
as in the AISI 3310 tribofilm? What roles did alloy com- surfaces. Part 2: mineral oil lubricant with sulfur- and phospho-
position (e.g., the presence of Ni in AISI 3310) play in rus-containing gear oil additives. Tribol. Trans. 48, 299–307
tribofilm formation and damage susceptibility? Future (2005)
4. Evans, R.D., Nixon, H.P., Darragh, C.V., Howe, J.Y., Coffey,
work is required to address these open issues, and the D.W.: Effects of extreme pressure additive chemistry on rolling
challenge remains to develop generalized scientific mech- element bearing surface durability. Tribol. Int. 40, 1649–1654
anisms that describe more than just one specific situation at (2007)
a time. 5. Reichelt, M., Weirich, T.E., Mayer, J., Wolf, T., Loos, J., Gold,
P.W., Fajfrowski, M.: TEM and nanomechanical studies on tri-
In addition, nanometer-scale tribochemical wear dam- bological surface modifications formed on roller bearings under
age has not yet been directly linked to macroscopic surface controlled lubrication conditions. J. Mater. Sci. 41, 4543–4553
damage modes like micropitting or point-surface origin (2006)
spalling. However, high-load, high slip in rolling/sliding 6. Ito, K., Martin, J.M., Minfray, C., Kato, K.: Low-friction tribo-
film formed by the reaction of ZDDP on iron oxide. Tribol. Int.
contact, and boundary lubricated contact conditions such as 39, 1538–1544 (2006)
those used to generate tribochemical wear damage in this 7. Laine, E., Olver, A.V., Lekstrom, M.F., Shollock, B.A., Bever-
study have been associated with micropitting [10, 11] in idge, T.A., Hua, D.Y.: The effect of a friction modifier additive
bearings and gears. Although the present results alone do on micropitting. Tribol. Trans. 52, 526–533 (2009)
8. Hamrock, B.J., Dowson, D.: Isothermal elastodynamic lubrica-
not definitively preclude the use of AISI 52100 in wind tion of point contacts, part III—fully flooded results. J. Lubr.
turbine bearings, the intention of this work is to show the Technol. 99, 264–276 (1977)
extent to which lubricant-material interactions and result- 9. Grossiord, C., Varlot, K., Martin, J.M., Le Mogne, Th., Esnouf,
ing microscopic damage can differ based on the material C., Inoue, K.: MoS2 single sheet lubrication by molybdenum
dithiocarbamate. Tribol. Int. 31, 737–743 (1998)
selection alone with all else constant. Admittedly, the ball- 10. Webster, M.N., Norbart, C.J.J.: An experimental investigation of
on-disk test settings in this study were aggressive in contact micropitting using a roller disk machine. Tribol. Trans. 38, 883–
stress and slip compared to typical wind turbine gearbox 893 (1995)
bearing conditions to accelerate the rate of tribofilm for- 11. Oila, A., Bull, S.J.: Assessment of the factors influencing
micropitting in rolling/sliding contacts. Wear 258, 1510–1524
mation within the 24-h test duration. However, this work (2005)
makes it clear that material selection and surface charac-
teristics can affect the propensity for tribochemical wear

123

You might also like