Wear 346-347 (2016) 614

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Temperature-dependent fretting damage of high strength stainless

steel sheets
Michael R. Hirsch a, Richard W. Neu a,b,n
a
b

George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, 801 Ferst Drive, Atlanta, GA 30332-0405, USA
School of Materials Science and Engineering, Georgia Institute of Technology, 771 Ferst Drive, Atlanta, GA 30332-0245, USA

art ic l e i nf o

a b s t r a c t

Article history:
Received 3 April 2015
Received in revised form
20 October 2015
Accepted 24 October 2015
Available online 31 October 2015

Changes in temperature can have a large effect on the fretting damage response for materials with
properties that are highly temperature sensitive, such as austenitic stainless steel. The fretting damage
response of thin sheets of AISI 301 stainless steel in the full hard condition in contact with AISI 52100
steel is investigated for temperatures ranging from 20 C to 250 C. Changes in material resistance to
fatigue damage and wear are characterized by determination of mechanical properties and oxidation
behavior with changing temperature. The drivers for wear and fatigue are evaluated via nite element
analysis over a range of temperatures. The net effect of changes in the wear and fatigue damage drivers
and resistance with changing temperature is discussed.
& 2015 Elsevier B.V. All rights reserved.

Keywords:
Fretting
Fatigue
Frictional energy dissipation
Stainless steel
Glaze oxide

1. Introduction
Fretting is a low amplitude oscillatory motion between contacting bodies which results in damage of the materials in contact.
Wear and fatigue damage can occur, with the type and extent of
damage being sensitive to many parameters [1]. Changes in temperature affect many aspects of the interaction, including the
mechanical properties and oxidation behavior, which in turn affect
the friction response and the extent of fatigue damage.
The mechanical properties of austenitic stainless steels are
highly temperature sensitive. Austenitic stainless steel is comprised austenite in the annealed state, and can undergo a transformation to martensite due to a combination of thermal treatment and/or mechanical loading. Martensite has a higher hardness
than austenite, resulting in an increase in strength. The occurrence
of the transformation in regions of high deformation results in
local hardening and reduction in the local plastic instability, thus
increasing ductility [2]. The stability of the austenite increases
with increasing temperature, and is sensitive to the composition
[36]. The temperature above which martensite cannot be formed
by mechanical loading, Md, has been reported as 100 C for 301
stainless steel [7,8]. Therefore, the benecial effects associated
n
Corresponding author at: George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, 801 Ferst Drive, Atlanta, GA 30332-0405,
USA. Tel.: 1 404 894 3074.
E-mail address: rick.neu@gatech.edu (R.W. Neu).

http://dx.doi.org/10.1016/j.wear.2015.10.007
0043-1648/& 2015 Elsevier B.V. All rights reserved.

with the transformation are not present at temperatures above

100 C, and the ductility is drastically reduced [9].
The wear behavior has a strong inuence on the extent of fatigue
damage by means of the material removal rate and the tangential
force. A high material removal rate results in the removal of surface
material which was subjected to fatigue damage, thus increasing
the fatigue life [1013]. A high material removal rate also increases
the contact conformity, thus increasing the contact area which
reduces the contact stresses and increases fatigue life [10,14,15].
However, a large wear depth may result in the formation of a notch
which increases fatigue damage from a remotely applied load. Also,
a deep wear scar may cause a reduction in fatigue life if the wear
depth is signicant relative to the thickness of the component [16].
An adhesive wear mode is typically associated with a high tangential force, which increases the cyclic stresses and therefore
increases the level of fatigue damage. A high oxidation environment
can reduce the likelihood for adhesive wear, thus lowering the
friction and increasing the fretting fatigue life [17,18]. Thus the level
of fatigue damage is sensitive to the wear behavior, which is related
to the oxidation behavior.
There are many implications on the tribological behavior due to
the phase transformation that cause a high temperature sensitivity. The low hardness of austenite relative to martensite promotes
adhesive wear, associated with galling and a higher coefcient of
friction (COF), with a higher wear rate than for martensite [19,20].
This suggests that the wear rate would be lower for fretting at
temperatures below Md, however, only the surface material

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

transforms to martensite during fretting below Md, leading to a

hard surface layer being supported by a soft substrate which leads
to fracture and a high wear rate [21]. AISI 301 stainless steel in the
full hard condition has an initial martensite content of approximately 60% [22,23], which results in martensite presence in the
wear debris for fretting at temperatures above Md and more
support for the hardened surface layer at temperatures below Md.
The wear behavior is further complicated by a transition in the
oxide composition and morphology that occurs with changing
temperature. A wear-resistant glaze oxide is found to form at
elevated temperature with a transition in oxide composition from
Fe2O3 at 20 C to Fe3O4 at approximately 200 C, which drastically
decreases the wear rate [9,24]. Thus transitions occur in the
mechanical properties and oxide properties over the same temperature range between 20 C and 250 C which each affect the
fretting damage behavior. There are currently no studies available
in the open literature which address the inuence of these changes on the level of fatigue damage due to fretting.
This work addresses the changes in the fretting damage of austenitic stainless steel sheets caused by changing temperature with a
focus on the change in fatigue damage due to an increase of temperature from 20 C to 250 C. Changes in the inuential factors are
identied as either a change in the driver for damage or a change in
the material resistance to wear and fatigue through plain fatigue
tests, fretting tests, and nite element modeling. The net effect is
determined and discussed in terms of the role of each factor.

2. Materials and method

2.1. Materials
The materials used in this study were sheets of AISI 301
stainless steel in the full hard condition and bars of AISI 52100
steel in the annealed condition. The chemical composition of these
materials is shown in Table 1. The 301 stainless steel was received
with a rolling reduction of approximately 50%. Specimens were
machined into at dog-bone specimens with a 76.2 mm length,
12.7 mm tab section width, 4.762 mm gage section width, and
0.205 mm thickness with the length of the specimen aligned
parallel to the direction of rolling. The surface of the specimens is
left in the as-rolled condition, which corresponded to an Ra of
0.10 m. The edges of the specimens were ground with P300 grit
SiC paper to remove approximately 50 m from each edge to
remove material affected by machining. The edges were then
polished with P4000 grit SiC paper in the longitudinal direction to
eliminate surface roughness as a cause of fatigue crack formation
at the edge.
Specimens of 52100 steel were machined into cylinders with a
diameter of 20.0 mm and a length of 25 mm. Specimens were
polished to the nal diameter using P1200 SiC paper. The hardness
was measured as 20 RHB at 20 C.

2.2. Experimental method

Tensile tests were performed on specimens of 301 stainless
steel at temperatures from 20 C to 800 C to determine the
change in mechanical properties as a function of temperature. The
details of the tests are provided in previous work [9]. The modulus
of 52100 steel was measured as a function of temperature for use
in nite element modeling. The modulus of annealed 304 stainless
steel, which was used for the fretting xture, was also measured as
a function of temperature for use in nite element modeling.
Plain fatigue tests were performed on the 301 stainless steel
specimens at 20 C and 250 C to determine the baseline fatigue
performance at these temperatures. Specimens were tested in
force control at 10 Hz with a force ratio of R 0.05 in an oven.
Various stress amplitudes were applied to generate stress-life
curves, with a maximum test duration of 106 cycles.
Fretting experiments were performed for contact between 301
stainless steel specimens and 52100 steel specimens over a range
of normal forces from 100 N to 375 N, displacement amplitudes
from 10 m to 200 m, atmospheres of air and Ar (0.3 70.2% O2),
and temperatures from 20 C to 550 C at 10 Hz using a unique
experimental conguration for fretting testing of thin sheets as
shown in Fig. 1. This setup allows for independent control of the
displacement amplitude, normal force, and temperature. PTFE was
used as the lubricant between the stainless steel specimens and
the specimen holder for tests performed at 20 C and 250 C, and a
BN based lubricant was used for tests performed at higher temperatures. Comparison of the results obtained using PTFE is not
compared to results obtained using BN in this work due to the
difference in response which may occur. A thermocouple was
welded to the clamping plate near the contact rather than to the
specimen to avoid damage to the specimen that would affect the
subsequent fatigue results. More details on the experimental
conguration are available in previous work [16,25]. Tests were
performed for 104 cycles, which was sufcient to reach a stabilized
friction response for approximately 90% of the test duration.
Damage characterization was performed to determine the level
of wear and fatigue damage resulting from each condition tested.
Wear was measured by stylus prolometry of the wear scars.
Three line scans were performed in the direction of fretting
motion at locations spaced evenly along the full width of the
fretting scar. Wear volumes were calculated by averaging the three
wear proles and multiplying by the scar width. Scars were also
inspected using optical microscopy and scanning electron microscopy to determine wear characteristics. The oxide composition
was measured using X-ray diffraction.
Fatigue damage due to fretting was characterized by performing uniaxial fatigue tests on samples that were subjected to fretting to determine the subsequent fatigue life. Subsequent fatigue
tests were performed at 20 C in force control at 10 Hz using a
force ratio of R 0.05 and a stress amplitude of 468 MPa, which
corresponds to room temperature fatigue limit. The subsequent
fatigue life indicates the extent of fatigue damage that occurred
during fretting, with a longer subsequent fatigue life corresponding to less fatigue damage during fretting and vice versa.

Table 1
Material composition in weight percent.
Element Weight Percent

AISI 301 Stainless Steel

AISI 52100 Steel

Mn

Cr

Si

Ni

Mo

Cu

Al

0.15
0.931.05

2.00
0.250.45

16.018.0
1.351.60

1.00
0.150.35

0.05
0.025 Max

0.03
0.015 Max

6.08.0
0.25 Max

0.10

0.10 Max

0.30 Max

0.0015 Max

0.050 Max

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

Fig. 1. Fretting test conguration.

Table 2
Elastic constants used for material models.
E (GPa)

Stationary specimen
Backing plate
Moving specimen
PTFE layer

Fig. 2. Finite element model geometry [16].

3. Modeling method
3.1. Model formulation
A nite element model of the experimental conguration was
created using ABAQUS/Standard to determine the local cyclic
stressstrain behavior and local frictional energy dissipation. The
geometry of the model is shown in Fig. 2. The model is twodimensional plane-strain and consists of three bodies with two
contact interactions. A Lagrangian contact formulation was used
for contact between the moving specimen and the stationary
specimen to enforce exact stick. A Penalty contact formulation was
used for contact between the stationary specimen and the PTFE
layer to aid convergence. A compliant layer was added to the sides
of the backing plate to account for the compliance of the test
conguration. The sides of the compliant layers and the bottom of
the backing plate were constrained against motion. Normal force
and displacement boundary conditions were applied to a reference
node in a rigid layer at the top of the moving specimen. A mesh
size of 2  2 m2 was used at the contact interface between the

20 C

250 C

20 C

250 C

167
195
210
0.5

142
182
200
0.03

0.30
0.30
0.30
0.46

0.30
0.30
0.30
0.46

moving specimen and stationary specimen, with the size

increasing with increasing distance from the contact. A total of
approximately 105 elements were used. This mesh size was
determined to be sufcient during a mesh convergence study
during verication using the Johnson model for sliding of Hertzian
contact [26].
Simulations were performed to represent experiments performed at 20 C and at 250 C. Simulations of experiments at 20 C
were performed using isotropic elastic material models. An elastic
approximation was found to be appropriate by analysis of the
maximum Von Mises stresses determined from the simulations.
Simulations of experiments performed at 250 C were performed
using isotropic elastic models for the materials other than the 301
stainless steel moving specimen, which was modeled using isotropic hardening to t the experimental stressstrain response
from a tensile test. The choice of hardening law was not critical
because the stressstrain behavior was nearly elastic-perfectly
plastic. The elastic constants used for each material at each temperature are shown in Table 2. The measured values of the elastic
moduli were used for materials other than PTFE. The values used
for PTFE were in agreement with the literature [27]. The modulus
of the compliant layers was determined by comparison of the
hysteresis loops calculated from the model to the hysteresis loops
measured in the corresponding fretting test, and iteration of the
compliant layer modulus until correspondence was achieved.
The COF for the contact between the moving specimen and the
stationary specimen was uniform over the surface and was specied as the value measured during the fretting test of the corresponding conditions. The COF for contact between the stationary

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

specimen and the PTFE used was 0.04 for 20 C and 0.03 for
250 C, which was in agreement with the literature [28].
Simulations were performed for three cycles, which was found
to be sufcient to stabilize the model response, which was
determined by a negligible difference in plastic strain between
cycles three and four for the most highly loaded test condition.
Model results for the third cycle were used for analysis.
3.2. Fretting damage modeling
The results from the model were used to calculate damage
parameters to determine the changes in damage drivers with
changing test conditions. The stressstrain and slip results from
the simulation of each test condition were exported from ABAQUS
and post-processed using MATLAB. The FatemiSocie (FS) criticalplane based multiaxial fatigue damage parameter [29] was used to
indicate the local driver for fatigue damage as a function of location over the model. The FS parameter was found in previous work
[16] to better correlate the fretting damage for the 20 C conditions investigated here than the SmithWatsonTopper parameter
[30]. The FS value was averaged over a 15 m radius, approximately equal to the grain size, to account for strain gradient sensitivity. The local energy dissipation per cycle, adapted from the
Ruiz Fretting Damage Parameter (FDP) [31], was used to indicate

Fig. 3. Plain fatigue response of 301 stainless steel (full hard) at 20 C and 250 C,
R 0.05.

the local driver for wear over the surface of the stationary
specimen.

4. Experimental results
4.1. Baseline material performance
The stress-life results for plain fatigue of 301 stainless steel at
20 C and 250 C are shown in Fig. 3. The increase in temperature
resulted in a decrease in the fatigue limit by approximately 510%.
The decrease in the fatigue strength was more signicant at lower
cycles to failure. The larger reduction in fatigue strength at lower
cycles to failure is attributed to the reduction in ductility which
occurs due to the increase in temperature. This reduction in ductility
between 20 C and 250 C was found to be 80% in previous work [9].
This has a more signicant effect for shorter lives because of the
larger plastic strain amplitude for the higher stress amplitudes.
4.2. Friction response
The COF evolution at 20 C and 250 C for various displacement
amplitudes and normal forces is shown in Fig. 4. The COFs were
obtained from the ratio of tangential to normal force measured in
the gross slip stage of the hysteresis loops. Tests conducted at
20 C had a low initial COF which increased to reach a maximum
value after 102103 cycles and then remained approximately
constant for the remaining duration of the test. For tests conducted at 250 C, the COF typically reached a maximum value early
in the test, at less than 10 cycles, before decreasing to a stabilized
value prior to reaching 102 cycles. Tests conducted in the mixed
slip regime had different contact condition evolution which is
related to this difference in COF evolution. At 20 C, a mixed slip
running condition consisted of an initial period of gross slip followed by a transition to stabilized partial slip. Conversely, a mixed
slip running condition at 250 C consisted of an initial period of
partial slip followed by a transition to stabilized gross slip. The
difference in behavior is related to the formation of a glaze oxide
on the surface in the tests conducted at elevated temperature. This
is discussed in a subsequent section.
The stabilized values of the maximum tangential force ratio
(TFR), dened as the ratio of the magnitude of the maximum tangential force during a cycle to the imposed normal force, for each
condition at 20 C and 250 C is shown in Fig. 5. The steady state
slip amplitude was determined as half of the width of the stabilized

Fig. 4. COF evolution for tests conducted at (a) 20 C and (b) 250 C.

10

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

Fig. 5. Steady state tangential force ratio versus steady state slip amplitude.

Fig. 6. : Accumulated dissipated energy at 104 cycles.

hysteresis loop at mid-height (tangential force equal to zero).

Representation using the steady state slip amplitude helps to
remove the effect of test system stiffness change with changing
temperature to make a more direct comparison. However, this does
not account for the decrease in contact pressure associated with the
increase in contact area caused by a decrease in material stiffness
with increasing temperature. The stabilized tangential force ratio
was higher for tests conducted at 20 C for each test condition.
The total energy dissipated during each test conducted at 20 C
and 250 C is shown in Fig. 6. This was calculated based on the
area of the stabilized hysteresis loop, which represented at least
90% of the test hysteresis loops for each test. Therefore, this result
includes energy dissipated by means other than friction between
the moving specimen and stationary specimen (e.g. friction
between the stationary specimen and PTFE), necessitating nite
element modeling to determine the local frictional energy dissipation. The amount of dissipated energy was higher for tests
performed at 20 C. This is a result of the higher tangential force at
20 C. The decrease in COF at elevated temperature decreases the
tangential force, but also increases the amount of slip. This tends
to offset the difference in dissipated energy.
4.3. Wear behavior
The wear volume for each specimen resulting from fretting
with a normal force of 255 N and a displacement amplitude of

Fig. 7. Wear volume versus temperature including results from the literature for
fretting of annealed 304 stainless steel against itself [24].

200 m as a function of temperature is shown in Fig. 7. A signicant decrease in the wear rate was found to occur between
20 C and 250 C. This is in agreement with results for fretting of
annealed 304 stainless steel against itself [24]. This is attributed to
a transition in the oxide composition from Fe2O3 to Fe3O4, which
was identied via X-ray diffraction (XRD) in our earlier work [9].
Fretting at elevated temperature resulted in the formation of a
glaze oxide layer that resulted in the decrease in wear, partially
due to a decrease in the COF. The measured increase in total
volume (negative wear volume) is attributed to impacted iron
oxide, which has a higher volume than iron [24]. The decrease in
the wear volume was greater than the decrease in the amount of
dissipated energy at elevated temperature, thus the wear resistance of the material increased.
Cross-sections of 301 stainless steel subjected to fretting at 20 C
and 550 C are shown in Fig. 8. The specimen subjected to fretting at
20 C was etched to show the phase distribution. A layer of martensite (dark phase) was formed at the surface due to fretting.
Martensite is also present throughout the sample due to prior cold
work. The glaze oxide layer responsible for the decreased friction
and wear rate was observed on the sample subjected to fretting at
550 C. Of the fretting scars and cross-sections inspected by optical
and scanning electron microscopy, no clearly distinguishable cracks
were detected in the microstructurally-altered volumes where
fretting had occurred.
4.4. Fatigue damage due to fretting
The subsequent fatigue lives for specimens subjected to fretting
using a normal force of 255 N at 20 C and 250 C are shown in
Fig. 9. Fatigue tests were conducted on the fretting damaged
specimens at a stress amplitude of 468 MPa, which is 85% of the
room temperature fatigue strength, both based on R0.05. Fatigue
damage due to fretting was found to be more severe at 250 C than
at 20 C for displacement amplitudes below 60 m. The fatigue
damage due to fretting was found to be less severe for tests conducted at 250 C than at 20 C for displacement amplitudes of
60 m and above. The 20 C results do not seem to demonstrate
the expected benecial effects associated with high wear at the
higher displacement amplitudes as typically shown in a Vingsbo
plot [32]. This is due to the signicant wear depth for the 20 C
tests which causes a reduction in the cross-section of the specimens by as much as 20%. This causes a decrease in the subsequent
fatigue life due to the increased stresses and appears as an
increase in the level of fatigue damage, however it is an artifact

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

11

Fig. 8. Subsurface structure from fretting at (a) 20 C and (b) 550 C [9] (reprinted with permission from Elsevier).

needs to be analyzed further. Since no cracks could be readily

detected after fretting, a detailed investigation of the subsequent
crack nucleation and propagation during the fatigue test is needed
but is beyond the scope of the current project.

5. Modeling results

Fig. 9. Uniaxial fatigue life after fretting with a 255 N normal force.

Fig. 10. Subsequent fatigue life versus the steady state maximum tangential force.

which only occurs for specimens which are thin relative to the
wear depth [16].
The subsequent fatigue lives as a function of the maximum
steady state tangential force is shown in Fig. 10. It would be
expected that if wear did not have a signicant effect on the level
of fatigue damage due to fretting then the maximum steady state
tangential force would correlate well with the level of fatigue
damage due to fretting [16]. There is not a strong correlation, thus
suggesting that the wear behavior may play an important role in
the level of fatigue damage due to fretting for these conditions and

An example of the output of the accumulation of frictional

energy dissipation at the surface of the stationary specimen
throughout a cycle is shown in Fig. 11. The total energy dissipated
per cycle is equal to the area under the curve at the end of the
cycle. The dissipated energy distribution has a maximum value in
the center of the contact for the gross slip case. This is expected
and is the reason for the typical "U" shaped wear prole for gross
slip conditions. The dissipated energy distribution for the partial
slip condition is maximum near the edges of contact, which is the
only portion of the contact that undergoes slip. This distribution
leads to the typical "W" shaped wear prole.
The total frictional energy dissipation on the surface of the
stationary specimen determined from the model for a normal
force of 255 N is shown in Fig. 12(a). The total energy dissipation is
higher for the tests conducted at 20 C for displacement amplitudes above 40 m. Representation in terms of the energy dissipated per area of material over which the energy was dissipated
(contact width plus displacement range) shows a larger difference
between the results at 20 C and 250 C as shown in Fig. 12(b). This
is due to the larger contact area at 250 C resulting from a decrease
in the modulus of the materials. Having a similar total dissipated
energy would suggest a similar total wear volume if the material
resistance to wear was unchanged. The greater difference in the
dissipated energy density suggests that the difference in the wear
depth would be greater than the difference in the total wear
volume, since the similar total amount of energy is concentrated
on a smaller area. This would lead to a more signicant decrease in
the subsequent fatigue lives for high wear cases due to the
increased reduction in thickness, and would lead to an increase in
subsequent fatigue lives for lower wear cases where the additional
wear could increase the benecial effects.
The maximum FS values calculated from the model for each
simulation are shown in Fig. 13. The max FS values for a normal
force of 255 N are higher for the 20 C cases than for the 250 C
cases for displacement amplitudes up to 150 m. This is partly due
to the increased contact area at elevated temperature due to the
decrease in modulus, and partly due to the difference in the COF
values which are the closest in value at the highest displacement
amplitudes. The max FS values at 250 C calculated for a 100 N
normal force were higher than or similar to the values for the
255 N normal force. This was not expected since the tangential
force was higher for the 255 N case. Similarly, the max FS values
for the 255 N normal force were similar to or higher than the

12

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

Fig. 11. Frictional energy dissipation accumulation for a (a) partial slip condition and (b) gross slip condition.

Fig. 12. Frictional energy dissipation per cycle on the surface of the stationary specimen in terms of (a) the total dissipation and (b) dissipation per unit contact area.

Fig. 13. Maximum FatemiSocie values.

values for the 375 N normal force. The increased contact area
resulting from the higher normal force causes a reduction in the
pressure, which is enhanced by the presence of the compliant
PTFE layer, helps explain this result.
A comparison of the model prediction for fatigue damage and
the experimental subsequent fatigue lives is shown in Fig. 14. The
subsequent fatigue results are presented as the inverse of subsequent life so that it is proportional to fatigue damage. The

prediction for the level of fatigue damage was fairly consistent for
the tests conducted at 20 C for all displacements. It was expected
that the actual fatigue damage would be more severe than the
prediction at high amplitudes because of the reduction in thickness, rather than the expected result for a thick specimen which
would have a higher damage than predicted due to the benecial
effects of wear. This suggests that the detrimental effect of thickness reduction and the benecial effects of material removal
nearly offset one another.
The actual level of fatigue damage at 250 C was relatively
higher than predicted by the model, especially at the lower displacement amplitudes. This is due to the reduction in fatigue
strength of the material at elevated temperature. At higher displacements, the predicted level of damage increases because the
COF values increase with increasing amplitude. However, the
actual level of fatigue damage did not continue to increase with
increasing displacement. The glaze oxide layer which is more
prominent at higher displacements may have provided a protective effect.
6. Discussion
The changes that occur with changing temperature can be
categorized as being either a change in material resistance or a

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

Fig. 14. Maximum FS values and experimental subsequent fatigue results.

change in the driver for fatigue damage. The combination of the

various changes results in the net effect on the fretting damage
response. The increase in temperature causes a decrease in the
ductility of 301 stainless steel due to the stabilization of austenite,
which prohibits the transformation to martensite which is
responsible for the increased ductility. The increase in temperature
does not have as signicant of an effect on the strength. The effects
are evident from the difference in the plain fatigue results at 20 C
and 250 C. The reduction in ductility has a larger affect on the
higher amplitude lives, which were found to have a larger difference than for the lower amplitude cases. Thus the material resistance to fatigue damage was decreased by the increase in
temperature.
The increase in temperature causes a change in the oxide
composition and morphology which affects the COF response and
the wear behavior, which are closely related. The reduction in the
COF due to the formation of the glaze oxide caused a decrease in
the dissipated energy, and thus a decrease in the driver for wear.
Analysis of the energy dissipation and wear volume determined
that the reduction in wear was not only related to a reduction in
the driver for wear but also an increase in the material resistance
to wear, since the volume of material removed per unit dissipated
energy decreased. Thus the driver for wear was reduced and the
material resistance to wear was increased.
The decrease in COF due to the formation of the glaze oxide
also results in a decrease in the driver for fatigue damage due to
the reduction in the cyclic stress state. The increase in contact area
identied by the model due to the decreased modulus further
decreases the contact stresses and the driver for fatigue damage.
Thus the increase in temperature reduced the driver for fatigue
damage.
The formation of the glaze oxide removed the benecial effects
of wear, however it also provided a protective effect which
resulted in a net increase in the material resistance to fatigue
damage resistance.

7. Conclusions
The inuence of temperature on the fatigue damage in 301
stainless steel sheets due to fretting against 52100 steel was
investigated. Using experimental observations and measurements
coupled with nite element modeling, the drivers for damage and
changes in the material resistance as temperature are increased
from 20 C to 250 C were elucidated:
1. The resistance of the 301 stainless steel to wear improved at the
higher temperature due to the formation of a glaze oxide layer.
The glaze oxide layer formed as a result of a change in the oxide
composition from Fe2O3 to Fe3O4.

13

2. The driver for wear decreased due to the reduction of coefcient

of friction (COF) caused by the formation of the glaze oxide
layer. As a result, the energy dissipated by friction on the surface
of the stationary specimen was reduced.
3. However, the resistance to fatigue damage was reduced at the
higher temperature. This was attributed to the reduction in
ductility of 301 stainless steel caused by the increased stability
of the austenite and the absence of the additional deformation
mechanism associated with transformation of austenite to
martensite.
4. In contrast, the driver for fatigue damage was reduced. This was
attributed to both the lower contact stresses from the reduction
in the elastic modulus as temperature increases and the
decrease in the tangential force from the reduction of the COF
due to the formation of the glaze oxide layer. The potency of the
fatigue driver can be described by the FatemiSocie parameter
determined via a nite element model of the fretting contact
using the temperature-dependent material properties.

Acknowledgments
This project was conducted with the nancial support of the
Victor-Reinz Division of Dana Holding Corporation.

References
[1] J.M. Dobromirski, Variables of fretting process: are there 50 of them? in:
H.M. Attia, R.B. Waterhouse (Eds.), Standardization of Fretting Fatigue Test
Methods and Equipment, ASTM STP 1159, American Society for Testing and
Materials, Philadelphia, 1992, pp. 6066.
[2] Z. Tourki, H. Bargui, H. Sidhom, The kinetic of induced martensitic formation
and its effect on forming limit curves in the AISI 304 stainless steel, J. Mater.
Process. Technol. 166 (2005) 330336.
[3] A. Grifths, J. Wright, Mechanical properties of austenitic and metastable
stainless steel sheet and their relationships with pressforming behaviour, Iron
and Steel Institute, Publication 117 (1969) 5165.
[4] P. Maxwell, A. Goldberg, J. Shyne, Inuence of martensite formed during
deformation on the mechanical behavior of FeNiC alloys. Metallurgical and,
Mater. Trans. B 5 (1974) 13191324.
[5] H. Mughrabi, H.J. Christ, Cyclic deformation and fatigue of selected ferritic and
austenitic steels: specic aspects, ISIJ Int. 37 (12) (1997) 11541169.
[6] G. Olson, M. Azrin, Transformation behavior of TRIP steels, Metall. Mater.
Trans. A 9 (1978) 713721.
[7] G. Baudry, A. Pineau, Inuence of strain-induced martensitic transformation
on the low-cycle fatigue behavior of a stainless steel, Mater. Sci. Eng. 28 (1977)
229242.
[8] A. Pineau, R. Pelloux, Inuence of strain-induced martensitic transformations
on fatigue crack growth rates in stainless steels, Metall. Mater. Trans. B 5
(1974) 11031112.
[9] M.R. Hirsch, R.W. Neu, Inuence of temperature on the fretting response
between AISI 301 stainless steel and AISI 52100 steel, Tribol. Int. 68 (2013)
7784.
[10] S. Malkin, D.P. Majors, T.H. Courtney, Surface effects during fretting fatigue of
Ti6Al4V, Wear 22 (1972) 235244.
[11] S.C. Gordelier, T.C. Chivers, A literature review of palliatives for fretting fatigue,
Wear 56 (1979) 177190.
[12] R. Magaziner, O. Jin, S. Mall, Slip regime explanation of observed size effects in
fretting, Wear 257 (2004) 190197.
[13] K. Nishioka, K. Hirakawa, Fundamental investigations of fretting fatigue IV. The
effect of mean stress, Bull. Jpn. Soc. Mech. Eng. 12 (1969) 408414.
[14] J.J. Madge, S.B. Leen, I.R. McColl, P.H. Shipway, Contact-evolution based prediction of fretting fatigue life: effect of slip amplitude, Wear 262 (2007)
11591170.
[15] O. Jin, S. Mall, Effects of slip on fretting behavior: experiments and analyses,
Wear 256 (2004) 671684.
[16] M.R. Hirsch, R.W. Neu, Fretting damage in thin sheets: analysis of an experimental conguration, Tribol. Int. 44 (2011) 15031510.
[17] A. Iwabuchi, K. Kato, K. Hokkirigawa, T. Suzuki, Fretting damage of SUS304 in a
vacuum (Discussion of the condition of crack generation, JSME Int. J. 30 (1987)
13191325.
[18] K. Kato, T. Suzuki, A. Iwabuchi, K. Hokkirigawa, Fretting Damage of SUS 304 in
a Vacuum, JSME Int. J. 30 (1987) 330336.
[19] W. Wang, M. Hua, X. Wei, Friction behavior of SUS 304 metastable austenitic
stainless steel sheet against DC 53 die under the condition of friction coupling
plastic deformation, Wear 271 (2011) 11661173.

14

M.R. Hirsch, R.W. Neu / Wear 346-347 (2016) 614

[20] Z.Y. Yang, M.G.S. Naylor, D.A. Rigney, Sliding wear of 304 and 310 stainless
steels, Wear 105 (1985) 7386.
[21] K.L. Hsu, T.M. Ahn, D.A. Rigney, Friction, wear and microstructure of unlubricated austenitic stainless steels, Wear 60 (1980) 1337.
[22] J. Luksza, M. Ruminski, W. Ratuszek, M. Blicharski, Texture evolution and
variations of -phase volume fraction in cold-rolled AISI 301 steel strip,
J. Mater. Process. Technol. 177 (2006) 555560.
[23] P. Marshall, Austenitic Stainless Steels-microstructure and Mechanical Properties, Elsevier Applied Science, London; New York, 1984.
[24] T. Kayaba, A. Iwabuchi, The fretting wear of 0.45% C steel and austenitic
stainless steel from 20 to 650 C in air, Wear 74 (1981) 229245.
[25] M.R. Hirsch, R.W. Neu, Fretting behaviour of AISI 301 stainless steel sheet in
full hard condition in contact with AISI 52100 steel, Tribol.-Mater. Surf.
Interfaces 2 (2008) 39.
[26] K.L. Johnson, Contact Mechanics, Cambridge University Press, Cambridge,
New York, 1985.

[27] Z. Zhang, X. Chen, Y. Wang, Uniaxial ratcheting behavior of polytetrauoroethylene at elevated temperature, Polym. Test. 29 (2010) 352357.
[28] B.J. Briscoe, Y. Lin Heng, T.A. Stolarski, The friction and wear of poly(tetrauoroethylene)-poly (etheretherketone) composites: An initial appraisal of
the optimum composition, Wear 108 (1986) 357374.
[29] A. Fatemi, D.F. Socie, A critical plane approach to multiaxial fatigue damage
including out-of-phase loading, Fatigue Fract. Eng. Mater. Struct. 11 (1988)
149165.
[30] K.N. Smith, P. Watson, T.H. Topper, A stressstrain function for the fatigue of
metals, J. Mater. 5 (1970) 767778.
[31] C. Ruiz, P. Boddington, K. Chen, An investigation of fatigue and fretting in a
dovetail joint, Exp. Mech. 24 (1984) 208217.
[32] O. Vingsbo, S. Soderberg, On fretting maps, Wear 126 (1988) 131147.