MASTER'S THESIS
Tribological Performance of PTFE Composites at Breakaway in Sliding
Lubricated Contacts
Arash Golchin
Master of Science Mechanical Engineering
Luleå University of Technology
Department of Engineering Sciences and Mathematics
Tribological Performance of PTFE Composites at Breakaway in Sliding Lubricated Contacts
Arash Golchin
Luleå University of Technology
Department of Applied Physics and Mechanical Engineering
Division of Machine Elements
31 March 2010
iii
A BSTRACT
Babbitt has long been used as the lining material in hydrodynamic sliding bearings. However, with the new operating conditions imposed on hydro-electric power plants due to grid frequency regulation, the power plants undergo more frequent starts and stops which increases the need for an alternative material with better friction characteristics at start-up compared to conventionally used white-metal (Babbitt). Polytetrafluoroethylene (PTFE) based materials have potential to provide significant improvements in hydrodynamic sliding bearings through their compliant and breakaway characteristics under loading. However, while pure PTFE can provide excellent performance, it is somewhat limited in extreme loading situations. This study is therefore aimed at investigating the tribological characteristics at the initiation of sliding (breakaway friction) of several polytetrafluoroethylene (PTFE) based materials including virgin PTFE, PTFE filled with 25% black glass, PTFE filled with 40% bronze, PTFE filled with 25% carbon, and PTFE filled with 20% glass fibre and 5% Molybdenum disulphide, as well as standard Babbitt material in lubricated sliding contact with a steel counter-face. Experiments were carried out using a reciprocating tribo-meter in the block on plate configuration with the specific goal of determining the friction characteristics at breakaway under various conditions. Apparent contact pressures of 1 to 8 MPa were applied with oil temperature levels of 25° to 85°C.Bronze- and carbon-filled PTFE and virgin PTFE were found to provide generally lower and more stable breakaway friction over the course of testing than the other materials. Breakaway friction tests after an extended stop under loading showed a maximum change in breakaway friction of 0.07 for bronze filled, carbon filled, and virgin PTFE whereas Babbitt produced an increase of 0.32 in the first cycle after the extended stop, reaching the friction level of more than three times of those of PTFE based composites. Breakaway friction for the four tested materials after an extended stop returned to the pre-stop values after 1 stroke. The effect of materials on the steel counter face was also examined using an optical profilometer finding that only glass filled composites (black glass filled PTFE, and glass fiber and MoS2 filled PTFE) produced significant polishing of the steel surface under high loads. No measurable polishing was detected for other tested materials.
v
P REFACE
The research presented in this thesis has been carried out as part of a project in conjunction with the Swedish Hydropower Center – SVC. SVC has been established by the Swedish Energy Agency, Elforsk and SvenskaKraftnät in partnership with academic institutions. This thesis is based upon studies conducted during April 2009 to March 2010 at the Department of Applied Physics and Mechanical Engineering, Division of Machine Elements at Luleå University of Technology, Sweden.
Conducting research in a world class laboratory as Tribolab was a unique opportunity for me to learn how to perform high quality laboratory work. The latter could not be achieved without the assistance and positive attitude of every one helping me through this process. I would like to express my sincere gratitude to my leading supervisor Professor Sergei Glavatskih, whose enthusiasm, knowledge, and aim at perfection and honest science will never stop amazing me.
His support, criticism, and suggestions during this work are greatly appreciated. Further I would like to thank my co-supervisor Mr. Gregory Simmons for his great co-operation and help.
Without his advice and unique support this thesis would never have become a reality.
Finally, I wish to express my greatest thanks to my wife Nassim for her support and patience during my studies. I also wish to thank my parents for all the support and showing me the value of hard work.
Luleå, 31 March 2010
Arash Golchin
vii
C ONTENT
1: Introduction ... 1
1.1. Hydrodynamic Bearings ... 2
1.1.1.Journal Bearings ... 2
1.1.2.Tilting Pad Journal Bearings ... 3
1.1.3. Tilting Pad Thrust Bearings ... 5
2: Literature Review ... 7
3: Methodology ... 11
3.1. Choice of the Test Rig ... 11
3.2. Initial Test Configurations ... 12
3.2.1. Steel Block on Polymer Plate Configuration ... 12
3.2.2. Steel Cylinder on Polymer Plate Configuration ... 14
3.2.3. Polymer Block on Steel Plate Configuration ... 15
3.3. Frequency and Stroke Length ... 16
3.4. Breakaway Friction ... 16
4: Experimental Setup ... 19
4.1. Materials ... 19
4.2. Instrumentation ... 20
4.3. Test Specimens ... 21
4.4. Test Conditions ... 21
4.4.1. Pressure and Temperature Effect Tests ... 22
4.4.2. High Pressure-High Temperature Tests ... 22
4.4.3. Prolonged Stop and Restart ... 23
5: Results ... 25
5.1. Effect of Contact Pressure ... 25
5.2. Effect of Temperature ... 28
5.3. High Pressure – High Temperature Tests ... 31
5.4. Prolonged Stop and Restart ... 32
5.5. Wear of the Steel Counter-surface ... 33
5.6. SEM investigation of block specimens ... 35
5.6.1 Babbitt ... 35
5.6.2 Black-glass filled PTFE ... 35
5.6.3 Fiberglass and MoS2 filled PTFE ... 36
5.6.4 Bronze filled PTFE ... 37
5.6.5 Carbon filled PTFE ... 37
5.6.6 Pure PTFE... 38
6: Discussion ... 39
7: Conclusions... 43
Future Work ... 45
References ... 47
Appendix A ... 49
Appendix B ... 51
Appendix C ... 55
1
1
Introduction
With the strong push towards production of more renewable energy such as wind, wave, solar and tidal energy which are intermittent in nature, the role of hydropower has altered from base electricity production to more participation in regulation and stabilizing the grid frequency through primary and secondary regulation. Primary regulation in hydropower plants is done by automatic adjustment of guide vane angle while secondary regulation is done by start-up and shut-down of the whole hydroelectric power plant which causes turbines to start and stop more frequently than they used to. The new operating conditions imposed on power plants have great impact on mechanical components especially hydro-dynamically lubricated bearings which are susceptible to damage from frequent starts and stops of the turbine shaft. In steady state operation, the shaft and bearing surfaces are separated by a hydro-dynamic lubricant film, providing low friction and long service life of the mating components. During the transient phases of start-ups and stops, a contact between shaft and bearing surfaces is inevitable resulting in wear. Considerable start-up torque is required to overcome the breakaway friction in a steel- Babbitt contact.
To prevent bearing wear and reduce start-up friction, a hydraulic jacking system may be employed. This provides a hydrostatically pressurized oil film between the shaft and bearing surfaces.
In an earlier study [1] conducted on some PTFE based composites it was shown that these composites have great potential to replace the conventionally used Babbitt faced bearings.
2 Introduction PTFE based composites may simplify the start-up of large turbines while at the same time have the potential to eliminate the need for complex hydraulic jacking systems currently used in hydropower stations.
Therefore, this study is aimed at investigation of breakaway friction of some commercially available PTFE based composites under lubrication of synthetic ester turbine oil at a wide range of operating pressures and temperatures.
The choice of the tested materials is based on their ready availability and previous works on these materials [1]. The results are compared with those for pure PTFE and Babbitt material.
1.1. Hydrodynamic Bearings
Hydrodynamic bearings support a rotating shaft by carrying the shaft on a hydrodynamic fluid film. The rotating shaft pumps the lubricant into the converging gap between the surfaces, creating a fluid wedge which results in a full hydrodynamic fluid film at certain operating conditions.
The hydrodynamic fluid film makes it possible for this type of bearings to operate at high loading with very low friction. Hydrodynamic bearings which are utilized in hydropower plants are journal bearings, tilting pad journal bearings, and tilting pad thrust bearings.
1.1.1. Journal Bearings
A journal bearing is comprised of a sleeve or shell (Fig. 1.1) supporting a rotating shaft or journal. Journal bearings experience much lower contact pressures compared to rolling element bearings (often lower by three orders of magnitude). This is due to the larger contact area created by the conforming surfaces of the journal and the bearing.
Introduction
Figure (1.1) Plain journal bearing sleeves
The mean pressure in the load supporting area of a journal bearing is determined as the weight or load supported by the bearing divided by the
diameter multiplied by the length of the bearing).
Babbitt (also referred to as white metal) is used as a lining material in hydro shaft damage, due to its lower hardness into contact with the shaft during start
Babbitt has high embedability and can take care of the particles in them below the contact surface of the bearing
particles can cause. However, it mixed and boundary lubrication
undergo frequent start-ups and stops of the shaft.
1.1.2. Tilting Pad Journal Bearing
Tilting pad journal bearings consist of several pads
tilt to conform to the shaft which is influenced by dynamic loads from the lubricant. Tilting pad
ure (1.1) Plain journal bearing sleeves [www.ibuonline.com]
The mean pressure in the load supporting area of a journal bearing is determined as the weight or load supported by the bearing divided by the projected load area of the bearing (the bearing diameter multiplied by the length of the bearing).
eferred to as white metal) is an amalgam of tin, lead, copper and antimony which hydrodynamically lubricated journal bearings. Babbitt
due to its lower hardness compared to steel, in the event that th during start-ups and shut-downs.
Babbitt has high embedability and can take care of the particles in the contact region and embed them below the contact surface of the bearing, thus protecting the shaft from
. However, it has inadequate frictional characteristics when operating in the and boundary lubrication which makes Babbitt less desirable material for bearings
ups and stops of the shaft.
Tilting Pad Journal Bearings
Tilting pad journal bearings consist of several pads as shown in Figure (1.2).The pads
which is influenced by dynamic loads from the lubricant. Tilting pad 3
The mean pressure in the load supporting area of a journal bearing is determined as the weight or load area of the bearing (the bearing
tin, lead, copper and antimony which lubricated journal bearings. Babbitt prevents the bearing comes
contact region and embed the shaft from damage these frictional characteristics when operating in the material for bearings that
.The pads are free to which is influenced by dynamic loads from the lubricant. Tilting pad
4
journal bearings are made with different number
the load exerted either on a pad or between the pads as sh
Figure (1.2) Tilting pad journal bearing
Figure (1.3) Different
a) Load on pad configuration
different numbers of tilting pads arranged in configurations with the load exerted either on a pad or between the pads as shown in Figure (1.3).
Figure (1.2) Tilting pad journal bearing [www.marunda.com.sg]
Different load configurations of tilting pad journal bearings.
pad configuration, b) Load between pads configuration
Introduction of tilting pads arranged in configurations with
Introduction 5 Tilting pad journal bearings are more complex and more expensive compared to the plain journal bearings; however these bearings are widely incorporated in rotating machines due to their higher stability compared to plain hydrodynamic journal bearings.
1.1.3. Tilting Pad Thrust Bearings
A thrust bearing is designed to support axial load of a rotating shaft and transfer it to the machine foundation. The bearing is comprised of several pads supported in a carrier ring (Fig. 1.4). Each pad can tilt and adjust its angular position so that a self-sustaining hydrodynamic fluid film is formed.
Figure (1.4) Tilting pad thrust bearing [www.marunda.com.sg]
Thrust load, shaft rotational speed, lubricant viscosity, incorporation of a hydraulic jacking system, shaft/bearing materials, shaft’s diameter through the bearing etc., determine the design and geometry of a thrust bearing to be employed for a specific application. Thrust bearings are often integrated with hydraulic jacking systems to ensure that an appropriate fluid film exists between contact surfaces of thrust runner and bearing pads at start-up.
6 Introduction
7
2
Literature Review
It is well known that polytetrafluoroethylene (PTFE) exhibits low friction coefficient in dry sliding conditions [3, 4] However it has the highest wear rate among the crystalline polymers [5].
In [6-8] it is shown that special fillers can considerably reduce the wear rate of polymeric composites in dry sliding conditions. However; the choice of type of filler and filler content should be determined through experiments because the mechanism of filler action in increasing the wear resistance is not well understood. There are many other factors such as the shape, size, and aspect ratio of the filler particles as well as the relative hardness of the filler and counter face materials that complicate the situation further. There are several explanations for the action of fillers in reducing the wear rate. One of them is based upon observation of excessive filler concentration on the composite surface after prolonged sliding, suggesting that the load is mainly supported by the fillers, resulting in reduced wear rate of the composite. Another explanation suggests that fillers in some way improve the adhesion of the transfer film to the counter-surface and thus increase the wear resistance of the polymers [9].
Research has also been conducted on tribological properties of bronze/carbon filled PTFE dry journal bearings, showing that the friction coefficient and wear rate are strongly influenced by bearing pressure, temperature, sliding distance and speed. It was found that the friction coefficient decreases as bearing pressure increases as a result of a more continuous dispersion, smoother and thicker transfer film. It was also found that friction increases with temperature [10, 11] resulting in prevention of PTFE composite films from forming on the surface by causing the filler material to become uncovered [12].
In another investigation of wear of fiber reinforced polymer composites, it was found that wear rate is dependent on fiber orientation. Wear rate of polymers with fiber orientation normal to the
8 Literature Review contact surface was found to be significantly higher than that of polymers with fibers parallel or vertical to the sliding direction [6].
Work has also been accomplished to investigate the effect of carbon aspect ratio on the friction and wear of PTFE-based composites, finding that the friction coefficient is not affected by the carbon aspect ratio but wear rate is. It was determined that the composites filled with oleophilic graphite which has a relatively high aspect ratio, wear at a rate which is about eight times greater than that for composites filled with either polar or nuclear graphite which have a relatively low aspect ratio [7].
Further study on the effect of tribochemical reaction of PTFE transferred film with steel, copper and aluminum substrates on its wear behavior revealed that although tribochemical reaction of PTFE transferred film with substrates was detected, the reactivity had no effect on wear rate of PTFE, indicating that PTFE transferred films are detached from the middle layer of the films and not from the interface between the film and substrate during the rubbing process [13].
In an investigation of tribological properties of CuS and graphite filled PTFE composites, it was found that graphite filler reduced the wear rate of PTFE by two orders of magnitude while it increased the friction coefficient by about 30%, in contrast CuS provided an equally large reduction in wear rate but did not increase the friction coefficient [9].
Although a lot of research has been carried out regarding friction and wear of PTFE-based composites in dry sliding conditions [9, 13, 14], little work has been done under lubricated conditions [15,16]. In [3] the friction coefficient of PTFE + 15% SiO2 + 5% MoS2 was investigated in lubricated conditions under various loads against steel counter faces with different surface roughness, finding that a rough mating surface resulted in higher friction and higher wear rate of the composite compared to that with a smooth mating surface. In this study it was further demonstrated that the friction coefficient decreased with increasing load until a certain level of load was reached over which friction increased as load increased. This is explained to be due to higher contact temperatures at high loads resulting in less contribution of lubricant in reduction of friction coefficient [3].
Literature Review 9 In [7] friction and wear properties of metal powder filled PTFE were studied under oil lubricated conditions. It was found that Cu, Pb, and Ni fillers increase the load carrying capacity of the PTFE composites, and also improve transfer of the PTFE composites onto the counter faces, thus reducing the wear rate of the PTFE composites. It was further demonstrated that friction of PTFE composites in liquid paraffin can be decreased by one order of magnitude and the wear rate by 1 or 2 orders of magnitude compared to dry conditions [17].
These studies, however, characterize the kinematic friction coefficient of tested material combinations during continuous motion at mixed or in some cases full film lubrication [3, 15, 17]. In hydrodynamic bearing applications, the tribological behaviour of mating materials at initiation of relative motion (breakaway) is also of interest. In an earlier study on behaviour of plain hydrodynamic journal bearings during starting and stopping [2], it was found that negligible wear takes place during the shut-down phase due to the fact that the shaft did not contact the bearing until some time after it stopped. It was further demonstrated that during the start-up phase the contact time of shaft and bearing surfaces is very short, of an order of a small fraction of a second. Although the contact time is very short; most of the wear occurs during this phase due to the asperity contact of the mating surfaces.
Although some studies have been carried out on breakaway friction of Babbitt faced bearings [18], very little work has been done on breakaway friction characteristics of PTFE based composites, which have great potential to replace Babbitt as a lining material in hydrodynamic sliding bearings.
The purpose of this study is to investigate the tribological performance at initiation of relative motion (breakaway) of some commercially available PTFE based composites, under start-up conditions typical for hydrodynamic sliding bearings.
Experiments include testing of PTFE based composites along with pure PTFE and white metal (Babbitt) over a wide range of contact pressures and oil temperatures. The effect of mating materials left under loading for an extended period of contact time, such as in the case of a machine stop, is also studied.
10 Literature Review
11
3
Methodology
3.1. Choice of the Test Rig
Friction and wear rate are not material properties but depend on the mating materials and operating conditions. Precise tribological behavior of mating materials may be determined in field tests by the use of real machine components. However, since field tests are very expensive and time consuming; tribo-meters are employed to investigate the relative tribological behavior of different material combinations.
Test conditions should approach the real working environment as much as possible, including the micro and macro structure of the material sliding couple, contact geometry, contact pressures, type of sliding motion, etc. [19]. Most of the experiments conducted for tribological testing of polymeric materials under lubricated conditions are performed using a block-on-ring test rig.
Although this configuration may seem to comply with the continuous sliding motion and overall macro configuration found in journal bearings; it was not used in this study due to some drawbacks explained in the next section.
In a block-on-ring configuration, the contact area gradually grows due to wear of the block, resulting in gradual decrease in contact pressure. In contrast, a flat-on-flat configuration provides constant contact area and pressure over the entire testing time [19]. Besides, a converging gap found in a block-on-ring contact, might affect friction due to build-up of hydrodynamic pressure.
In addition, such block-on-ring test configuration is “line contact” which differs from the conformal contact found in large journal bearings and thrust bearings.
12 Methodology A Cameron-Plint reciprocating test rig has been used in this study to investigate the breakaway friction coefficient of the selected materials. There are some drawbacks of using this test rig such as: mechanical removal of transfer films by front-edge surface of block specimen which might influence formation of transfer films on steel counter-surface. The reciprocating motion which does not represent the type of motion found in hydrodynamic journal and thrust bearings, might lead to surface fatigue of polymer specimens after numerous cyclic loads. However, characterizing the test materials upon breakaway friction coefficient requires a test rig which is able to provide a sufficient number of starts and stops during the test period with constant contact pressure and oil temperature, together with a contact geometry representing those found in journal and thrust bearings. These objectives were readily achieved by using the Cameron-Plint reciprocating test rig during the experiments.
3.2. Initial Test Configurations
Three contact configurations were briefly tested during initial tests to determine the most optimal configuration for further experiments. These configurations included a steel block sliding against a polymer plate, a steel cylinder sliding against a polymer plate, and a polymer block sliding against a steel plate.
3.2.1. Steel Block on Polymer Plate Configuration
Using this configuration the apparent contact pressure (P) can be calculated by the following formula:
=
(3.1)
Where F is normal load, and A is apparent contact area.
Flat-on-flat configuration is supposed to provide an even load distribution over the contact region. In practice, even load distribution is only achieved at very small loads, when there is a small elastic deformation of the polymer plate. At high loads, the steel block penetrates into the
Methodology
polymer plate due to low modulus of elasticity of concentration and higher contact pressures
Figure (3
Brief tests revealed that due to the
plates there was a significant ploughing effect edges of the steel block, led to
shows the worn surface of the polymer plate which had been run against couple of cycles.
Figure (3.2) Worn surface of polymer plate after a couple of configuration
Steel block
odulus of elasticity of the polymer. This result
higher contact pressures in the regions exposed to the edges of the steel block.
3.1) Steel block on polymer plate configuration
the large difference in elasticity of the steel block and
ploughing effect, which together with stress concentration at the bulk material separation from the polymer plate. Figure ( polymer plate which had been run against a steel block for only a
ace of polymer plate after a couple of cycles with steel block on polymer plate test Load
Motion Polymer
plate Steel block
13 results in stress exposed to the edges of the steel block.
steel block and polymer which together with stress concentration at the plate. Figure (3.2) steel block for only a
cycles with steel block on polymer plate test
14 Methodology
3.2.2. Steel Cylinder on Polymer Plate Configuration
In this configuration, the contact geometry and pressure at low loads can be approximated by the Hertzian formula for line contact, from which the width of the contact, D is:
= 2 = 8 ⁄ (3.2)
Where F is the normal load, R is the radius of cylinder, is the length of the cylinder and is the effective modulus of elasticity which can be calculated by the following formula:
= 2 1 −
+1 −
(3.3)
Where is poissons ratio and E is modulus of elasticity.
Mean pressure (Pm) and maximum contact pressure (Pmax) which is achieved in the middle of the contact width can be calculated by:
=
2 , ! = 4
(3.4)
Figure (3.3) Steel cylinder on polymer plate configuration
Preliminary tests using this configuration showed no bulk separation of polymer material.
However, there is an uneven pressure distribution in the contact region. Significant deformation
Load
Polymer plate Steel cylinder
Motion
Methodology
of the contact zone along the stroke mechanical characteristics of PTFE elastic behavior of this type of polymer Figure (3.4).
Figure (3.4) Plastic deformation of polymer plate along the stroke path using cylinder on plate test configuration
3.2.3. Polymer Block on Steel P
Using this configuration, neither
polymer blocks was observed. The polymer specimens were which limited deformation of the
account the advantages of using the latter configuration, it was decided to conduct further experiments using the polymer block
Figure (3 Polymer
contact zone along the stroke path was observed which altered the
PTFE such as compression modulus and hardness due to visco lastic behavior of this type of polymer. The worn surface of the polymer plate tested
) Plastic deformation of polymer plate along the stroke path using cylinder on plate test configuration
Steel Plate Configuration
neither macroscopic ploughing effect, nor considerabl The polymer specimens were clamped in the holder of
the polymer blocks along the reciprocating direction. Taking into account the advantages of using the latter configuration, it was decided to conduct further
polymer block against steel plate configuration.
3.5) Polymer block on steel plate configuration Load
Motion
Steel plate Polymer block
15 the physical and hardness due to visco-
tested is shown in
) Plastic deformation of polymer plate along the stroke path using cylinder on plate test configuration
considerable deformation of the holder of the test rig polymer blocks along the reciprocating direction. Taking into account the advantages of using the latter configuration, it was decided to conduct further
16 Methodology Drawings of test specimens, specimen holder and oil bath are included in Appendix A.
3.3. Frequency and Stroke Length
The average speed of the reciprocating motion (V) can be calculated by:
# = 2$, (3.5)
Where is stroke length, and $ is the reciprocating frequency.
The speed of a polymer block sliding against the steel plate was kept as low as possible to avoid hydrodynamic effects in the contact region. This was obtained through performing tests at very low frequency together with short stroke length. Therefore, experiments were carried out at reciprocating frequency of 0.1 Hz which was achieved by installing a gear box with a transmission ratio of 20:1.
The stroke length was set to 5 mm, equal to dimension of block specimen parallel to the sliding direction. This could provide a possibility for removal of wear debris from the contact region.
3.4. Breakaway Friction
Breakaway friction, which is the maximum static friction prior to relative motion, is determined from the friction data stream provided by the test rig. The value of breakaway friction is taken from the maximum value of each stroke occurring when the reciprocating test specimen goes through the stop and start at the end of each stroke. A data stream of friction coefficient of carbon filled PTFE at 4 MPa loading is shown in Figure (3.6).
Methodology 17
Figure (3.6) Data stream of friction coefficient of carbon filled PTFE at 4MPa loading for three cycles
An average of these values over the length of the test is used to determine the average breakaway friction coefficient for the materials. Variation of breakaway friction coefficient (i.e. ∆µ=µmax- µmin) during the entire test period has been analysed to characterize the stability and trends in breakaway friction coefficient for various pressures and oil temperatures. This method was used as some of the materials tested did not provide a steady state friction after running-in. Further supporting this decision is the application of this test program to hydrodynamic sliding bearings, which only experience breakaway friction at start-up. These bearings typically experience a couple of starts and stops per day over the course of their several years’ service life thus causing any variation of breakaway friction, to have significant consequences on machine operation.
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
0 10000 20000 30000 40000 50000
Data points
Coefficient of frictionn
18 Methodology
19
4
Experimental Setup
4.1. Materials
Experiments were carried out using four commercially available PTFE-based composites together with pure PTFE and Babbitt material. Details of PTFE-based composites are given in Table (4.1). Compression modulus of materials has been determined on-site. The materials have been selected due to their availability and previous study on PTFE-based composites [1].
The counter-face against which the specimens were run was, in all cases, common carbon steel polished to a surface roughness of 0.4 µm (CLA) parallel to sliding direction. The lubricant used in all experiments was SE LV synthetic ester turbine oil. Details of physical characteristics of lubricant are given in Table (4.2). Further information about lubricant characteristics is given in Appendix B.
Table (4.1) Characteristics of PTFE-based composites
Materials Compression modulus[GPa]
Specific density
Virgin PTFE 0.46 2.13
PTFE+40% bronze 0.99 3.96
PTFE+25% carbon 0.85 2.09
PTFE+25% black Glass 0.66 2.21
PTFE+20% glass fibre+5% MoS2 0.78 2.28
20 Experimental Setup
Table (4.2) Physical characteristics of SE LV synthetic ester turbine oil
Characteristics Methods Units Values
Density at 15ºC ASTM D 4052 kg/m3 940.8
Density at 40ºC Calculated kg/m3 924.55
Density at 100ºC Calculated kg/m3 855.5
Kinematic viscosity at 40 ºC ASTM D 445 m2/s 31.00 e-6 Dynamic viscosity at 40 ºC Calculated m.Pa.s 28.66 Kinematic viscosity at 100 ºC ASTM D 445 m2/s 6.17 e-6 Dynamic viscosity at 100 ºC Calculated m.Pa.s 5.47
4.2. Instrumentation
Experiments were carried out on a TE-77 Cameron-Plint reciprocating tribo-meter (Fig. 4.1) under SE LV synthetic ester lubricated condition. The reciprocating motion is provided by a moving arm coupled to an eccentric power transmission, which converts the rotational motion provided by a controlled variable speed electric motor into a linear motion with an adjustable sliding stroke. The moving arm transfers the reciprocating motion to a holder which clamps the polymer block, preventing it from tilting as the block moves back and forth against a fixed steel plate, so that a more even load distribution is achieved in the contact.
Figure (4.1) TE-77 Cameron-Plint reciprocating tribo-meter [www.ltu.se]
Experimental Setup 21 The steel plate is fixed to a stationary oil bath, which keeps the lubricant in the contact during the experiment and is supported by leaf springs with high stiffness in normal direction and appropriate flexibility in the horizontal direction. The normal force is provided by a manual loading system with a transverse bridge that is mechanically applied. An electrical force transducer in contact with the stationary steel plate measures the horizontal reaction force (friction force) that the moving specimen exerts on the steel plate [19]. A heating device with a feedback from a thermocouple mounted in the oil bath allows experiments at elevated temperatures with maximum temperature variation of ±3 ºC. The oil temperature is continuously measured during experiments.
4.3. Test Specimens
The polymer blocks are prepared in two different sizes to achieve desired contact pressures with the practical load limit of the test rig, which is in the range of 80N to 400N. The dimensions of the apparent contact area of polymer blocks for 1MPa tests are 5mm parallel and 16 mm perpendicular to the sliding direction. For the rest of experiments the dimensions of apparent contact area are 5mm parallel and 8mm perpendicular to the sliding direction. This allows for apparent contact pressures of 1MPa to 8MPa. The polymer blocks are cleaned in an ultrasonic bath using heptane and ethyl alcohol and dried in air prior to each test.
The steel plates are ground to a surface roughness of 0.4 µm (CLA) parallel to the reciprocating direction using 150 grade emery paper and later cleaned by a cloth dipped in ethyl alcohol to remove abrasive particles from frictional surfaces.
4.4. Test Conditions
The experiments include four test categories: Pressure Effect Tests (PET), Temperature Effect Tests (TET), High Pressure - High Temperature tests (HPHT), and tests with Prolonged Stop and Restart (PSR).All tests are carried out in lubricated contacts with synthetic ester turbine oil. The experimental conditions are summarized in Table (4.3).
22 Experimental Setup
Table (4.3) Experimental conditions for Pressure Effect Tests (PET), Temperature Effect Tests (TET), High Pressure-High Temperature tests (HPHT), and tests with Prolonged Stop and Restart (PSR)
Test category
Load [N]
Contact Pressure [MPa]
Oil Temperature
[ºC]
Test Duration
[hr]
Stroke Length [mm]
Reciprocating Frequency
[Hz]
PET 80-320 1,2,4,6,8 25
3 5 0.1
TET 80 2 25,45,65,85
HPHT 320* 8* 85
PSR 80 2 Room temp. 72**
*Babbitt was tested at load of 160 N representing apparent contact pressure of 4 MPa due to overloading of test rig at 320 N and 240 N normal load.
** Prolonged stop under loading after 5 minutes of running-in.
4.4.1. Pressure and Temperature Effect Tests
Tests have been carried out with reciprocating frequency of 0.1 Hz and duration of 3 hours with stroke length of 5 mm, which result in sliding distance of 10.8m and 2160 starts and stops in total. Each test has been carried out three times to ensure repeatability. Results are averaged from three test runs for each sliding condition.
4.4.2. High Pressure-High Temperature Tests
Experiments have been carried out to investigate the breakaway friction coefficient of carbon filled PTFE and bronze filled PTFE, which have exhibited better tribological performance in pressure and temperature effect tests compared to black glass- and fibreglass- filled PTFE. The results have been compared to those of pure PTFE and Babbitt.
Experimental Setup 23
4.4.3. Prolonged Stop and Restart
Lubricant is squeezed away from the contact region, when mating surfaces are left under loading without relative motion for extended period of time. The poorly lubricated contact provides higher levels of friction at the initiation of relative motion. Therefore tests with a prolonged stop have also been carried out in this study to investigate the degree of the change in breakaway friction coefficient of tested materials upon pressing a specimen against steel plate without relative motion for an extended time.
The tests are done with pure PTFE, Babbitt, carbon filled PTFE and bronze filled PTFE, which demonstrated better frictional characteristics at various pressures and temperatures compared to those of black glass- and fibreglass- filled PTFE composites.
In order to run-in the new surfaces and obtain a more stable friction coefficient, the specimens are run against steel plates at 2 MPa apparent contact pressure in lubricated contacts for five minutes. Then the specimens are left at rest under a load of 80N (2MPa). The tests restarted after 72 hours and ran for ten minutes to investigate the trend in breakaway friction coefficient after such a prolonged stop.
All tests prior to and after the idle period were carried out with synthetic oil at room temperature with reciprocating frequency of 0.1 Hz and stroke length of 5 mm. Each test was carried out two times to ensure the reproducibility of test data. The results are averaged for each material.
24 Experimental Setup
25
5
Results
5.1. Effect of Contact Pressure
Figure (5.1) shows the breakaway friction coefficient of tested materials during the test time at various pressures. Each curve is averaged from three test runs. Individual friction curves over three test runs for each material and for each tribological condition are included in Appendix C.
Figure (5.1) Breakaway friction vs. time at room temperature and various apparent contact pressures Black glass filled PTFE
0.00 0.05 0.10 0.15 0.20 0.25
0 3600 7200 10800
Time [S]
µ
1 MPa 2 MPa 4 MPa
6 MPa 8 MPa
Bronze filled PTFE
0.00 0.02 0.04 0.06 0.08 0.10
0 3600 7200 10800
Time [S]
µ
1 MPa 2 MPa 4 MPa
6 MPa 8 MPa
Carbon filled PTFE
0.00 0.02 0.04 0.06 0.08 0.10 0.12
0 3600 7200 10800
Time [S]
µ
1 MPa 2 MPa 4 MPa
6 MPa 8 MPa
Glass fiber and MoS2 filled PTFE
0.00 0.05 0.10 0.15 0.20
0 3600 7200 10800
Time [S]
µ
1 MPa 2 MPa 4 MPa
6 MPa 8 MPa
Pure PTFE
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
0 3600 7200 10800
Time [S]
µ
1 MPa 2 MPa 4 MPa
6 MPa 8 MPa
Babbitt
0.00 0.05 0.10 0.15 0.20 0.25 0.30
0 3600 7200 10800
Time [S]
µ
1 MPa 2 MPa 4 MPa
6 MPa 8 MPa
26 Results Table (5.1) shows a time average of breakaway friction coefficient of tested materials obtained at various apparent contact pressures ranging from 1MPa to 8MPa under lubricated conditions.
Table (5.1) Time average of breakaway friction coefficient of tested materials at various apparent contact pressures
Material 1Mpa 2MPa 4MPa 6MPa 8MPa
Babbitt 0.22 0.24 0.23 0.22 0.21
Glass fiber + MoS2 + PTFE 0.12 0.15 0.14 0.13 0.11 Black glass + PTFE 0.17 0.11 0.13 0.11 0.12 Bronze + PTFE 0.08 0.07 0.07 0.06 0.06 Carbon + PTFE 0.10 0.10 0.08 0.08 0.07
Pure PTFE 0.05 0.06 0.05 0.06 0.04
A graphical illustration of the data given in Table (5.1) along with standard deviation of average breakaway friction coefficients obtained from each test run is shown in Figure (5.2).
Figure (5.2) Time average of breakaway friction coefficient of tested materials at various contact pressures and room temperature
It was found that pure PTFE consistently provided the lowest friction. Bronze-filled PTFE also provided low friction. Carbon-filled PTFE exhibited higher friction compared to pure PTFE and bronze-filled PTFE, still its friction being in low levels of 0.1 or even less at various contact pressures. Although black glass- and fiber glass- filled PTFE provided highest breakaway friction among PTFE based composites, all PTFE based materials showed significantly lower friction than Babbitt. Friction levels were often less than one half of that of Babbitt.
0.00 0.05 0.10 0.15 0.20 0.25 0.30
1 2 3 4 5 6 7 8
Apparent contact pressure [MPa]
Coefficient of frictionn
Babbitt
Glass fiber+MoS2+PTFE Black glass+PTFE Bronze+PTFE Carbon+PTFE Pure PTFE
Results 27 In general, a trend of slightly decreasing breakaway friction coefficient was observed with Babbitt, MoS2 and glass fiber filled-, bronze filled- and carbon filled-PTFE. No consistent trend and no pressure dependence were noted for black glass filled and pure PTFE.
Figure (5.3) shows the difference between minimum and maximum breakaway friction coefficients (i.e. ∆µ=µmax-µmin) obtained during three hour test period. The lowest variation in breakaway friction coefficient occurs for pure PTFE followed by carbon filled PTFE. Both produced a marginal variation in breakaway friction coefficient of less than 0.03during the test period for all contact pressures.
Figure (5.3) Variation of breakaway friction coefficeint (∆µ=µmax-µmin) of tested materials during 3h test period at various contact pressures and room temperature
While a trend of low and constant variation in breakaway friction coefficient with pressure was observed with carbon-filled PTFE and pure PTFE, no such trend in breakaway friction coefficient was observed with the rest of the materials.
0.00 0.02 0.04 0.06 0.08 0.10 0.12
1 2 3 4 5 6 7 8
Apparent contact pressure [MPa]
Variation of frictionn
Babbitt
Glass fiber+MoS2+PTFE Black glass+PTFE Bronze+PTFE Carbon+PTFE Pure PTFE
28 Results
5.2. Effect of Temperature
Breakaway friction curves for tested materials during the test time at various oil temperatures are shown in Figure (5.4). Each curve is an average from three test runs. Individual friction curves over three test runs for each material and for each test condition are included in Appendix C.
Figure (5.4) Breakaway friction coefficient at 2 MPa apparent contact pressure and various oil temperatures
Table (5.2) gives a time average of breakaway friction coefficient of tested materials obtained at oil temperatures ranging from 25°C to 85°C.
Black glass filled PTFE
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
0 3600 7200 10800
Time [S]
µ
25 C 45 C
65 C 85 C
Bronze filled PTFE
0.00 0.02 0.04 0.06 0.08 0.10
0 3600 7200 10800
Time [S]
µ
25 C 45 C
65 C 85 C
Carbon filled PTFE
0.00 0.02 0.04 0.06 0.08 0.10 0.12
0 3600 7200 10800
Time [S]
µ
25 C 45 C
65 C 85 C
Glass fiber and MoS2 filled PTFE
0.00 0.05 0.10 0.15 0.20
0 3600 7200 10800
Time [S]
µ
25 C 45 C
65 C 85 C
Pure PTFE
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
0 3600 7200 10800
Time [S]
µ
25 C 45 C
65 C 85 C
Babbitt
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
0 3600 7200 10800
Time [S]
µ
25 C 45 C
65 C 85 C
Results 29
Table (5.2) Time average of breakaway friction coefficient of tested materials at different oil temperatures Material 25 ºC 45 ºC 65 ºC 85 ºC
Babbitt 0.24 0.25 0.29 0.31
Glass fiber + MoS2 + PTFE 0.15 0.11 0.11 0.17 Black glass + PTFE 0.11 0.11 0.10 0.09 Bronze + PTFE 0.07 0.06 0.04 0.04 Carbon + PTFE 0.10 0.09 0.09 0.04
Pure PTFE 0.06 0.05 0.04 0.05
Figure (5.5) shows a graphical illustration of the data given in Table (5.2) along with standard deviation of average friction coefficients obtained from each test run.
Figure (5.5) Average breakaway friction coefficients at 2 MPa apparent contact pressure and various oil temperatures
All PTFE based materials provided significantly lower breakaway friction coefficients compared to that of Babbitt.Friction levels were often less than one half of that of Babbitt.
The highest friction coefficients were obtained with Babbitt. The lowest friction coefficients were obtained with pure PTFE and bronze filled PTFE, with low and pressure independent friction levels of less than 0.08 at all oil temperatures.
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
25 45 65 85
Oil temperature [°C]
Coefficient of frictionn
Babbitt
Glass fiber+MoS2+PTFE Black glass+PTFE Bronze+PTFE Carbon+PTFE Pure PTFE
30 Results The breakaway friction of Babbitt increased with oil temperature whereas neither considerable, nor consistent trends in breakaway friction of PTFE based materials was observed with oil temperature increase.
An increasing trend in breakaway friction with temperature was observed with Babbitt. This was most likely a result of decreased lubricant viscosity that could have allowed for more severe contact with the counter-surface at higher temperatures. Although more severe contact between PTFE materials and steel counter surface is expected at lower oil viscosities, low shear strength at friction surface of PTFE based materials may explain the difference between frictional behaviour of PTFE based composites and that of Babbitt with temperature increase.
Figure (5.6) shows the difference between maximum and minimum breakaway friction coefficients (i.e. ∆µ=µmax-µmin) during a three hour test period for temperature effect tests.
Figure (5.6) Variation in breakaway friction coefficeint (∆µ=µmax-µmin) during a 3h test period at 2 MPa apparent contact pressure and various oil temperatures.
Babbitt along with glass fiber and MoS2 filled PTFE were observed to yield the highest variation in breakaway friction during the test period reaching the maximum levels of variation at 85 ºC.
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
25 45 65 85
Oil temperature [°C]
Variation of frictionn
Babbitt
Glass fiber+M oS2+PTFE Black glass+PTFE Bronze+PTFE Carbon+PTFE Pure PTFE
Results 31 Pure PTFE and carbon filled PTFE demonstrated very low variation in breakaway friction.
Bronze filled PTFE provided the same level of variation as pure PTFE and carbon filled PTFE at higher temperatures.
5.3. High Pressure – High Temperature Tests
Breakaway friction from high pressure-high temperature tests is displayed in Figure (5.7). Tests were conducted with Babbitt, carbon-filled, bronze-filled, and pure PTFE. The poor performance of black glass- and fibreglass-filled PTFE in earlier experiments disqualified these materials from further investigation. Oil temperature and contact pressure were set to 85°C and 8MPa respectively to investigate their combined effect on friction.
Figure (5.7) Breakaway friction vs. time at 85 °C oil temperature and 8 MPa* apparent contact pressure
*The friction curve demonstrated for Babbitt is obtained at 4MPa due to overloading of test rig at 8 and 6 MPa apparent contact pressures
It was found that Babbitt produced the highest breakaway friction coefficient among the tested materials. It was about 4-6 times higher than those of bronze-filled, carbon-filled and pure PTFE.
While an increasing trend in breakaway friction coefficient was observed for Babbitt during the test time, much lower and very stable breakaway friction coefficients were obtained with PTFE- based materials. The friction levels for PTFE based materials were lower than those obtained at high pressure or high temperature alone.
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
0 3600 7200 10800
Time [S]
µ
Babbitt Bronze+PTFE Carbon+PTFE Pure PTFE
32 Results
5.4. Prolonged Stop and Restart
Testing of the effects of mating materials left under loading for an extended period of time, such as in the case of a machine stop, were conducted with Babbitt, bronze filled-,carbon filled-, and pure- PTFE. The poor performance of black glass- and fibreglass-filled PTFE in the earlier testing disqualified these types of fillers from further testing and so they were not included in the time consuming extended stop tests. Results showing the average breakaway friction coefficient before and the breakaway friction coefficient for each cycle after the stop are displayed in Figure (5.8). The friction values at ‘0’ cycle, are the average from the one minute prior to the stop, while
‘1’ and ‘2’ are the first and second cycles after the stop and so forth. The average friction values at ‘0’ cycle are in agreement with the friction coefficients measured for these materials at 2 MPa contact pressure and t=5 minutes (Figure 5.1).
Figure (5.8) Breakaway friction before and after 72 hour loading without relative motion at room temperature and 2 MPa apparent contact pressure
One of the most notable features of these results is that the maximum friction returns to values measured prior to the breakaway after just one stroke, regardless of the material. However, in regards to the materials, it was observed that Babbitt exhibited an increment of 0.32 in breakaway friction compared to the steady state value measured before the prolonged stop, providing a breakaway friction coefficient of 0.52, while PTFE-based composites exhibited
0.00 0.10 0.20 0.30 0.40 0.50 0.60
0 1 2 3 4 5 6
Number of cycle after 72h idle time
Coefficient of frictionn
Babbitt Bronze +PTFE Carbon +PTFE Pure PTFE
Results 33 increase of maximum 0.07 in breakaway friction, reaching less than one third of friction levels obtained with Babbitt after the prolonged stop. It is felt that the high increase ratio of Babbitt was caused by the large difference in breakaway friction coefficient of Babbitt in dry and lubricated contacts. In the case of PTFE materials, reduction of amount of lubricant in the contact is believed to allow PTFE to operate in the ranges of low friction that it is known for PTFE in dry contacts.
5.5. Wear of the Steel Counter-surface
The effects the materials had on the steel counter-surface were analysed using an optical profilometer. Roughness average (Ra) was measured at the same points in the center of the wear track both before and after testing to determine the degree of polishing or roughening experienced by the surface through the test. Figure (5.9) and (5.10) show the steel counter-face before and after testing respectively with glass fibre filled PTFE at 6 MPa apparent contact pressure.
Figure (5.9) Steel counter-face before test with glass fibre filled PTFE at 6 MPa apparent contact pressure and room temperature
34 Results
Figure (5.10) Steel counter-face after test with glass fibre filled PTFE at 6 MPa apparent contact pressure and room temperature
Figure (5.11) Change in Ra surface roughness of steel counter-surface in tests at 1 and 6 MPa contact pressures
These results are displayed in Figure (5.11) showing that only the glass-filled PTFE based materials (black glass- and fibre glass- filled PTFE) had a significant effect on the counter- surface at high load. Given the significant surface polishing provided by these materials at 6 MPa load level it is felt that the trend observed at 1 MPa loading also represents polishing by the glass-based materials even though the variation is very near to the uncertainty of the measurement method. Although the change in surface roughness of steel when tested against bronze filled PTFE was marginal, a transfer film of bronze could be detected on the steel
1 MPa 6 MPa -350
-300 -250 -200 -150 -100 -50 0 50 100
Roughness change (nm)
Babbitt
Glass fiber+M oS2+PTFE Black glass+PTFE Bronze+PTFE Carbon+PTFE Pure PTFE
Results 35 counter-faces by the naked eye; however such transfer film could not be observed in cases of carbon filled PTFE and pure-PTFE.
5.6. SEM investigation of block specimens
Investigations of block specimens were carried out utilizing a scanning electron microscope (SEM). The sliding direction in all SEM images is approximately in the vertical direction.
5.6.1 Babbitt
SEM evaluation of the worn Babbitt surface revealed a significant wear scar with grooves cut into the material by the steel counter-surface. A representative image of the worn Babbitt surface is displayed in Fig. 5.12. Counter-surface material was not detected on the worn surface signifying that wearing occurred on the Babbitt material only. Similar wear scars were seen on the Babbitt specimens for both high and low loads regardless of temperature.
Figure (5.12) SEM image of worn Babbitt
5.6.2 Black-glass filled PTFE
Investigation of the black-glass filled PTFE revealed the reason for the change in surface roughness observed on the counter-surfaces. The fibers of the black glass abraded the steel surface, and collections of iron particles were piled up on the friction surface of the PTFE. This effect is clearly demonstrated in Fig 5.13.bin which the darkest areas are PTFE, the gray areas
36 Results are fibers and the white areas were found to be rich in iron. Given that the PTFE material did not initially contain iron, it can only be assumed that the iron presence was the result of wearing on the counter surface. In this case, it appears that the PTFE did very little to reduce the friction in the contact as large portions of the load appear to have ridden on the harder glass fibers.
a) b)
Figure (5.13) SEM images of black glass filled PTFE a) unworn b) tested at 8MPa apparent contact pressure
5.6.3 Fiberglass and MoS
2filled PTFE
Analysis of the worn fiberglass and MoS2 filled PTFE revealed large areas of blended MoS2 and iron as well as sharp fibers. As in the case of black glass, these sharp fibers tended to collect the wear debris in the direction of sliding and produced significant polishing on the counter-surface.
This effect is shown in Fig. 5.14. While the MoS2 appeared to have smoothed out the PTFE surface, it does not appear to have reduced friction in lubricated conditions as the friction levels were greater for MoS2and fiberglass filled PTFE than black glass filled PTFE.
a) b)
Figure (5.14) SEM images of fiberglass and MoS2 filled PTFE a) unworn b) tested at 8MPa contact pressure
Results 37
5.6.4 Bronze filled PTFE
The wear characteristics of bronze filled PTFE were very consistent between high and low pressures and high and low temperatures. In all cases, the bronze particles had scratches parallel with the sliding motion. However, depending on the load, the bronze particles and the PTFE surrounding them appeared to have worn at different rates. This difference is highlighted in Fig.
5.15 showing bronze filled material tested at 2 and 8 MPa. At 2 MPa, it appears that the surfaces of the worn bronze particles and the bulk material are approximately level with smooth transitions between filler and bulk material. This is not the case at 8 MPa where gaps can be distinguished around the edges of the filler particles.
Additionally, the bronze particles occupy a much larger area of the contact region at higher pressure than at lower pressure loading. It is thought that the bronze particles are more wear resistant than the PTFE material and so they collect at the surface and are pushed into the bulk material as it is steadily worn away.
a) b) c)
Figure (5.15) SEM images of bronze filled PTFE, a) unworn b) tested at 2 MPa contact pressure c) tested at 8MPa contact pressure
5.6.5 Carbon filled PTFE
Carbon filled PTFE had consistent performance at low and high pressures and low and high temperatures. In all cases, the carbon particles wore evenly with the PTFE bulk material as shown in Fig. 15.16 in which the darker patches are carbon and the lighter gray areas are PTFE.
The primary difference between heavier and lighter loads was observed in the smoothness of the worn surface with higher loads yielding a markedly smoother surface than lighter loads.
38 Results
a) b) c)
Figure (5.16) SEM images of carbon filled PTFE a) unworn b) tested at 2 MPa contact pressure c) tested at 8MPa contact pressure
5.6.6 Pure PTFE
Pure PTFE showed very similar wear characteristics to carbon filled PTFE with an even smoothing of the surface in general without the complex characteristics seen for the composite materials. As could be expected, counter-surface particles were not observed as in the case of the glass filled materials.
a) b) c)
Figure (5.17) SEM images of pure PTFE a) unworn b) tested at 2 MPa contact pressure c) tested at 8MPa contact pressure
39
6
Discussion
Discussion
In tests with variation in pressure, an initial increase in breakaway friction followed by a decrease and increase was observed with both glass-filled PTFE composites at higher contact pressures (6 and 8 MPa) forming an “N” shape friction curve as shown in Figure(5.1).
This is thought to be due to the combined effect of wear of PTFE matrix, polishing of steel counter-face, and embedment of steel wear debris on polymer surface at different stages of the test period.
As wear of the PTFE matrix progresses, more fillers are uncovered and engaged with the steel counter surface resulting in increased friction coefficient of the mating materials at the initial stages of relative motion. Subsequent reduction in breakaway friction coefficient is thought to be due to decreased surface roughness of the steel counter-face as the hard fillers tend to wear off the asperities of the steel counter-face resulting in less ploughing and reduced breakaway friction.
It should be noted that although polishing of the steel surface by hard fillers reduce the friction;
generation and embedment of steel wear debris on the contact surface of the PTFE composites, as shown in SEM images of worn glass filled PTFE, lead to increased adhesion and abrasion of the mating materials. The latter may explain the trend of increased friction coefficient of the mating materials at the final stage of the test period.
