<%BANNER%>

Spectrum loading and multidimensional sliding of PTFE with a pin-on-flat tribometer

University of Florida Institutional Repository

PAGE 1

SPECTRUM LOADING AND MULTIDIRECTIONAL SLIDING OF PTFE WITH A PIN-ON-FLAT TRIBOMETER By DARREN MCGUIRE A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2003

PAGE 2

ACKNOWLEDGMENTS I would like to acknowledge all those that helped make the culmination of my academic career a reality. I cannot express the debt of gratitude I feel nor stress the importance of every persons contribution to my thesis project. However, I would at least like to point out a few that made it possible. I thank Howard Purdy for his machining efforts, Gregory Sawyer for his guidance, knowledge, and friendship, Dr. Ziegert Dr. Arakere, and Dr. Schueller for their review of this thesis and assistance in the classroom and lab throughout the course of my entire education, and members of the tribology lab for their suggestions, perspectives, and support. Most of all I would like to thank my parents and family. Emotions can leave words far in the distance and I find none that can explain what they mean to me. I love them more than life itself. ii

PAGE 3

TABLE OF CONTENTS Page ACKNOWLEDGMENTS..................................................................................................ii LIST OF TABLES...............................................................................................................v LIST OF FIGURES...........................................................................................................vi ABSTRACT.....................................................................................................................viii CHAPTER 1 INTRODUCTION........................................................................................................1 2 REVIEW OF LITERATURE.......................................................................................4 2.1 Proposed Mechanisms of Wear for PTFE.........................................................4 2.2 Attempted Improvements in Wear Characteristics of PTFE..........................13 2.3 Wear of UHMWPE.........................................................................................18 3 ENGINEERING APPROACH...................................................................................21 3.1 Six-Station Pin-on-Flat Tribometer.................................................................21 3.1.1 Table and Drive System..........................................................................21 3.1.2 Pneumatic Control..................................................................................22 3.1.3 Sample Holders.......................................................................................24 3.2 Motion Paths and Loading Patterns.................................................................27 3.3 Counterface Preparations and Handling..........................................................30 3.4 Polymer Sample Preparation............................................................................31 4 EXPERIMENTAL RESULTS...................................................................................34 4.1 Electro-Pneumatic Performance Data..............................................................34 4.2 Variations in Wear Rate and Sliding Conditions.............................................35 4.2.1 Transfer Film Formation.........................................................................35 4.2.2 Wear Rate Comparisons.........................................................................40 4.3 Cycle dependence on Wear..............................................................................46 4.4 Images of Wear................................................................................................47 iii

PAGE 4

5 SURFACE CHARACTERIZATION AND SUBSURFACE STRESS MODELING49 6 DISCUSSION.............................................................................................................58 6.1 Delamination....................................................................................................58 6.2 Reversal Zones.................................................................................................62 6.3 Cycle Dependence in Wear Rate.....................................................................63 7 CONCLUSIONS........................................................................................................65 APPENDIX A MOTION PATH PROGRAMS..................................................................................67 B RAW DATA...............................................................................................................71 C SHOP DRAWINGS....................................................................................................74 D SURFACE METROLOGY........................................................................................78 LIST OF REFERENCES...................................................................................................79 BIOGRAPHICAL SKETCH.............................................................................................81 iv

PAGE 5

LIST OF TABLES Table page 2-1. Collection of PTFE wear rate data from previous authors........................................17 3-1. Equipment register for pin-on-flat tribometer...........................................................26 3-2. Polishing steps for raw counterface samples..............................................................31 3-3. Polishing steps for used counterface samples.............................................................31 4-1 Raw data with calculated wear rates............................................................................36 v

PAGE 6

LIST OF FIGURES Figure page 2.1. Vertical pin-on-disk configuration.............................................................................12 3.1. Photograph of pressure gauges (top) backpressure to cylinders (bottom) pressure to electro pneumatics....................................................................................................22 3.2. Arrangement of electro pneumatic gauges................................................................23 3.3. Photograph of pneumatic cylinder arrangement........................................................24 3.4. Assembly drawing of entire pin-on-flat tribometer...................................................25 3.5. Schematic of polymer sample holder.........................................................................27 3.6. Loading spectrum applied to samples for wear testing..............................................29 3.7. Motion paths used for wear testing.............................................................................30 3.8. Schematic of polymer sample....................................................................................32 4.1. Electro-pneumatic performance data output from load cells.....................................34 4.2. Optical micrograph of PTFE transfer film characteristic of high wear rates (a) wear debris (b) end of wear path (c) top middle (d) bottom middle.................................38 4.3. Optical micrograph of PTFE transfer film characteristic of low wear rates (a) end of wear path (b) top middle (c) bottom middle (d) middle...........................................38 4.4. Optical micrograph of PTFE transfer film deposited by 14.9 mm diameter circular wear path (a) top (b) bottom (c) left (d) right...........................................................39 4.5. Optical micrograph of PTFE transfer film deposited by diamond pattern (a) soft corner (b) sharp corner.............................................................................................40 4.6. Wear rate as a function of load for 670 meters of linear reciprocating sliding.........41 4.7. Effects of varying load on wear rate compared with effect of loading spectrum on wear rate...................................................................................................................42 4.8. Effects of load and diameter on wear rate for circular motion..................................43 vi

PAGE 7

4.9. Ratio of circular motion wear rates over linear reciprocating motion wear rates......44 4.10. Wear rates for diamond pattern sliding motion as a function of inclusion..............46 4.11. Wear rate as a function of number of cycles............................................................47 4.12. Photograph of polymer samples before 4 hour linear reciprocating wear test at 176 N (left), and after (right)...........................................................................................47 4.13. PTFE wear post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235 N, right 176 N...............................................................................................................48 4.14. Wear paths post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235 N, right 176 N...............................................................................................................48 5.1. Asperity peak configuration at the surface of the steel counterface..........................50 5.2. Radius of asperity peak at surface of steel counterface.............................................51 5.3. Pressure profile applied to PTFE surface when in contact with counterface.............52 5.4. Sigma X compressive stress in subsurface of PTFE..................................................54 5.5. Sigma Z compressive stress in subsurface of PTFE..................................................55 5.6. Tau XZ shear stress in subsurface of PTFE...............................................................56 5.7. Plot of subsurface shear stress along x = 0 indicating shear max..............................57 6.1. Presents of subsurface cracks within polymer pin under stress.................................59 6.2. Subsurface cracks begin to propagate and link up.....................................................59 6.3. Ejections of polymer wear debris resulting from large subsurface crack..................60 6.4 Modes of crack propagation a) mode I b) mode II.....................................................61 6.5. Model showing wear rate transition at some critical number of cycles n c ................63 vii

PAGE 8

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science SPECTRUM LOADING AND MULTIDIRECTIONAL SLIDING OF PTFE WITH A PIN-ON-FLAT TRIBOMETER By DARREN MCGUIRE August 2003 Chair: W. Gregory Sawyer Cochair: John Ziegert Major Department: Mechanical and Aerospace Engineering Wear is a process of gradual breakdown or removal of material by relative motion between two contacting surfaces. In this report, the wear of polytetrafluoroethylene (PTFE) under spectrum loading and multidirectional sliding was investigated. Wear tests revealed a strong proportional dependence on load but showed no dependence on sinusoidal fluctuations in load. Abrupt changes in the direction of motion, known as reversal zones, correlated with increases in wear, meaning the more reversal zones present on a motion path, the more severe the wear was compared to a motion path with no reversal zones (circular). Multidirectional sliding appeared to have no influence on overall wear. This differs greatly with what is known about the wear behavior of ultra high molecular weight polyethylene (UHMWPE). The difference may stem from a delamination process that takes place during PTFE sliding. Using data collected from a viii

PAGE 9

white light interferometer, a model of the counterface surface was constructed, and when applied to the polymer pin using Hertzian contact assumptions induced a pressure profile that yields subsurface shear stresses. The magnitude of the subsurface shear stresses shows no dependence on direction of sliding. This is offered as an explanation for the lack of increased wear during multidirectional sliding over unidirectional sliding. Finally a correlation between number of cycles incurred over a wear path and wear rate is observed. A simple model using rules of linear mixing was constructed for wear predictions based on single point mass loss measurements. ix

PAGE 10

CHAPTER 1 INTRODUCTION Wear is defined by the dictionary as to impair, deteriorate, or consume gradually by use or any continued process. From an engineering standpoint, wear maybe be described as the gradual breakdown or removal of material stemming from relative motion between two contacting surfaces. Wear of special polymers such as polytetrafluoroethylene (PTFE) and ultra high molecular weight polyethylene (UHMWPE) has become a major concern due to their effectiveness as biocompatible materials. Despite the success of polymers as bearing surface for total knee or hip replacements in treatment of end stage arthritis, the gradual breakdown of these surfaces has led to component loosening and an eventual need for revision surgery. The mechanisms that cause removal of material can be resolved into abrasive, adhesive, and fatigue wear. Abrasive wear occurs when asperity peaks on the harder surface scratch away material on the softer surface. Adhesive wear occurs when bonding forces between the two surfaces are strong enough to pull material away from one or both surfaces as relative motion occurs between them. Fatigue wear occurs when material is removed as a result of cyclic stresses that exceed the fatigue strength of the material [50]. Each mechanisms contribution to wear is difficult to observe when all three factors are acting simultaneously. Minimizing the wear mechanisms that are not of interest allows the remaining mechanisms to dominate the wear process. Therefore, an observation of a 1

PAGE 11

2 single desired mechanism or a combination of desired mechanisms affects could be made. Throughout this paper adhesive wear will always be a central focus. Several wear tests will be run with the interest of investigating behavior of adhesive wear as well as behavior of adhesive and fatigue wear in combination. The tests presented in this paper all involve PTFE in contact with stainless steel counterfaces. For the case in which one of the two surfaces in contact is significantly harder than the other surface, minimizing abrasive wear requires that the harder surface be highly polished, thereby limiting the material scratched away from the softer surface to an insignificant amount. Under dry testing, wear rates for PTFE and UHMWPE are known to increase dramatically when in contact with counter faces displaying and average roughness (Ra) greater than 0.1m. Fatigue wear can be virtually eliminated by maintaining the load holding the two surfaces in contact constant. However, some cyclical effects will be present as the frictional force vector changes direction along with changes in the sliding path direction. Wear rate is a key factor in characterizing a materials performance in wear testing. If wear rate is known prior to testing, predictions about how much material will be lost during the course of a test can be made. However, wear rate does not always remain constant and may change depending on the testing parameters implemented. Although the molecular structures for both PTFE and UHMWPE are similar, their wear behavior differs drastically. UHMWPE wear is dominated by a surface phenomenon that is highly dependent on changes in the direction of sliding. However, PTFE shows no directional dependence and appears to wear from a subsurface phenomenon known as delamination. A process where subsurface imperfections or cracks are forced to propagate due to

PAGE 12

3 subsurface stresses incurred during sliding. The cracks propagate and link up with one another forming larger cracks. These cracks continue to propagate and eventually turn towards the surface breaking off a flake of wear debris. This process repeats through out the duration of sliding. However, the wear process has been shown to depend on the development of a transfer film. This indicates that the wear process is not completely driven by the bulk material, but depends on the surface interactions between bulk and counterface.

PAGE 13

CHAPTER 2 REVIEW OF LITERATURE 2.1 Proposed Mechanisms of Wear for PTFE With all the recent interest surrounding UHMWPE and its use in human joint replacements many experiments have been compiled in hopes of developing a predictive wear model. Sliding wear of UHMWPE is described as a surface wear phenomenon where the molecular chains align themselves with the direction of sliding. Once the molecular chains align themselves with the direction of sliding, wear rate drops significantly. If the direction of sliding changes, wear rate increases until the chains can reorient themselves and wear rate drops once again. If the direction of sliding is always changing there is no preferred molecular orientation and wear rate remains significantly higher than under unidirectional sliding. Although PTFE and UHMWPE have similar molecular composition their wear behaviors differ drastically in that PTFE is considered to have poor wear resistance. Tanaka, Uchiyama, and Toyooka [1] performed early investigations into the wear of PTFE. They observed that when slid against a glass plate, PTFE deposited a fibrous thin film with long bands and striations perpendicular to the length of the bands. At a sliding speed of 20 cm/s, the PTFE wear was shown to increase linearly with increasing sliding distance while the friction measured during sliding initially decreased before settling to a constant value. The initial drop off in friction is attributed to formation of the PTFE transfer film. Tanaka later demonstrated that varying the sliding rate produced 4

PAGE 14

5 a change in the PTFE wear rate. Sliding a 5 cm/s resulted in an initial high wear rate that quickly transitioned to a slower wear rate. He also showed that at temperatures below 100C the wear rates would peak for a given sliding speed, and that these peaks shifted to higher speeds as the temperature was increased. Whereas for temperatures above 100C the peaks tended to shift towards lower speeds as temperature was increased. Friction decreased with increasing temperature and increased with increasing sliding speed. Tanaka also used a thermocouple to measure the temperature rise at the sliding surface, and concluded that the temperature rise due to frictional heating was insignificant. This was attributed to the rapid removal of PTFE during sliding and the transfer films aid in heat dissipation. Inspection of the worn surfaces revealed that the same wear mechanism of PTFE acted under all circumstances. The molecular orientation and the uniform separations of striations in the transfer films indicate that the PTFE fibers are held together by lateral connections. Tanaka states that the explanation for PTFEs high wear rate stems from the lack of melting at the PTFE/Substrate interface and the easy removal of film from the substrate. Inspection of the transfer film reveals that a thin layer of amorphous PTFE is likely removed from the bulk during sliding. The shearing that occurs during sliding causes slippage in the amorphous region and the PTFE is then deposited in slices. Given this, Tanaka attributes the effect of sliding speed on friction and wear to the likelihood of viscoelastic behavior between these slices. When abrasive wear is present the slippage between slices does not occur and no film is observed. This is due to the severe damage caused by abrasive wear. In 1977 Tanaka and Miyata [2] published a study of PTFE cylinders slid on glass plates. The study into the friction and transfer of semi-crystalline polymers reveled that

PAGE 15

6 adhesion of the molecular chains in PTFE is stronger when the direction of sliding is perpendicular to the orientation of the molecular chains. This was contradictory to previously understood results. Tanaka and Miyata examined the transfer films after several increments in traverse. They observed that the transfer film thickness increased with increasing traverses. However, they limited the number of traverses to twenty. They were also able to observe that the kinetic friction value remained constant for all traverses. This suggests that shearing occurs within the PTFE transfer film. Tanaka and Miyata also suggest that the banded structure of the PTFE film as well as adsorption of water molecules on the PTFE molecular chains reduces shearing strength on the PTFE surface. However, they go on to suggest that the friction behavior and transfer of PTFE during sliding is mainly influenced by the molecular profile and not by the banded structure. Tanaka previously found that film thickness formation of about 300 angstroms is based on mutual slippage of the crystalline slices in the banded structure, but was unable to apply these findings to the current experiments. The prevailing model for PTFE transfer from bulk to substrate under reciprocating conditions is that of very long, straight, and crystalline ribbons. Using a oxide-covered Si wafer, M. Schott [3] observed that these ribbons cover nearly the entire substrate and stated that the morphology of the films depends on the temperature at which sliding takes place, the speed at which the PTFE bulk slides relative to the substrate, and the normal load applied to the PTFE bulk while sliding. Schott chose to indicate the loading conditions as weights in order to deduce the nominal pressures. Both the PTFE and substrate were kept at the same temperature and sliding speed was kept low and smooth (0.4 to 2 mm/s). Schotts experiments were conditioned by allowing the starting PTFE

PAGE 16

7 material to be removed before the transfer film was formed. Using IR transmission data, nuclear reaction analysis, and atomic force microscopy Schott was able to confirm the presence of a highly oriented PTFE transfer film. Bodo and Schott [4] did a second investigation into the formation of highly oriented PTFE films in which a PTFE rod was slid against an oxide-covered silicon wafer under controlled load and temperature. The PTFE samples were conditioned by prior sliding to eliminate the initial rapid wear phase of PTFE. Schott again saw the classical characteristics of a PTFE transfer film. However, he was able to generalize the films into three different categories. The first group consisted of irregular ribbons and low coverage (20% of wear path). Both the second and third groups showed long, straight, and parallel ribbons, but the second group had incomplete coverage while the third had complete coverage. Schott reported that the first group appeared during tests run at temperatures under 150C. The ribbons showed kinks and branching. He also states that although the ribbons consisted of bundles of polymer chains running parallel to the ribbons, the ribbons appeared considerably longer than the macromolecule in the PTFE. Schott goes on to state that the ribbons are formed during or after film deposition. The second type of film occurred during tests run above 150C and the third occurred above 220C and 600g loads. However, the coverage decreased to 76% when the load was increased to 2425g. The film cross-sectional thickness increased only slightly with increasing temperature. Schott acknowledges that film variations may result from variations in sample-substrate contact. He suggests that at high temperatures and loads, the variations diminish due to higher plastic deformation. Schott also reports that the film morphology is independent of substrate material and is a property of PTFE only.

PAGE 17

8 Using the concept of dimensional analysis, Kar and Bahadur developed an equation describing the wear of polymers in 1974. The Kar and Bahadur equation was a function of pressure P, speed v, time T, modulus of elasticity E, surface energy thermal conductivity K and specific heat C p The dimensional analysis yielded an equation for volume loss represented by: zpyxayyzyxKCETkP33 V (1) where x, y, and z are exponents determined experimentally. In 1995 Viswanath and Bellow [5] extended the Kar and Bahadur equation to include counterface surface roughness. Viswanath and Bellow again used dimensional analysis to derive a dimensionally homogeneous equation that included the specified variables. Viswanath and Bellow included five dimensionless groups comprised of wear volume V, surface energy modulus of elasticity E, specific heat C p thermal conductivity K, contact force W, counterface roughness sliding speed v, and time T. The groups signify dependence on interface contact and deformation (VE 3 / 3 ) normal load and strength characteristics (WE/ 2 ), speed and temperature (TEC p /K), counterface roughness (E/), and thermal contributions (C p /vK) yielding the dimensionless function: 0,,,,233EKTECCvKWEVEpp (2) Viswanath and Bellow then applied a non-linear relationship to map their experimental results. qrpsqpsrpsrqpKCETkW233 V (3)

PAGE 18

9 where p, q, r, and s are exponents determined experimentally. They found, for a pin-on-disk configuration with steel counterfaces, Equation (3) allows a predictive model for volume losses based on operating and material factors know to affect wear. In 1990, Yang et al. [6] investigated the effect of temperature on the formation of transfer films. Yang and Hirvonen used a pin-on-plate configuration to produce sliding of a PTFE specimen on a stainless steel counterface. They used -ray yield curves to verify the amount of PTFE transferred during sliding. They were able to observe that the deposition rate was higher during the first traverse of unidirectional sliding, and that prolonged sliding could be modeling using a linear equation: 10 nxx (4) where x is the amount of PTFE transferred, x 0 is the amount transferred after the first traverse, v is the steady-state transfer rate, and n is the number of traverses. They also observed that transfer rate depended on temperature. Their normalized data indicated that thickness increased less rapidly with increasing temperature as the number of traverses was increased. They model this using the following equation: TgnTxnTxrt,, (5) where x(T,n) and x(T rt ,n) are the thicknesses of the transferred PTFE after n traverses at room temperature and temperature T, and g(T) is a temperature enhancement factor. Despite variations in transfer film appearance at certain temperatures and qualitative disagreement with transfer and PTFE hardness, Yang and Hirvonen concluded that transfer of PTFE during the first traverse increases non-linearly by approximately an order of magnitude when increasing temperature from room to 200C. However, the

PAGE 19

10 steady-state transfer rate at room temperature and 100C are virtually the same while transfer at 200C is increased significantly. Using a pin-on-disk configuration, Blanchet and Peng [7] published an investigation into irradiated PTFE wear on stainless steel. Irradiating PTFE causes primary and secondary radicals to form by severing the C-C and C-F bonds. The radicals then recombine to form a cross-linked network that has been shown to offer better wear resistance than the original linear PTFE molecular structure. Blanchet and Peng showed that the wear rate of PTFE decreased as irradiation dose increased from 0 to 30 Mrad then increased somewhat as the irradiation dose increased from 30 to 100 Mrad. They also observed that the hardness of PTFE increased from R h = 43 to 53 over the same 0 to 30 Mrad dose. Friction was also found to increase as the irradiation dose increased from 0 to 5 Mrad. However, friction was then observed to decrease as to dose increased to 20 Mrad then settled to a constant value that was comparable to the coefficient of friction of unirradiated PTFE. In addition to these findings, they also observed a change in the wear debris morphology. Unirradiated PTFE is known to produce large plate like wear debris where as the irradiated PTFE yielded a very fine debris. Blanchet and Peng concluded that the difference in properties between unirradiated and irradiated PTFE was due to the branched and cross-linked network preventing molecular orientation. Molecular orientations of PTFE chains result in low shear strength and the subsequent formation of transfer films and wear debris. Briscoe proposed that polymer wear falls into two categories known as cohesive wear and interfacial wear [8]. Briscoe describes cohesive wear as wear involving dissipation of frictional work such as abrasion and fatigue. The level of damage due to

PAGE 20

11 sliding is prescribed by surface asperities and surface traction largely governed by the cohesive strength of the polymer. Briscoe describes interfacial wear as wear involving dissipation of frictional work by transfer and is more damaging than cohesive. He also points out that the current research indicated a correlation between wear rate and the reciprocal of the work to rupture (1/ y y ) for a given polymer. Briscoe goes on to state that polymers considered ductile, such as PTFE, are far less affected by rough surface anomalies than glassy polymers such as PMMA, and that PTFE can accommodate more strain before rupture. He gives an equation for ductile abrasion based on a theory of asperity resistance: tan2 (6) HKWztan (7) Where is the coefficient of friction, is the slope of a conical asperity, W is normal load, H is hardness, and K is probability constant. In regard to PTFE, Briscoe also points out that molecular orientation affects transfer wear, and that the transfer is largely unchanged polymer. Most polymers form a thick lumpy film during transfer. However, PTFE transfer is inconsistent forming either a relatively thin film or a patchy uneven film once steady sliding has occurred. There also appears to be an upper limit for sliding velocity at a given temperature for which transfer can occur. Briscoe also states cracking and delamination may be responsible for the transfer film often appearing rumpled and having patches of plate-like polymer removed from the film. He also states that this debris acts like a lubricant in the case of PTFE. To conclude, Briscoe points out the need

PAGE 21

12 for continued research but suggests that wear rate is a function of the material and a dominant wear mechanism. Briscoe [9] also investigated the effects of combined rotating and linear sliding motions on the wear and transfer of PTFE. Briscoe contends that the increasing wear of PTFE with increasing number of traverses is dependent on kinematics of relative motion between the sample and the substrate surface. Wear is dependent on sliding distance, and normally increases with increasing distance. For the vertical pin-on-disc machine shown in Figure 2.1 wear rate actually decreases with increasing number of traverses. Figure 2.1. Vertical pin-on-disk configuration. PTFE wear shows a behavioral change depending on the angular velocity of the pin. Effective linear sliding is achieved when the pin is rotated with the same angular velocity of the disc but in the opposite direction. Wear remains a function of sliding distance, but decreases as the pins angular velocity is varied to simulate pure linear sliding to rotational sliding with a radius equal to the radius of the PTFE pin. Briscoe showed that the combination of spins inputs energy into the system that results in wear rate fluctuations. A relative rotation between the pin and disk inhibits molecular orientation

PAGE 22

13 and results in increased wear and friction. Briscoe also states that motion aligned with polymer fibers leads to fibers being pulled away from the bulk. Wear and friction of PTFE are known to undergo a transition at approximately 0.1 km. Vijayan and Biswas [10] observed a slight expansion in the unit cell up to this distance of sliding and unit cell shrinkage with increased sliding distance. They also observed a steady increasing in wear rate and friction until a steady-state wear and friction were reached around 3 km of sliding. They were also able to confirm that no change in the atomic structure of the PTFE surface took place over this distance. Although Vijayan and Biswas contend there is a strong correlation between wear rate and unit cell volume, they were unable to correlate the change in crystallite sizes to wear rate. Brainard and Buckley [11] discovered that PTFE would transfer to metals under static conditions if a normal mass of 100 g were applied (pin dimensions not given). Due to the increased area of contact under load, it is not clear whether or not the transfer is load dependent. They also were able to conclude that PTFE transfer is possible with or without an oxide layer covering the metal substrate. Therefore, the chemical behavior of the substrate is not significant. The adhesion force was non-existent during the first two minutes of loading. Therefore, Brainard believes that adhesion is strongly influenced by creep. 2.2 Attempted Improvements in Wear Characteristics of PTFE Cheng et al. [12] improved the wear properties of PTFE by preparing a composite with 60% wt. lead and 5% wt. glass fiber. Xue and Xie recorded friction coefficients and mass loss of PTFE composite samples sliding on steel using a pin-on-disc configuration with an oil lubricant. They concluded that PTFE containing modified glass fiber demonstrated better wear resistance and possessed a lower friction coefficient than virgin

PAGE 23

14 PTFE. They credited the lead and glass fillers with decreasing wear by controlling the size and shape of the wear debris. The glass fiber also provides load support at the sample/substrate interface that decreases wear of the PTFE composite. There have been several attempts to optimize the surface features of sliding counterfaces. Optimizing these features would achieve the lowest possible polymer wear and friction. Wieleba [13] studied the frictional and wear relationship between PTFE composites and the surface texture of steel counterfaces. While maintaining sliding velocity, pressure, and sliding distance constant throughout all tests, Wieleba carried out his experiments under dry sliding conditions. Using regression functions, Wieleba showed that wear rate is more dependent on surface features than friction. The regression functions predict that friction is most greatly influenced by the mean spacing between asperities. The larger the spacing between peaks the higher the friction value became. Wieleba suggests that large spacing makes formation of a PTFE transfer film more difficult. He showed friction was similarly influence by average asperity slope. The smaller the average slope the larger area of contact, consequently, adhesion forces friction to increase. Wear rate was most strongly influenced by asperity peak heights and the peaks average radius of curvature. The higher the asperity peaks as well as the sharper the peaks the more severe the wear. This is explained by the increase in mechanical interactions between the counterface and the PTFE under such conditions. The combination of factors affecting wear and friction presented by Wieleba suggest that optimal tribological conditions will yield low wear in conjunction with low friction.

PAGE 24

15 Jintang [14] proposed that PTFE forms metal fluorides during sliding contact with stainless steel. He claims extremely complex chemical reactions caused by compression, tension, and shear may take place and allow PTFE its solid lubricant attributes. PTFE transfer is initiated when a polymer particle is removed from the bulk and adheres to the counterface. The particle then pulls the rest of the PTFE molecular chain from the bulk and eventually forms a film. Pleskachevsky and Smurugov [15] investigated the significance of thermal fluctuations on PTFE transfer film formation. They were able to show that the friction upon restart of a paused sliding test is dependent on the duration of the pause. After short pauses friction resumes the same as before the pause. As the pause duration increases the restart friction increases until it reaches the initial sliding value. Many researchers have suggested that the transfer film formation causes a drop in the friction value, however, after long pauses the polymer behaves as if there were no transfer film. Pleskachevsky and Smurugov suggest that the friction is more dependent on the PTFE/substrate interface temperature. They believe this explains the behavior of PTFE after pauses in sliding. They also observed that static and dynamic friction decrease with increasing test temperature. This is significant because there is no film transfer when static friction is occurring. Since both static and dynamic friction are affected, temperature can not be given credit for reducing friction by improving PTFEs self-lubricating properties. Pleskachevsky and Smurugov also show that test temperature becomes less significant as normal load increases. Other tests have shown that increasing load produces a thinner more efficient transfer layer. After failing to produce an irradiated PTFE/unirradiated PTFE composite with wear properties favorable to unirradiated PTFE, Blanchet and Peng [16] were able

PAGE 25

16 manufacture a composite PTFE/irradiated Fluoropolymer (FEP) with improved wear properties to PTFE. The composite contained 50% FEP irradiated to 30 Mrad and 50% unirradiated PTFE. Their experiments were run using a three-pin-on-disk configuration on a stainless steel counterface. They found the composite to wear considerably less and have a lower friction coefficient than PTFE under the same sliding conditions. The large plate-like wear debris commonly resulting from PTFE sliding is suspected to increase friction by adding to surface ploughing and deformation. The composite sliding yielded only thin oriented films. Blanchet and Peng suggest this is evidence that the wear debris morphology is indicative of friction. Also, the irradiated/unirradiated composite was non-abrasive to the steel counterface and possessed stronger creep resistance. Sui [17] demonstrated that friction coefficient of a PTFE composite under sliding has a non-linear dependence on speed. The composites friction was lower at slow speeds than at high speeds. They also confirmed a strong dependence on contact stress. In order to avoid severe wear they determined sliding must take place at speeds less than 20 m/s along with contact stresses below 0.5 Mpa. Contact stress was shown to decrease while sliding after an initial run in period. The stress is believed to drop with increasing contact width. PTFE wear rate data was collected from the aforementioned authors and is displayed in Table 2A below. Wear rates along with the major test parameters known to influence wear rate are displayed.

PAGE 26

17 Table 2-1. Collection of PTFE wear rate data from previous authors. Load (N)Pin GeometryV (m/s)Ra (um)K (mm /Nm) Avg.Distance (km)PaperCounterfaceTanaka et. al.Glass0.314.73 mm diameter9.83 mm diameterGlass* 6.6E-46.0E-45.6E-45.6E-45.7E-40.20.40.60.81.00.050.51.01.52.02.33.8E-42.2E-41.5E-41.2E-41.1E-40.20.51.01.51.759.84.9E-44.8E-44.9E-44.9E-4Sliding in atmospheric conditions and room temperatureCircular wear pathCircular wear pathCircular wear path3 mm diameter 14.714.714.7 14.70.30.30.30.33 mm diameter3 mm diameter3 mm diameter3 mm diameterGlassGlassGlassGlass9.89.89.89.80.050.050.050.053 mm diameter3 mm diameter3 mm diameter3 mm diameter3 mm diameter3 mm diameter3 mm diameterGlassGlassGlassGlassGlassGlassGlassGlass 9.89.89.80.20.20.2 Test run in vacuum Pre-rubbing of samplePin on dis k Tanaka et. al.Pin on dis k Tanaka et. al.Pin on dis k 79.60.53.755.910.614.413 mm diameter0.2Mild Steel7.6E-49.7E-47.5E-47.5E-417.00.20.751.42.12.93.66 mm diameter0.2Mild Steel1.2E-31.3E-31.3E-31.2E-31.3E-3 Briscoe combined rotationsLinear wear pathCircular wear path79.679.679.60.50.50.50.20.20.213 mm diameter13 mm diameter13 mm diameterMild SteelMild SteelMild Steel 17.017.017.017.00.20.20.20.20.20.20.20.26 mm diameter6 mm diameter6 mm diameter6 mm diameterMild SteelMild SteelMild SteelMild Steel Blanchet0.050.01304 S. Steel1044 mm X 4 mm1.1E-35.51040.010.14 mm X 4 mm0.01304 S. Steel9.6E-40.28.7E-40.38.3E-40.48.9E-40.51.0E-3Circular wear pathCircular wear path 1041041041040.010.010.010.014 mm X 4 mm4 mm X 4 mm4 mm X 4 mm4 mm X 4 mm0.010.010.010.01304 S. Steel304 S. Steel304 S. Steel304 S. Steel et. al.Blanchet et. al.Briscoeet. al.Briscoeet. al.Pin on diskPin on diskPin on disk Briscoe, Evans, Pelillo, and Sinha [18] used a scratching technique to investigate the energies related to surface deformation of polymers. Their research focused on the characterizing wear and adhesion of polymers that produce thin transfer films including PTFE. They indicate that the elastic properties of polymers such as PTFE impart a restoring force to asperities that scratch the polymer surface, and that the nature of this mechanism is a minimization of energy dissipated for material displacement. They also indicate that PTFE experienced low levels of strain and responded elastically to a blunt indenter.

PAGE 27

18 2.3 Wear of UHMWPE In regards to ultra-high molecular weight polyethylene (UHMWPE), Wang [19] produced a theoretical wear model. Wangs theory originated from the observance of fibrils composing most of the wear debris, and the importance of multi-directional sliding on wear rate. Multi-directional sliding leads to shear and tensile stresses acting in concordance. Any molecular orientation in the direction of one motion has weak resistance to motion in another direction. Wang states that if enough energy is put into the UHMPE surface a fibril may be pulled from neighboring fibrils. The work of friction must act perpendicular to the fibril alignment in order for wear to take place. Wang also states that there exists a critical value of coefficient of friction below which fibril removal will not occur. There also exists a critical cross-link density above which fibril removal will not occur. While improving the wear resistance of UHMWPE, crosslinking has also resulted in a decrease of material toughness. Muratoglu [20] investigated the crosslinking effects on UHMWPEs wear behavior with a pin-on-disk tribometer. They demonstrated that there was a linear relationship involving molecular weight between crosslinkings and wear rate. Muratoglus findings agree with previous findings that wear rate decreases with increasing crosslinking density. They also state that crosslinking affects the polymers ability to orient its molecular chains while sliding. Therefore, crosslinking improves wear behavior under multi-directional sliding. In their review of current concepts in wear of total hip and knee replacements, Schmalzried and Callaghan [21] state that polyethylene wear is different from creep. Creep deforms the polyethylene but does not produce any wear particles. Oxidation reduces the ability of irradiated UHMWPE to form crosslinks between molecular chains.

PAGE 28

19 Therefore, oxidized UHMWPE suffers from higher fatigue rates and delamination. UHMWPE wear is also highly sensitive to scratches on the counterface surface. Wear rate increases thirty to seventyfold when scratches of two micrometers in depth are present [21]. Wear tests by Burroughs and Blanchet [22] in 2000 showed that shelf-aged irradiated UHMWPE was less wear resistant than melt (200C) vacuum post-irradiation storage by approximately three-fold. The shelf-aged irradiated UHMWPE displayed wear similar to unirradiated UHMWPE under multi-directional sliding against polished surfaces. They also demonstrated that under multi-directional sliding UHMWPE does not undergo the initial run-in wear period observed during unidirectional sliding. Given the motion present in hip and knee joints, behavior during multi-directional sliding is of more concern than unidirectional sliding. The tests run by Blanchet and Burroughs consist of circular motion tests and rectangular motion tests. The wear rates for both these forms of testing proved comparable. This suggests that multi-directional motion of any kind is sufficient to produce higher wear than unidirectional motion. Suh [23] presented the delamination theory of wear in 1973 for metals. The theory is centralized around the assumption that subsurface material cold-works more than material near the surface due to a greater dislocation density in the subsurface. Cold working causes cracks to develop in the subsurface that eventually join together. When the cracks reach a critical length the material between the crack and the surface will shear and wear debris is thus produced. Suh also states that the same wear mechanisms are responsible for adhesive and fatigue wear. Although this theory was developed for metals, Briscoe [8] later indicated that delamination might play a role in polymer wear.

PAGE 29

20 PTFE wear debris morphology appears in a plate-like form that is characteristic of delamination. Delamination is driven by subsurface shear stresses that are not influenced by the direction of sliding only that sliding exists.

PAGE 30

CHAPTER 3 ENGINEERING APPROACH 3.1 Six-Station Pin-on-Flat Tribometer The objective of this project was to elucidate some of the wear mechanisms associated with load fluctuation and multi-directional motion of special polymers (PTFE). Testing was performed using a pneumatically load controlled six-station pin-on-flat tribometer developed at the University of Florida (UF) and presented in the thesis by Aaron Ison. Several PTFE wear tests were performed using the six-station test rig prior to commencement of this project to confirm the validity of this device as a tribometer. The test rig was equipped with linear voltage differential transducers (LVDT) in order to confirm translation and position of the movable stage. Load cells were used to verify that the pneumatic pressure devices supplied the intended load to the samples. The results of the preliminary testing indicate that the tribometer offers motion and load control accurate enough for this project. 3.1.1 Table and Drive System The six-station tribometer was designed to provide a means of creating multi-directional motion while allowing in-situ variable load capabilities. The sample counterfaces were translated using a multi-axis drive system. The system consists of a stage set atop two linear tables stacked on top of one another and configured in a perpendicular alignment. The top table is translated by a linear microstepper while the stage is translated along the top table by another linear microstepper. This allows the 21

PAGE 31

22 stage to travel anywhere in a horizontal x y plane. A controller manages the position, motion, velocity, and acceleration of the tables. Programs describing the motion path are sent from a computer to the controllers local memory. The controller then determines the necessary signals needed to produce the desired motion and sends them to the microsteppers. 3.1.2 Pneumatic Control Six pneumatic cylinders are used to apply normal load to the polymer samples. Pressure is supplied at 1.3 MPa by a 120 gallon compressor and is regulated by two 0 207 kPa pressure gauges. These pressure gauges can be seen in Figure 3.1 below. Figure 3.1. Photograph of pressure gauges (top) backpressure to cylinders (bottom) pressure to electro pneumatics. The top mounted pressure gauge supplies backpressure to the cylinders as a means of separating the polymer samples from the counterfaces. The bottom pressure gauge supplies air to six electro-pneumatic pressure transducers. Each electro-pneumatic is

PAGE 32

23 connected to one of the six pneumatic cylinders, and can supply pressure ranging from 0 138 kPa. A 0 5 Vdc signal sent to the electro-pneumatics controls the pressure supplied to the cylinders. The electro pneumatic arrangement can be seen in Figure 3.2. Figure 3.2. Arrangement of electro pneumatic gauges. The output pressure is linearly proportional to the voltage supplied. Each electro-pneumatic is controlled independently by separate signals sent from a computer that can either be programmed to fluctuate or hold constant. For all the tests presented in this paper, the pressure supplied to the electro-pneumatics was held constant at 207 kPa. Voltage signals ranged from 0 5 V depending on desired test parameters, and were used to produce output pressures ranging from 0 138 kPa sent to the cylinders. The signals can be manipulated to model any desired loading pattern within the limits of the electro-pneumatic pressure range and time response.

PAGE 33

24 The pneumatic cylinders are all aligned perpendicular to the movable stage in two rows of three. The arrangement can be seen in Figure 3.3. Figure 3.3. Photograph of pneumatic cylinder arrangement. The cylinders possess a 2-inch inner diameter and have a threaded stem used to attach the polymer sample holders. Each cylinder can apply a maximum load of 68 lbs (293 N) directly to the polymer samples. An illustration showing the entire pin-on-flat tribometer is displayed in Figure 3.4. Finally, the equipment log detailing the components needed to operate the tribometer is displayed in Table 3-1. 3.1.3 Sample Holders Six polymer sample holders were machined out of stainless steel to provide a rigid brace for the polymer samples. The sample holders were designed specifically for use with the pneumatic cylinders and can be viewed in Figure 3.5. A shop drawing of the sample holder can be viewed in Appendix C.

PAGE 34

25 Figure 3.4. Assembly drawing of entire pin-on-flat tribometer.

PAGE 35

26 Table 3-1. Equipment register for pin-on-flat tribometer. EquipmentManufacturer/Supplier(Product Number)QuantityDescriptionLinear TableParker Automation2Square rail bearing with linear(406100XRMS)AmplifierParker (Compumotor)2Amplifier for use with motors(OEMZL4)Pressure GaugeOmega2Regulate supply pressure toelectro pneumatics(PRG501-30)Motors & Motor toIndexer/Drive CableParker (Compumotor)2Load CellOmegadyne3Signle-axis load cell for measuringnormal load(LCKD-100)Connector BlockNational Instruments1Allows computer to collect data(SCB-68)ConditionerOmega3Condition voltage signal formLCKD-100 to load(DP25-S)Power SupplyOmegasnap2Convert wall voltage to 24 Vdc signal@ 850 mA to power electropneumat(DRN-PS-1000)Daq Board CableNational Instruments(SH68-68-EP)2Communication between computerand data aquistion boardsDaq BoardNational Instruments1Analog output board for controllingelectro pneumatics(PCI-6713)Daq BoardConnector BlockNational Instruments1Allows computer to send data(CB-68LP)National Instruments1Analog Input board for data collectio n Electro PneumaticTransducerOmega6Provides active control of supplypressure to cylinders(IP413-020)IndexerParker (Compumotor)1(6K8)8 axis indexer drivescrew8 Amp drive microstepper usedto translate counterfacesfrom tribometerto tribometer (PCI-6034E) Pneumatic CylinderBimba6Provide normal load ranging from0 63 lbs to the samples(NR-311-D)

PAGE 36

27 Figure 3.5. Schematic of polymer sample holder. The sample holders consist of a inch inner diameter tapped to fit the cylinder stems with a inch thru hole cut through the centerline. The inch diameter runs inch deep and bottom shelf flattened by an end-mill. The inch hole is cut through the shelf and is inch deep. The large diameter end of the polymer sample fits snuggly inside the sample holder with the polymer stem extending through the inch hole. The inch hole allows the polymer to contact the counterface and lends side support to the polymer stem during sliding. The sample holders are numbered along with the cylinders for bookkeeping purposes. 3.2 Motion Paths and Loading Patterns All sliding tests were performed in air under several motion paths and loading patterns. The first motion path consisted of simple linear reciprocating through a stroke

PAGE 37

28 length of 40.6 millimeters and an average sliding velocity of 46.5 mm/s. This motion was used for several wear tests and in each case ran for 4 hours producing a sliding distance of 670 meters. For the first series of tests all six cylinders were held at constant pressure. The test pressures at the cylinders were 56, 84, and 112 kPa corresponding to loads of 117, 176, and 235 N respectively. These loads equate to nominal pin pressures of 3.69, 5.56, and 7.42 Mpa. The second series of tests consisted of the same motion path, but applied loads oscillating at 6 second cycles. The range of load oscillation was different for each test. The first test oscillated from 148 206 N the second 117 235 N and the third 59 295 N. Each of these tests had an average load of 176 N. Finally a third group of tests consisting of a random selection of load peaks and valleys cycling once every 110 seconds was run with this motion path. The loading spectrum used in this test was modeled after a rain-flow spectrum presented by J. A. Collins [26], but was modified to achieve an average load of 176 N. A graph of the loading spectrum is shown in Figure 3.6. A second linear reciprocating motion path was used to provide the longest possible stroke length that could be performed with the current counterfaces. The path ran diagonal to the rectangular counterface with each stroke spanning a distance of 63.5 millimeters. Again, the test ran for 4 hours but was operating at a sliding velocity of 48 mm/s and a total sliding distance of 690 meters. The loading used for this test was held constant at 176 N. Three diamond patterns with varying degrees of crossing were also used as motion paths. The first diamond path ran at 30 of crossing with lengths of 16.6 millimeters.

PAGE 38

29 Figure 3.6. Loading spectrum applied to samples for wear testing. The second diamond path ran at 60 of crossing with lengths of 16.2 millimeters, and the third diamond path ran at 90 of crossing with lengths of 15.8 millimeters. All of the diamond pattern tests were run for 4 hours with all six cylinders at a constant load of 176 N. The path lengths were calculated to match the wear path areas for all diamond patterns to the diagonal reciprocating test. The change in length of the diamond sides is due to the increase in the polymer pins wear path area contribution. Due to the pins circular geometry, an increase in degree of crossing causes an increase in wear path area whenever a change in sliding direction takes place. Finally, five circular motion paths were generated with different diameters. Each test was run at a sliding velocity of 50 mm/s for 4 hours resulting in a sliding distance of 720 meters. For each of the five circular motion paths tested, all six cylinders were functioning with constant loads being applied to the polymer samples. However, the load

PAGE 39

30 was different from cylinder to cylinder. In groups of two, the cylinders applied loads of 117, 176, and 236 N respectively. The diameters for the five circle patterns were 6.35, 10.6, 15.0, 25.4, and 36.4 millimeters respectively. Figure 3.7 displays the motion paths described in this paper. Figure 3.7. Motion paths used for wear testing. The programs used to generate these motion paths are located in Appendix A. 3.3 Counterface Preparations and Handling The counterfaces were cut from 440 stainless steel bar stock. Counterface material was chosen to provide a significantly harder surface than that of the polymer sample. Due to the high hardness of 440 stainless steel (Rockwell 54 C), wear of the counterface could be neglected while sliding against the much softer polymer samples (PTFE 58 R, UHMWPE 63 Shore D). The bar stock was cut into 6 rectangular plates 3 X 2.75 X 0.125 inches. The counterfaces were then prepared by a series of polishing steps as described in Table 3B., beginning with 220 grit sandpaper and ending with a 0.3 m particle slurry.

PAGE 40

31 Table 3-2. Polishing steps for raw counterface samples. After polishing, the counterfaces were washed using water and a mild detergent then rinsed with methanol. Finally the surfaces of the counterfaces were characterized using an optical profilometer. Based on data collected from the profilometer the polishing technique described above yields an average surface roughness of approximately 0.02 m. Following each wear test, the counterfaces were washed with water and detergent then polished, cleaned, and characterized using the steps described in Table 3C. Table 3-3. Polishing steps for used counterface samples. Each of the six stainless steel counterfaces are constrained to the movable stage by four 4-40 facets, and are numbered to correspond with the cylinders and sample holders. 3.4 Polymer Sample Preparation The PTFE samples were cut from molded inch rod stock of virgin Teflon. Receiving coordinate information from the G code presented in the appendix of this paper, the samples were cut using a CNC mini milling machine. A schematic of the

PAGE 41

32 polymer sample can be viewed below in Figure 3.8 or a shop drawing can be viewed in Appendix C. Figure 3.8. Schematic of polymer sample. The polymer sample has a large diameter base of inches and is approximately inches in thickness. The stem has a smaller diameter of inches and extends 2/5 inches from the base. Allowing the stem to extend 2/10 of an inch from the sample holder. The milling machine ensures that the base shelf and the stems top surface are parallel to each other. The samples base shelf sits flat against the inner surface of the sample holder. The samples stem extends through the hole cut in the sample holder to contact the counterface surface. The sample is locked into position when the sample holder is threaded onto the cylinder. The UHMWPE samples were cut in the same manner and with the same dimensions as the PTFE samples. The unirradiated UHMWPE was cut from 1-inch stock samples 3/8 inch in diameter. The irradiated UHMWPE was cut from the interior of a shelf aged puck shaped sample. Prior to each wear test, the samples were placed in their respective sample holders and weighed individually. Once the wear test was completed, loose debris was removed from the polymer stem. The samples would

PAGE 42

33 remain inside the sample holders as they were weighed. Mass loss during the wear test could then be calculated and used to determine wear rates.

PAGE 43

CHAPTER 4 EXPERIMENTAL RESULTS 4.1 Electro-Pneumatic Performance Data Load cell data collected to evaluate the performance of the electro-pneumatics is displayed in Figure 4.1. Input SignalLoad Cell Output Figure 4.1. Electro-pneumatic performance data output from load cells. The electro-pneumatics demonstrate the ability to follow a sinusoidal loading cycle that ranges from 0 to 80 % of maximum pressure and cycles once every 6 seconds. The 34

PAGE 44

35 conditioners used to translate data received from the load cells were not capable of handling frequencies over 0.333 Hz. Therefore, there was inconclusive evidence regarding the electro-pneumatics ability to follow a 3 second cycle that spanned 0 to 80 % of the max pressure range. To ensure reliable testing parameters, all dynamic loading tests performed were constrained to frequencies equal to or below 0.167 Hz. 4.2 Variations in Wear Rate and Sliding Conditions Along with the respective sliding conditions, the wear rates from every PTFE wear test performed are shown in Table 4-1. The raw data collected for these calculations can be viewed in Appendix B. 4.2.1 Transfer Film Formation Although the tests involving PTFE were fairly consistent, data outliers occasionally appeared in the results. Out of the 93 tests run, where mass loss was the measured quantity, only 5 data points varied by more than 1 standard deviation from the mean for that test. The outlier appeared as an unexpectedly high wear rate 2 times, and appeared as an unexpectedly low wear rate 3 times. Unwaveringly, the appearance of outliers corresponded with two very distinct transfer films. Under reciprocating motion, high and normal wear always corresponded with the transfer film shown in Figure 4.2. The film appeared patchy and uneven with portions of the counterface still exposed. The patches appeared to be drawn out in the direction of sliding, but vary in width and length. Some patches even appeared to be deposited on top of a previous patch. In contrast, low wear always occurred in conjunction with the appearance of the smooth transfer film shown in Figure 4.3.

PAGE 45

36 Table 4-1. Raw data with calculated wear rates.

PAGE 46

37 Although these transfer films displayed regions of light and dark patches that appeared to run in the direction of sliding, there is no gross exposure of the underlying counterface. Unlike the high and normal wear films, the low wear films appeared very smooth and even. However, the regions of light and dark patches most likely indicate regions of varying thickness within the transfer film. The wear debris associated with both kinds of transfer films are plate-like in geometry with most of the longer debris strips folded in an accordion fashion. However, the wear debris appeared slightly smaller under low wear than it did under high wear. Another glairing contrast between the two films is the extension of transfer film over the wear path. For all reciprocating motion tests, the wear path consists of a long portion of constant width capped by semi-circular portions at both ends. In cases of high and normal wear, the transfer film extended only over the long portion of the wear path and large amounts of wear debris were deposited at the end points of the wear path. Whereas low wear transfer films cover the entire wear path. Transfer films would also appear for circular sliding motion. However, the long drawn out patches that appeared during high and normal wear under linear reciprocating motion were now assembled into circular patches with diameters approximately equal to the width of the wear path. The transfer films were all consistent in appearance and no

PAGE 47

38 (a) (b) (c) (d) Figure 4.2. Optical micrograph of PTFE transfer film characteristic of high wear rates (a) wear debris (b) end of wear path (c) top middle (d) bottom middle. (a) (b) (c) (d) Figure 4.3. Optical micrograph of PTFE transfer film characteristic of low wear rates (a) end of wear path (b) top middle (c) bottom middle (d) middle.

PAGE 48

39 major difference could be observed in the wear debris. Several images of circular motion transfer films can be seen in Figure 4.4. (a) (b) (c) (d) Figure 4.4. Optical micrograph of PTFE transfer film deposited by 14.9 mm diameter circular wear path (a) top (b) bottom (c) left (d) right. Finally, images of the diamond motion transfer films are displayed in Figure 4.5. These transfer films possessed characteristics resembling that of linear reciprocating along the sides, meaning that the transfer film is consistent and appears drawn out in the direction of sliding. At the corners, motion comes to a stop then restarts in a different direction. This change in sliding direction is similar to what the polymer pin experienced at the end points of the linear sliding tests. At all the corners the transfer film appeared to breakup and deposit itself as small irregular shaped patches with sections of the counterface exposed. Large amounts of the plate-like wear debris observed in previous tests were deposited at the corners. Although the morphology of the transfer film

PAGE 49

40 appeared different from the circular sliding tests, the wear debris showed no appreciable difference from either the circular or linear sliding tests. (a) (b) Figure 4.5. Optical micrograph of PTFE transfer film deposited by diamond pattern (a) soft corner (b) sharp corner. 4.2.2 Wear Rate Comparisons Comparisons between several wear tests are shown in Figures 4.6, 4.7, 4.8, 4.9, and 4.10 respectively. The volume of material lost during a wear test, v, was calculated by dividing the samples mass loss by its density. The wear rate, k, was then calculated by dividing the volume loss by the total sliding distance and the average normal load applied during the test as shown in equation 8. mksd (8) Figure 4.6 shows the variation in wear rate as load increases under linear reciprocating sliding motion. Each data point represents a sample slid for 670 meters under constant load. The data shows an obvious trend that wear rate is proportional to load, but variations in wear rate for each load do exist. At 117 N, the wear rate varied from 1.56E-4 to 3.64E-4 mm 3 /Nm with a standard deviation of 35 percent. The average of all six samples was 2.5E-4 mm 3 /Nm. With the exception of a single outlier at 235 N, the wear behavior is more consistent at 176 and 235 N with standard deviations of 10 and 18 percent, and average wear rates of 4.72E-4 and 5.05E-4 mm 3 /Nm respectively.

PAGE 50

41 050100150200250 Figure 4.6. Wear rate as a function of load for 670 meters of linear reciprocating sliding. The proportionality of increase in wear to increase in load for identical motion paths is similar to data presented by Tanaka [1]. For the same reciprocating wear path and sliding distance, data presented in Figure 4.7 indicates that oscillating load does not produce any change in wear behavior from that of constant load. Regardless of the magnitude of oscillation, the wear rates remained comparatively close to previously reported wear rates produced under a constant load equal to that of the oscillating loads average value. The average wear rates were 5.01E-4 for 58 N of oscillation, 4.06E-4 for 118 N of oscillation, and 4.57E-4 mm 3 /Nm for 234 N of oscillation. However, wear rates for the loading spectrum shown in Figure 3.6 of the Engineering Approach section of this paper were significantly higher than those produced with a constant load averaging 176 N. The average wear rate of all six samples was 6.35E-4 mm 3 /Nm. The load fluctuations in Figure 3.6 act at nearly the same

PAGE 51

42 frequency as those applied in the oscillating load test, however, the loading is more heavily weighted at the beginning of the cycle. Figure 4.7. Effects of varying load on wear rate compared with effect of loading spectrum on wear rate. Following the progression of wear rate and load shown in Figure 4.6, the wear results for spectrum loading more closely resemble those expected for a constant load of 293 N. A load of 293 N was applied for brief moments during the spectrum test. For the samples used in this test, 293 N of load borders on exceeding the yield strength of the material and may have altered the materials wear behavior. Figure 4.8 shows the effect of diameter on the wear rate for samples slid 720 meters in a circular motion path. As with reciprocating motion, the three constant load values were applied. Each test indicated that wear again increased with increasing load.

PAGE 52

43 Figure 4.8. Effects of load and diameter on wear rate for circular motion. An overall view of the circular motion tests indicated that wear increased at a fairly linear rate with increasing diameter. This behavior agrees with data presented by Briscoe [9] and what was expected given the increase in wear path area. Although Tanaka [1] provided data showing higher wear rates for smaller diameter circular wear paths, the tests loading and sliding speed were not identical. The only exception to the increasing wear with increasing diameter trend presented in this paper appears for the 36.4 mm diameter test when the wear rate decreased for 117 and 235 N loads from the 25.4 mm diameter test. All circular motion tests were preformed at the same sliding velocity. Therefore, the number of cycles increased with decreasing diameter. For tests with the same wear path area, wear rate proved to be higher under linear reciprocating motion than it did for a circular motion path. As can be seen in Figure 4.9, the wear rate for

PAGE 53

44 circular sliding motion does not become equal to that of linear sliding motion until the circular wear path area is nearly 1.75 times greater than the linear wear path. Figure 4.9. Ratio of circular motion wear rates over linear reciprocating motion wear rates. The wear rate ratios were calculated by taking the average wear rate from a circular motion path of given diameter and dividing that value by the average wear rate from linear reciprocating motion of the same load. The area ratios were calculated by determining the wear path area of a given diameter test and dividing that value by the wear path area created from a single pass under linear reciprocating motion. Although this data is consistent with findings by Briscoe [9], increasing wear with increasing diameter is in contrast to what was expected given the current theories regarding polymers such as UHMWPE and multi-directional sliding. However, this trend may be explained by the increasing number of cycles incurred by the polymer pin with decreasing wear path diameter and assuming a directional independent wear mechanism.

PAGE 54

45 Figure 4.10 shows the change in wear rate as a function of the angle of inclusion for several diamond shaped wear paths. All of the diamond pattern tests were run with a constant load of 176 N for a total of 4 hours. Due to changes in the motion path from test to test, the total sliding distance for 0 angle of inclusion differed from the other tests. The 0 angle of inclusion test slid for 695 meters while all other tests slid for 621 meters. The distances correspond to sliding speeds of 48.3 and 43.1 mm/s. Despite the small difference in speed, it is important to note that the higher sliding speed may result in an increase in wear relative to lower sliding speed tests. However, it is safe to assume that whatever change in wear rate resulted from different sliding speeds, it is not significant enough to produce a change in any trend observable in Figure 4.10. For 0 angle of inclusion the six samples had an average wear rate of 5.64E-4 mm 3 /Nm. The average wear rate at 15, 30, and 45 of inclusion were 6.82E-4, 6.08E-4, and 4.51E-4 mm 3 /Nm respectively. Although there was excellent consistency in the data from each set of tests, standard deviations of 4.78, 3.29, 1.74, and 3.88 % respectively, there was no consistency in the wear behavior from test to test. Wear rate increased from 0 to 15 of inclusion then decreased slightly at 30 of inclusion and finally decreased sharply at 45 of inclusion. With the exception of the 45 angle of inclusion, the wear rates collected under diamond shaped sliding motion were higher than any other wear test run with the same load. The wear path area was 4.35 cm 2 for all four tests. This wear path was 1.5 times greater than the wear path created under the previous linear reciprocating motion tests. However, wear rate was only shown to increase by this much when 15 of inclusion was incurred, and at 0 of inclusion wear rate increased by only 1.2 times that of the previous linear reciprocating tests.

PAGE 55

46 Figure 4.10. Wear rates for diamond pattern sliding motion as a function of inclusion. 4.3 Cycle dependence on Wear Wear rate shows an inverse dependence when plotted against the number of cycles incurred during sliding as shown in Figure 4.11 below. Regarding circular path wear tests, the least number of cycles run was approximately 8000 and corresponded to the highest wear rate for any such motion. As the number of cycles increased, the wear rate asymptotically approached some value around 0.18 x 10 ^-3 mm 3 /Nm. A similar trend appeared with the linear reciprocating and diamond path tests. Although the wear rates for this group were generally shifted up from the circular path wear curve, the final data point at 52000 cycles fell below what was expected from the circular path wear curve.

PAGE 56

47 00.10.20.30.40.50.60.70.8n 10,00020,00030,00040,00050,00060,00 0 reciprocatinghalf-circlefull-circle15 30 45 experimental data from tests with constantload and equal sliding distance curve-fit Figure 4.11. Wear rate as a function of number of cycles. 4.4 Images of Wear An image of the polymer samples before and after wear testing can be seen in Figure 4.12. Figure 4.12. Photograph of polymer samples before 4 hour linear reciprocating wear test at 176 N (left), and after (right).

PAGE 57

48 The amount of PTFE wear resulting from different loads after being ran for 720 meters at 50 mm/s in a circular sliding motion of diameter 36.4 mm can be seen in Figures 4.13 and 4.13. Figure 4.13 shows the sample holders marked 0 and 3 run at 117 N, holders marked 2 and 5 run at 176 N, and holders 1 and 4 run at 235 N. Figure 4.13. PTFE wear post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235 N, right 176 N. Figure 4.14 shows the PTFE wear paths resulting from the same test. Figure 4.14. Wear paths post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235 N, right 176 N.

PAGE 58

CHAPTER 5 SURFACE CHARACTERIZATION AND SUBSURFACE STRESS MODELING Based on data collected from optical stylus scans of the steel counterface prior to wear testing, the surface was assumed to be sinusoidal in nature with a 40 nm amplitude and a 30 m period. The optical scan information supporting these numbers is shown in the appendix of this report. The scan filtered out data with frequencies higher than 100 cycles per millimeter. This frequency was iteratively chosen because it gave a roughness value approximately equal to the average roughness of the entire counterface while eliminating any sharp asperities that would not actually support a significant amount of the normal load. The asperity density was modeled by assuming the pit-centered configuration shown in Figure 5.1. The area per asperity peak was calculated using equations 9 and 10. 2 (9) 2 (10) Where is the counterface area per asperity peak and is the length of one side of the square shown in Figure 5.1 above. The number of asperity peaks in contact with the polymer pin at any time was calculated using equation 11. pA (11) 49

PAGE 59

50 Figure 5.1. Asperity peak configuration at the surface of the steel counterface. Where is the number of asperity peaks in contact with the polymer pin and A p is the area of the pin face. Once the number of asperity peaks in contact with the pin is known the load per peak, f asp can be calculated by dividing the normal load by the number of peaks The peak radius, Rcp, shown in Figure 5.2 can be calculated before determining the Hertz elastic contact patch by using equation 12. 2124cpc R hh (12) Assuming that the polymer surface is identical to the counterface surface, the composite radius becomes 2cp R R (13) R can now be used with the Hertz contact patch equations to determine the size of the

PAGE 60

51 ch RcpAsperity Peak Figure 5.2. Radius of asperity peak at surface of steel counterface. contact patch between the polymer and counterface. 13'34asp f RaE (14) Where a represents the radius of a circular contact patch and E is the composite modulus of elasticity for steel and PTFE. R was shown to be orders of magnitude larger than the contact patch therefore confirming the Hertzian contact assumptions. The maximum pressure is then calculated using equation 15. max232aspfPa (15) The pressure profile along the surface of the polymer pin is shown in Figure 5.3. Given the low friction value between PTFE and steel, distortion of the pressure profile can be neglected once sliding has occurred [27]. The equation describing the pressure profile shown in Figure 5.3 is substituted in place of P(s) in the stress equations for an elastic half-space 16,17,18.

PAGE 61

52 Polymer PinSteel Counterface Direction of Sliding Incomplete Contact atPin/Counterface Junction Pmax 1a2s2 2s sa Figure 5.3. Pressure profile applied to PTFE surface when in contact with counterface. 232222222aaxaaPsxsPsxszdsdsxszxsz 2 (16) 3222222222aazaaPsPsxszzdsdsxszxsz (17) 2222222222aaxzaaPsxsPsxszzdsdsxszxsz (18) Where x and z describe the stress in the x and z directions while xz describes the shear stress in the xz plane. When the two materials in contact have modulus of elasticity an order of magnitude different, the presence of traction may cause a distortion of the pressure profile. However, the pressure profile here is assumed to be unaffected since the friction coefficient, of PTFE sliding against steel is not significant enough to distort the pressure profile from that shown in Figure 5.3. To analyze the subsurface affects of the asperities, a section of polymer 150 m long by 200 m deep was chosen to represent the entire polymer pin. Given the spacing of asperities along the surface, six asperities are in contact with this section of the pin.

PAGE 62

53 The section of polymer was discretized into regions 6 m long by 10 m deep. Using the stress equations above, the stress resulting from one of the six asperities was calculated for each of these regions within the main section of polymer. This process was repeated for all six asperities at their respective locations along the surface. The location of the regions remained the same while the distance to the corresponding surface asperity changed. Each asperitys increase in stress on a specific region was summed together to attain the combined stress of all six asperities on that region. Once the total stress of every region was calculated the results were plotted as shown in Figure 5.4, 5.5, and 5.6. Once the original stress and shear state was calculated, the Mohrs circle technique was implemented to find the maximum shear stress within the subsurface. The maximum shear stress at a given depth is given by 2xzAvg (19) 21xAvgR (20) (21) 22xzR 2212max R R (22) Where max is the shear stress of the maximum shear stress element. Once max is calculated for every location the results were plotted against the Z-axis to reveal the depth at which maximum shear stress occurs. These results can be seen in Figure 5.7.

PAGE 63

54 Figure 5.4. Sigma X compressive stress in subsurface of PTFE.

PAGE 64

55 Figure 5.5. Sigma Z compressive stress in subsurface of PTFE.

PAGE 65

56 Figure 5.6. Tau XZ shear stress in subsurface of PTFE.

PAGE 66

57 Subsurface Depth (m)(Pa) m Figure 5.7. Plot of subsurface shear stress along x = 0 indicating shear max.

PAGE 67

CHAPTER 6 DISCUSSION 6.1 Delamination Unlike UHMWPE, PTFE appears to wear at approximately the same rate regardless of any multi-directionality in the sliding motion. Results presented in tests 1-3 and 9 of Table 4A and Figure 4.8 indicate that there is considerable data overlap between wear rates resulting from unidirectional sliding and those resulting from multidirectional sliding. Images from the optical microscopes show that as PTFE wears it tends to form debris as thin plate shaped flakes of material. Wear of UHMWPE is often described as a surface wear process that is dependent on the orientation of the molecular chains relative to the direction of sliding. UHMWPE exhibits low wear under unidirectional sliding because the molecular chains eventually align themselves with the direction of motion, making the chains more difficult to remove from the bulk. The chains are more easily removed when experiencing shear, as is the case under multidirectional sliding. PTFE appears to wear as a result of a subsurface process known as delamination. Delamination occurs when a subsurface crack propagates long enough to linkup with other subsurface cracks until eventually one crack large enough to break from the bulk is present. The delamination process is depicted in Figures 6.1, 6.2, and 6.3 below. Once sliding begins stresses stemming from the combination of normal load and traction develop within the polymer pin. The stress equations used to model the subsurface stresses experienced by the polymer pin were presented in chapter 5 and have no 58

PAGE 68

59 Steel CounterfaceSubsurface CracksDirection of MotionNormal Load Polymer Pin Figure 6.1. Presences of subsurface cracks within polymer pin under stress. Steel CounterfaceDirection of MotionNormal Load Polymer Pin Subsurface Cracks Propagating Figure 6.2. Subsurface cracks begin to propagate and link up.

PAGE 69

60 Steel CounterfaceDirection of MotionNormal Load Polymer Pin Subsurface Cracks PropagatingDelamination Delamination Wear Flakes Figure 6.3. Ejections of polymer wear debris resulting from large subsurface crack. dependence on direction of motion. Therefore, the delamination process has no dependence on direction of sliding. This accounts for the lack of any directional dependence in PTFE sliding wear. Delamination may not influence UHMWPEs wear process because of the extremely low friction values incurred under lubricated sliding. The pressure profile applied to the polymer surface is the main factor influencing the subsurface stress. The pressure profile depends on both loading and surface topography. For the experiments pertaining to this paper, the only external load applied to the polymer was the normal load. Therefore, modeling the surface topography became the key factor when describing the pressure profile. Depending on the state of stress present in the polymer bulk compression, tension, and shear may be exuded on the imperfections present near the polymer surface. The imperfections include cracks that begin to propagate under either tension or shear. The state of stress depends on the loading and surface features at the polymer/counterface junction. Based on the surface model describing the counterfaces used during wear testing, the stress equations indicate

PAGE 70

61 that both compression and shear are present within the subsurface. Crack propagation may take place by either one or both of the two modes shown in Figure 6.4a, and b. x y TensionCrack a) xyCrackShear b) Figure 6.4 Modes of crack propagation a) mode I b) mode II. Tension, associated with mode I failure, will force the crack to pull apart and thus propagate. Shear will force the top and bottom halves of the crack to move in opposite directions. Mode II is characterized by motion perpendicular to the leading edge of the crack, whereas Mode III the motion is parallel. For the case describing both the counterface and polymer surface topography as sinusoidal with 40 nm amplitude and 30 m spacing, the subsurface stress model indicates the presence of shear and compression. Given a unidirectional sliding path, mode II propagation best agrees with the subsurface stress model, and indicates that the maximum shear takes place at about 10 m deep. The correlation between experimental data and theoretical modeling lend support to the theory of delamination as an explanation for the directionally independent wear behavior of PTFE sliding against a polished steel counterface. The subsurface shear present within the subsurface of the polymer results in mode II crack propagation. As the propagation begins to link up several cracks with one another it becomes easier to break a plate of material free from the bulk than to break the polymer/counterface junction. Once a section breaks from the bulk it is either ejected from the wear path as debris or deposited on to the counterface as a transfer film. Although delamination theory accounts for much of what was observed during PTFE wear, it is still unclear as to what mechanism causes

PAGE 71

62 the material to become part of a transfer film instead of ejected as debris. However, it is clear that a full description of PTFE sliding wear behavior must include an explanation for any third body interactions resulting from the transfer film. 6.2 Reversal Zones An inconsistency between wear tests that employ unidirectional sliding and those that employ multidirectional sliding involves reversal points in the motion path and the subsequent appearance of static friction. A reversal point is any location along the motion path where the polymer pin comes to a stop then restarts its motion in another direction. Such as the endpoints of a linear reciprocating wear path. These locations are of particular interest because they introduce static friction to the sliding process. Motion paths such as circles contain no reversal points since it is not necessary to stop the motion in order to change directions. As can be seen by comparing tests 1-3, 9, and 12 of Table 4A, under identical test conditions, tests that include reversal points have higher wear than tests with no reversal points. Tests 14-16 of Table 4A each include 4 reversal points and possess even higher wear rates than the tests with only 2 reversal points 1-3, and 12. However, the wear path area for tests 14-16 is greater than the area for tests 1-3, and 12. This is significant because an overall view of the data in Table 4A indicates that wear rate increases with increasing wear path area. A comparison could be made between test 13 and tests 14-16. Test 13 has the same wear path area as 14-16, but has only 2 reversal locations. On the average, the wear rate of 14-16 was higher than 13, but some data overlap did occur. Reversal points are identifiable on the transfer films as interruption areas of broken, patchy film. As stated in the results section of this report, the appearance of broken patchy transfer films was coincident with relatively high wear rates. For smooth

PAGE 72

63 transfer films with reversal points, the same could be said about wear rates over those areas of film interruption. Hence, wear tests have higher wear rates under smooth transfer film formation when reversal points are present in the sliding motion. 6.3 Cycle Dependence in Wear Rate Figure 4.10 of the results section shows that wear rate has a strong dependence on the number of cycles incurred during sliding. A simple model for this dependence is shown in Figure 6.5. number of cyclesncK2K1Transition of Wear Rates Figure 6.5. Model showing wear rate transition at some critical number of cycles n c This model depicts a transition from the initial high wear rate K 1 at some critical number of cycles n c to a lower steady state wear rate K 2 Therefore, any test that runs beyond n c will begin to appear more and more like the steady state wear rate. Using linear rules of mixing, a prediction of wear in terms of volume loss for single point measurements can be made using equation 23. 12()losscnncVnKFdKFdnn (23)

PAGE 73

64 Where d is the total sliding distance divided by the total number of cycles n. An expression based on wear rate of a single point measurement is given by equation 24. 21()losscspnVnKKKFDn 2K (24) Where K sp is the average wear rate over the entire test. Wear at any point during the test can be predicted if the number of cycles at that point is known.

PAGE 74

CHAPTER 7 CONCLUSIONS Changes in wear rate have been observed when changes in test parameters have been implemented. Factors influencing wear rate are load, wear path area, and number of cycles incurred during testing. The influence of number of cycles incurred indicates that wear is driven by the transfer film setup on the counterfaces. The calculated wear rate values from the pin-on-flat tribometer agreed closely with wear rate values calculated by previous authors. Preliminary testing of this tribometer indicated that the results are repeatable and that changes in wear rate can be attributed to changes in the testing parameters. Therefore, it is believed the tribometer used in this report is functioning properly and can be used to identify factors influencing wear. 1. The initial experiments with cyclic loading suggest that slowly varying cyclic loads have similar wear rates as produced by a constant load equal to the cyclic mean load. This has only been tested for a small range of load that has peak amplitudes that are within the same order of magnitude as the mean load. 2. The qualitative competitive rate models previously proposed for PTFE appear appropriate for explaining the dependence on the number of cycles on wear, but not for the development of the transfer film. 3. The number of reversals in a given wear path is related to wear rate. 4. The directionality of sliding shows significant differences in transfer film morphology within the reversal zones, but does not show significant differences in overall wear rate as compared to linear reciprocating sliding. There is no satisfactory explanation for how the transfer film develops. Questions concerning transfer film development as well as wear debris growth and expulsion still 65

PAGE 75

66 remain. A system designed specifically to observe the transfer films development under similar test conditions to those described in this paper is needed to help reveal useful information regarding wear behavior and would serve well as a point of future research. Such a system could be implemented on an existing tribometer with the addition of video imaging equipment and a proper scope setup.

PAGE 76

APPENDIX A MOTION PATH PROGRAMS ************************************************************************ constant ;Program for linear reciprocating motion path ************************************************************************ del constant ;Clear controller memory of previous program def constant ;Define new program drive10000 ;Activate tables ma00000 ;Set table to absolute coordinates a200,200 ;Set accelerations and decelerations ad200,200 v9.8425 ;Set velocities l ;Initiate loop d-200000 ;Command table number of units to move go1 ;Initiate table motion d200000 go1 ln ;End loop end ;End program ************************************************************************ circle ;Program for circular motion path ************************************************************************ del circle ;Clear controller memory of previous program def circle ;Define new program Drive11000 ;Turn tables on pv9.8425,9.8425 ;Set path velocities and accelerations and decelerations pa200 pad200 l1000 ;Initiate loop parcop0,0,0,62500 ;Define circle end points (x,y) and center points (x,y) ln ;End loop ;Before this program will run type pcomp circle into the terminal and enter ;Then type prun circle and enter end ;End program 67

PAGE 77

68 ************************************************************************ cir ;Program for loading and running circular motion path ************************************************************************ del cir ;Clear controller memory of previous program def cir ;Define new program pcomp circle ;Compile circle program into controller memory l200 ;Initiate loop for 200 cycles prun circle ;Run circle program ln ;End loop end ;End program ************************************************************************ arc2 ;Program for semi-circular motion path ************************************************************************ del arc2 ;Clear controller memory of previous program def arc2 ;Define new program drive11000 ;Turn tables on l ;Initiate loop pv9.5 ;Set path velocities and accelerations pa1000 ;Set path accelerations and decelerations pad1000 prtol5 parcp-127328,0,63664 ;Define arc startpoints (x,y) and end points (x,y) ln ;End loop ;Before this program will run type pcomp arc2 into the terminal and enter ;Then type prun arc2 and enter end ;End program ************************************************************************ constant1 ;Program for 0 o diamond motion path ************************************************************************ del constant1 ;Clear controller memory of previous program def constant1 ;Define new program drive11000 ;Activate tables ma00000 ;Set tables in absolute/ incremental mode a200,200 ;Set acceleration to 200 revs/sec^2 ad200,200 ;Set deceleration to 200 revs/sec^2 v7.3159,6.5843,1,1,1 l ;Initiate loop d-232279,209051 ;Command table number of units to move go11 ;Initiate table motion d232279,-209051 go11 ln ;ends loop end ;end main

PAGE 78

69 ************************************************************************ diamond1 ;Program for 15 o diamond motion path ************************************************************************ del diamond1 ;Clear controller memory of previous program def diamond1 ;Define new program drive11000 ;Activate tables ma00000 ;Set tables in absolute/ incremental mode a200,200 ;Set acceleration to 200 revs/sec^2 ad200,200 ;Set deceleration to 200 revs/sec^2 l ;Initiate loop v5.3625,8.2534 ;Set velocity to 9.8425 revs/sec d-44793,68942 ;Command table number of units to move go11 ;Initiate table motion v8.7707,4.4665 d-73263,37309 go11 v5.3625,8.2534 d44793,-68942 go11 v8.7707,4.4665 d73263,-37309 go11 ln ;ends loop end ;end main ************************************************************************ diamond2 ;Program for 30 o diamond motion path ************************************************************************ del diamond2 ;Clear controller memory of previous program def diamond2 ;Define new program drive11000 ;Activate tables ma00000 ;Set tables in absolute/ incremental mode a200,200 ;Set acceleration to 200 revs/sec^2 ad200,200 ;Set deceleration to 200 revs/sec^2 l ;Initiate loop v3.0436,9.3601 ;Set velocity to 9.8425 revs/sec d-24791,76241 ;Command table number of units to move go11 ;Initiate table motion v9.6279,2.0442 d-78422,16651 go11 v3.0436,9.3601 d24791,-76241 go11 v9.6279,2.0442 d78422,-16651 go11 ln ;ends loop end ;end main

PAGE 79

70 ************************************************************************ diamond3 ;Program for 45 o diamond motion path ************************************************************************ del diamond3 ;Clear controller memory of previous program def diamond3 ;Define new program drive11000 ;Activate tables ma00000 ;Set tables in absolute/ incremental mode a200,200 ;Set acceleration to 200 revs/sec^2 ad200,200 ;Set deceleration to 200 revs/sec^2 l ;Initiate loop v.51732,9.8289 ;Set velocity to 9.8425 revs/sec d4106,78017 ;Command table number of units to move go11 ;Initiate table motion v9.8289,.51732 d78017,4106 go11 v.51732,9.8289 d-4106,-78017 go11 v9.8289,.51732 d-78017,-4106 go11 ln ;ends loop end ;end main

PAGE 80

APPENDIX B RAW DATA TestCylinde r Mass LossNormal LoadSliding Distanc e Speed 47 47 47 47 47 47 176176176176176176 47 47 47 47 47 47 47 47 47 47 (g)(N)(m)(mm/s)(mm) 71

PAGE 81

72 49 49 49 117117176176176176 117117176176176176 49 49 49 49 49 49 49 49 49 49 49 49

PAGE 82

73 47 47 47 48 48 48 43 43 43 43 43 43 43 43 43 47

PAGE 83

APPENDIX C SHOP DRAWINGS Shop drawing of aluminum base mounted to tribometer stage. 74

PAGE 84

75 Shop drawing of counterface mount.

PAGE 85

76 Shop drawing of polymer pin sample.

PAGE 86

77 Shop drawing of polymer sample holder. R0.38

PAGE 87

APPENDIX D SURFACE METROLOGY Surface profiles of steel counterfaces. 78

PAGE 88

LIST OF REFERENCES 1. Tanaka K., Uchiyama Y., and Toyooka S., The Mechanism of Wear of Polytetrafluoroethylene, Wear 23, pp 153-172 (1973) 2. Tanaka K.,and Miyata T., Studies on the Friction and Transfer of Semi-Crystalline Polymers, Wear 41, pp 383-398 (1977) 3. Schott M., Preparation and Properties of Highly Oriented Polytetrafluoroethylene Films, Synthetic Metals 67, pp 55-61 (1994) 4. Bodo P., and Schott M., Highly Oriented Polytetrafluoroethylene Films: A Force Microscopy Study, Thin Solid Films 286, pp 98-106 (1996) 5. Viswanath N., and Bellow D. G., Development of an Equation for the Wear of Polymers, Wear 181-183, pp 42-49 (1995) 6. Yang E-L., Hirvonen J. P., and Toivanen R. O., Effect of Temperature on the Transfer Film formation in sliding Contact of PTFE With Stainless Steel, Wear 146, pp 367-376 (1991) 7. Blanchet T. A., and Peng Y-L., Wear-Resistant Polytetrafluoroethylene via Electron Irradiation, Lubr. Eng. 52, pp 489-495 (1995) 8. Briscoe B., Wear of Polymers: An Essay on Fundamental Aspects, Tribol. Int. 14, pp 231-243 (1981) 9. Briscoe B. J., and Stolarski T. A., Transfer Wear of Polymers During Combined Linear Motion and Load Axis Spin, Wear 104, pp 121-137 (1985) 10. Vijayan K., and Biswas S. K., Wear of Polytetrafluoroethylene: Some Crystallographic Observations, Wear 150, pp 267-273 (1991) 11. Brainard W. A., and Buckley D. H., Adhesion and Friction of PTFE in Contact With Metals as Studied by Auger Spectroscopy, Field Ion and Scanning Electron Microscopy, Wear 26, pp 75-93 (1973) 12. Cheng X., Xue Y., and Xie C., Tribological Investigation of PTFE Composite Filled With Lead and Rare Earths-Modified Glass Fiber, Materials Letters 4198, pp 1-5 (2002) 79

PAGE 89

80 13. Wieleba W., The Statistical Correlation of the Coefficient of Friction and Wear Rate of PTFE Composites With Steel Counterface Roughness and Hardness, Wear 252, pp 719-729 (2002) 14. Jintang G., Tribochemical Effects in Formation of Polymer Transfer Film, Wear 245, pp 100-106 (2000) 15. Pleskachevsky Y. M., and Smurugov V. A., Thermal Fluctuations at PTFE Friction and Transfer, Wear 209, pp 123-127 (1997) 16. Blanchet T. A., and Peng Y-L., Wear Resistant Irradiated FEP/Unirradiated PTFE composites, Wear 214, pp 186-191 (1998) 17. Sui H., Pohl H., Schomburg U., Upper G., and Heine S., Wear and Friction of PTFE Seals, Wear 224, pp 175-182 (1999) 18. Briscoe B. J., Evans P. D., Pelillo E., and Sinha S. K., Scratching Maps for Polymers, Wear 200, pp 137-147 (1996) 19. Wang A., A Unified Theory of Wear for Ultra-High Molecular Weight Polyethylene in Multi-Directional Sliding, Wear 248 (2001), pp 38-47 (2001) 20. Muratoglu O. K., Bragdon C. R., OConnor D. O., Jasty M., H. Harris W. H., Gul R., and McGarry F., Unified Wear Model for Highly Crosslinked Ultra-High Molecular Weight Polyethylenes, Biomaterials 20, pp 1463-1470 (1999) 21. Schmalzried T. P., and Callaghan J. J., Current Concepts Review Wear in Total Hip and Knee Replacements, The Journal of Bone and Joint Surgery, Incorporated 81-A, pp 115-136 (1999) 22. Burroughs B. R.,and Blanchet T. A., Factors Affecting the Wear of Irradiated UHMWPE, Society of Tribology and Lubrication Engineers Presentation NY (2000) 23. Suh N. P., The Delamination Theory of Wear, Wear 25, pp 111-124 (1973) 24. Steijn R. P., The Sliding Surface of Polytetrafluoroethlene: An Investigation with the Electron Microscope, Wear 12, pp 193-212 (1968). 25. Gong D., Xue Q., and Wang H., Physical Models of Adhesive Wear of Polytetrafluoroethylene and its Composites, Wear 147, pp 9-24 (1991) 26. Collins J. A., Failure of Materials in Mechanical Design, 2 nd Edition, A Wiley-Interscience Publication NY, NY (1993) 27. Johnson K. L., Contact Mechanics, Cambridge University Press Cambridge, UK (1985)

PAGE 90

BIOGRAPHICAL SKETCH Darren McGuire was born the second son of Joseph and Louise McGuire September 2 nd 1976. Darren lived in Wappingger Falls, New York until the age of 11 when he and his parents moved to beautiful Flagler Beach, Florida. Darren began his engineering career at Santa Fe Community College and then transferred to the University of Florida where he received his Bachelor of Science degree. After receiving the University of Florida alumni fellowship award he went on to pursue his Master of Science in the field of tribiology. 81


Permanent Link: http://ufdc.ufl.edu/UFE0001150/00001

Material Information

Title: Spectrum loading and multidimensional sliding of PTFE with a pin-on-flat tribometer
Physical Description: Mixed Material
Creator: McGuire, Darren ( Author, Primary )
Publication Date: 2003
Copyright Date: 2003

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001150:00001

Permanent Link: http://ufdc.ufl.edu/UFE0001150/00001

Material Information

Title: Spectrum loading and multidimensional sliding of PTFE with a pin-on-flat tribometer
Physical Description: Mixed Material
Creator: McGuire, Darren ( Author, Primary )
Publication Date: 2003
Copyright Date: 2003

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0001150:00001


This item has the following downloads:


Full Text












SPECTRUM LOADING AND MULTIDIRECTIONAL SLIDING OF PTFE WITH A
PIN-ON-FLAT TRIBOMETER















By

DARREN MCGUIRE


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2003















ACKNOWLEDGMENTS

I would like to acknowledge all those that helped make the culmination of my

academic career a reality. I cannot express the debt of gratitude I feel nor stress the

importance of every person's contribution to my thesis project. However, I would at least

like to point out a few that made it possible.

I thank Howard Purdy for his machining efforts, Gregory Sawyer for his guidance,

knowledge, and friendship, Dr. Ziegert, Dr. Arakere, and Dr. Schueller for their review

of this thesis and assistance in the classroom and lab throughout the course of my entire

education, and members of the tribology lab for their suggestions, perspectives, and

support.

Most of all I would like to thank my parents and family. Emotions can leave words

far in the distance and I find none that can explain what they mean to me. I love them

more than life itself.
















TABLE OF CONTENTS
Page

A C K N O W L E D G M E N T S .................................................................................................. ii

LIST OF TABLES ............. ...... ...... ..... ...... ..... .......... ........

LIST O F FIG U R E S .... .............................. ....................... ........ .. ............... vi

A B S T R A C T .......................................... .................................................. v iii

CHAPTER

1 IN TRODU CTION ................................................. ...... .................

2 REVIEW OF LITERATURE ......................................................... .............. 4

2.1 Proposed Mechanisms of Wear for PTFE ......................................................
2.2 Attempted Improvements in Wear Characteristics of PTFE ........................13
2 .3 W ear of U H M W P E ............................................ ....................................... 18

3 EN GINEERIN G APPROA CH .............................................................. ............... 21

3.1 Six-Station Pin-on-Flat Tribom eter ...................................... ............... 21
3.1.1 Table and D rive System ............................................................... 21
3.1.2 Pneum atic Control ......... ......................................... ....................22
3.1.3 Sample Holders ........................... .... .... .......... .... .............. 24
3.2 M otion Paths and Loading Patterns ............................. ............................. 27
3.3 Counterface Preparations and Handling .................................. ............... 30
3.4 Polymer Sample Preparation .................................... ...............31

4 EXPERIM ENTAL RESULTS ............................................................................34

4.1 Electro-Pneumatic Performance Data ........................................................34
4.2 Variations in Wear Rate and Sliding Conditions...........................................35
4.2.1 Transfer Film Form ation...................................... ........ ............... 35
4.2.2 W ear R ate Com prisons ........................................ ...... ............... 40
4.3 Cycle dependence on W ear......................................... ......................... 46
4.4 Im ages of W ear .................................................. .... .. ........ .... 47









5 SURFACE CHARACTERIZATION AND SUBSURFACE STRESS MODELING49

6 D IS C U S SIO N ...................................... ............................................... 58

6 .1 D elam in atio n ........................................................................ ................... 5 8
6 .2 R ev ersal Z on es .................. ........................................... .......... ......62
6.3 Cycle Dependence in W ear Rate ........................... ....... ........................... 63

7 CON CLU SION S ........................................................ ..............65

APPENDIX

A M OTION PATH PROGRAM S........................................................ ............... 67

B R A W D A TA ...................................... ............................... ................. 71

C SH O P D R A W IN G S.............................................................. 74

D SURFACE METROLOGY ............. ............................................. 78

L IST O F R E F E R E N C E S ......... ................. ........................................ .......................... 79

B IO G R A PH IC A L SK E T C H ...................................................................... ..................81






























iv
















LIST OF TABLES


Table page

2-1. Collection of PTFE wear rate data from previous authors.....................................17

3-1. Equipment register for pin-on-flat tribometer. .................................. ...............26

3-2. Polishing steps for raw counterface samples. .......... ................ ..............31

3-3. Polishing steps for used counterface samples................................. ............... 31

4-1 Raw data w ith calculated w ear rates....................................... ......................... 36
















LIST OF FIGURES


Figure pge

2.1. V ertical pin-on-disk configuration................................................. ....... ........ 12

3.1. Photograph of pressure gauges (top) backpressure to cylinders (bottom) pressure to
electro pneum atics .............................................................................. ...... 22

3.2. Arrangement of electro pneumatic gauges. .................................... .................23

3.3. Photograph of pneumatic cylinder arrangement............ ............................24

3.4. Assembly drawing of entire pin-on-flat tribometer. .............................................25

3.5. Schematic of polymer sample holder............... ..................... ................... 27

3.6. Loading spectrum applied to samples for wear testing ..........................................29

3.7. M otion paths used for w ear testing ........................................ ......................... 30

3.8. Schem atic of polym er sam ple ........................ ............ .......... .................... 32

4.1. Electro-pneumatic performance data output from load cells ...................................34

4.2. Optical micrograph of PTFE transfer film characteristic of high wear rates (a) wear
debris (b) end of wear path (c) top middle (d) bottom middle............................. 38

4.3. Optical micrograph of PTFE transfer film characteristic of low wear rates (a) end of
wear path (b) top middle (c) bottom middle (d) middle .......................................38

4.4. Optical micrograph of PTFE transfer film deposited by 14.9 mm diameter circular
wear path (a) top (b) bottom (c) left (d) right ......................................................39

4.5. Optical micrograph of PTFE transfer film deposited by diamond pattern (a) soft
corner (b) sharp co er. ............................ .................. ............... .... ............ 40

4.6. Wear rate as a function of load for 670 meters of linear reciprocating sliding. ........41

4.7. Effects of varying load on wear rate compared with effect of loading spectrum on
w ear rate. ............................................................................42

4.8. Effects of load and diameter on wear rate for circular motion. ................................43









4.9. Ratio of circular motion wear rates over linear reciprocating motion wear rates......44

4.10. Wear rates for diamond pattern sliding motion as a function of inclusion .............46

4.11. Wear rate as a function of number of cycles.......................... ........................47

4.12. Photograph of polymer samples before 4 hour linear reciprocating wear test at 176
N (left), and after (right) ......... ................................................... ........... ......47

4.13. PTFE wear post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235 N,
right 176 N .......................................... ............................ 48

4.14. Wear paths post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235 N,
right 176 N .......................................... ............................ 48

5.1. Asperity peak configuration at the surface of the steel counterface ........................50

5.2. Radius of asperity peak at surface of steel counterface. .................... ...............51

5.3. Pressure profile applied to PTFE surface when in contact with counterface. ...........52

5.4. Sigma X compressive stress in subsurface of PTFE............ .................................54

5.5. Sigma Z compressive stress in subsurface of PTFE. ............................................55

5.6. Tau XZ shear stress in subsurface of PTFE................... ....................56

5.7. Plot of subsurface shear stress along x = 0 indicating shear max ...........................57

6.1. Presents of subsurface cracks within polymer pin under stress ..............................59

6.2. Subsurface cracks begin to propagate and link up ..............................................59

6.3. Ejections of polymer wear debris resulting from large subsurface crack ...............60

6.4 Modes of crack propagation a) mode I b) mode II. .................................................61

6.5. Model showing wear rate transition at some critical number of cycles nc ...............63

















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Master of Science


SPECTRUM LOADING AND MULTIDIRECTIONAL SLIDING OF PTFE WITH A
PIN-ON-FLAT TRIBOMETER

By

DARREN MCGUIRE

August 2003

Chair: W. Gregory Sawyer
Cochair: John Ziegert
Major Department: Mechanical and Aerospace Engineering

Wear is a process of gradual breakdown or removal of material by relative motion

between two contacting surfaces. In this report, the wear of polytetrafluoroethylene

(PTFE) under spectrum loading and multidirectional sliding was investigated. Wear tests

revealed a strong proportional dependence on load but showed no dependence on

sinusoidal fluctuations in load. Abrupt changes in the direction of motion, known as

reversal zones, correlated with increases in wear, meaning the more reversal zones

present on a motion path, the more severe the wear was compared to a motion path with

no reversal zones (circular). Multidirectional sliding appeared to have no influence on

overall wear. This differs greatly with what is known about the wear behavior of ultra

high molecular weight polyethylene (UHMWPE). The difference may stem from a

delamination process that takes place during PTFE sliding. Using data collected from a









white light interferometer, a model of the counterface surface was constructed, and when

applied to the polymer pin using Hertzian contact assumptions induced a pressure profile

that yields subsurface shear stresses. The magnitude of the subsurface shear stresses

shows no dependence on direction of sliding. This is offered as an explanation for the

lack of increased wear during multidirectional sliding over unidirectional sliding. Finally

a correlation between number of cycles incurred over a wear path and wear rate is

observed. A simple model using rules of linear mixing was constructed for wear

predictions based on single point mass loss measurements.















CHAPTER 1
INTRODUCTION

Wear is defined by the dictionary as "to impair, deteriorate, or consume gradually

by use or any continued process." From an engineering standpoint, wear maybe be

described as the gradual breakdown or removal of material stemming from relative

motion between two contacting surfaces. Wear of special polymers such as

polytetrafluoroethylene (PTFE) and ultra high molecular weight polyethylene

(UHMWPE) has become a major concern due to their effectiveness as biocompatible

materials. Despite the success of polymers as bearing surface for total knee or hip

replacements in treatment of end stage arthritis, the gradual breakdown of these surfaces

has led to component loosening and an eventual need for revision surgery.

The mechanisms that cause removal of material can be resolved into abrasive,

adhesive, and fatigue wear. Abrasive wear occurs when asperity peaks on the harder

surface scratch away material on the softer surface. Adhesive wear occurs when bonding

forces between the two surfaces are strong enough to pull material away from one or both

surfaces as relative motion occurs between them. Fatigue wear occurs when material is

removed as a result of cyclic stresses that exceed the fatigue strength of the material [50].

Each mechanism's contribution to wear is difficult to observe when all three factors are

acting simultaneously. Minimizing the wear mechanisms that are not of interest allows

the remaining mechanisms to dominate the wear process. Therefore, an observation of a









single desired mechanism or a combination of desired mechanisms affects could be

made.

Throughout this paper adhesive wear will always be a central focus. Several wear

tests will be run with the interest of investigating behavior of adhesive wear as well as

behavior of adhesive and fatigue wear in combination. The tests presented in this paper

all involve PTFE in contact with stainless steel counterfaces. For the case in which one

of the two surfaces in contact is significantly harder than the other surface, minimizing

abrasive wear requires that the harder surface be highly polished, thereby limiting the

material scratched away from the softer surface to an insignificant amount. Under dry

testing, wear rates for PTFE and UHMWPE are known to increase dramatically when in

contact with counter faces displaying and average roughness (Ra) greater than 0.1 tm.

Fatigue wear can be virtually eliminated by maintaining the load holding the two surfaces

in contact constant. However, some cyclical effects will be present as the frictional force

vector changes direction along with changes in the sliding path direction.

Wear rate is a key factor in characterizing a materials performance in wear testing.

If wear rate is known prior to testing, predictions about how much material will be lost

during the course of a test can be made. However, wear rate does not always remain

constant and may change depending on the testing parameters implemented. Although

the molecular structures for both PTFE and UHMWPE are similar, their wear behavior

differs drastically. UHMWPE wear is dominated by a surface phenomenon that is highly

dependent on changes in the direction of sliding. However, PTFE shows no directional

dependence and appears to wear from a subsurface phenomenon known as delamination.

A process where subsurface imperfections or cracks are forced to propagate due to






3


subsurface stresses incurred during sliding. The cracks propagate and link up with one

another forming larger cracks. These cracks continue to propagate and eventually turn

towards the surface breaking off a flake of wear debris. This process repeats through out

the duration of sliding. However, the wear process has been shown to depend on the

development of a transfer film. This indicates that the wear process is not completely

driven by the bulk material, but depends on the surface interactions between bulk and

counterface.















CHAPTER 2
REVIEW OF LITERATURE

2.1 Proposed Mechanisms of Wear for PTFE

With all the recent interest surrounding UHMWPE and its use in human joint

replacements many experiments have been compiled in hopes of developing a predictive

wear model. Sliding wear of UHMWPE is described as a surface wear phenomenon

where the molecular chains align themselves with the direction of sliding. Once the

molecular chains align themselves with the direction of sliding, wear rate drops

significantly. If the direction of sliding changes, wear rate increases until the chains can

reorient themselves and wear rate drops once again. If the direction of sliding is always

changing there is no preferred molecular orientation and wear rate remains significantly

higher than under unidirectional sliding. Although PTFE and UHMWPE have similar

molecular composition their wear behaviors differ drastically in that PTFE is considered

to have poor wear resistance.

Tanaka, Uchiyama, and Toyooka [1] performed early investigations into the wear

of PTFE. They observed that when slid against a glass plate, PTFE deposited a fibrous

thin film with long bands and striations perpendicular to the length of the bands. At a

sliding speed of 20 cm/s, the PTFE wear was shown to increase linearly with increasing

sliding distance while the friction measured during sliding initially decreased before

settling to a constant value. The initial drop off in friction is attributed to formation of

the PTFE transfer film. Tanaka later demonstrated that varying the sliding rate produced









a change in the PTFE wear rate. Sliding a 5 cm/s resulted in an initial high wear rate that

quickly transitioned to a slower wear rate. He also showed that at temperatures below

1000C the wear rates would peak for a given sliding speed, and that these peaks shifted to

higher speeds as the temperature was increased. Whereas for temperatures above 100C

the peaks tended to shift towards lower speeds as temperature was increased. Friction

decreased with increasing temperature and increased with increasing sliding speed.

Tanaka also used a thermocouple to measure the temperature rise at the sliding surface,

and concluded that the temperature rise due to frictional heating was insignificant. This

was attributed to the rapid removal of PTFE during sliding and the transfer film's aid in

heat dissipation. Inspection of the worn surfaces revealed that the same wear mechanism

of PTFE acted under all circumstances. The molecular orientation and the uniform

separations of striations in the transfer films indicate that the PTFE fibers are held

together by lateral connections. Tanaka states that the explanation for PTFE's high wear

rate stems from the lack of melting at the PTFE/Substrate interface and the easy removal

of film from the substrate. Inspection of the transfer film reveals that a thin layer of

amorphous PTFE is likely removed from the bulk during sliding. The shearing that

occurs during sliding causes slippage in the amorphous region and the PTFE is then

deposited in slices. Given this, Tanaka attributes the effect of sliding speed on friction

and wear to the likelihood of viscoelastic behavior between these slices. When abrasive

wear is present the slippage between slices does not occur and no film is observed. This

is due to the severe damage caused by abrasive wear.

In 1977 Tanaka and Miyata [2] published a study of PTFE cylinders slid on glass

plates. The study into the friction and transfer of semi-crystalline polymers reveled that









adhesion of the molecular chains in PTFE is stronger when the direction of sliding is

perpendicular to the orientation of the molecular chains. This was contradictory to

previously understood results. Tanaka and Miyata examined the transfer films after

several increments in traverse. They observed that the transfer film thickness increased

with increasing traverses. However, they limited the number of traverses to twenty.

They were also able to observe that the kinetic friction value remained constant for all

traverses. This suggests that shearing occurs within the PTFE transfer film. Tanaka and

Miyata also suggest that the banded structure of the PTFE film as well as adsorption of

water molecules on the PTFE molecular chains reduces shearing strength on the PTFE

surface. However, they go on to suggest that the friction behavior and transfer of PTFE

during sliding is mainly influenced by the molecular profile and not by the banded

structure. Tanaka previously found that film thickness formation of about 300 angstroms

is based on mutual slippage of the crystalline slices in the banded structure, but was

unable to apply these findings to the current experiments.

The prevailing model for PTFE transfer from bulk to substrate under reciprocating

conditions is that of very long, straight, and crystalline ribbons. Using a oxide-covered Si

wafer, M. Schott [3] observed that these ribbons cover nearly the entire substrate and

stated that the morphology of the films depends on the temperature at which sliding takes

place, the speed at which the PTFE bulk slides relative to the substrate, and the normal

load applied to the PTFE bulk while sliding. Schott chose to indicate the loading

conditions as weights in order to deduce the nominal pressures. Both the PTFE and

substrate were kept at the same temperature and sliding speed was kept low and smooth

(0.4 to 2 mm/s). Schott's experiments were conditioned by allowing the starting PTFE









material to be removed before the transfer film was formed. Using IR transmission data,

nuclear reaction analysis, and atomic force microscopy Schott was able to confirm the

presence of a highly oriented PTFE transfer film.

Bodo and Schott [4] did a second investigation into the formation of highly

oriented PTFE films in which a PTFE rod was slid against an oxide-covered silicon wafer

under controlled load and temperature. The PTFE samples were "conditioned" by prior

sliding to eliminate the initial rapid wear phase of PTFE. Schott again saw the classical

characteristics of a PTFE transfer film. However, he was able to generalize the films into

three different categories. The first group consisted of irregular ribbons and low

coverage (20% of wear path). Both the second and third groups showed long, straight,

and parallel ribbons, but the second group had incomplete coverage while the third had

complete coverage. Schott reported that the first group appeared during tests run at

temperatures under 1500C. The ribbons showed kinks and branching. He also states that

although the ribbons consisted of bundles of polymer chains running parallel to the

ribbons, the ribbons appeared considerably longer than the macromolecule in the PTFE.

Schott goes on to state that the ribbons are formed during or after film deposition. The

second type of film occurred during tests run above 1500C and the third occurred above

2200C and 600g loads. However, the coverage decreased to 76% when the load was

increased to 2425g. The film cross-sectional thickness increased only slightly with

increasing temperature. Schott acknowledges that film variations may result from

variations in sample-substrate contact. He suggests that at high temperatures and loads,

the variations diminish due to higher plastic deformation. Schott also reports that the

film morphology is independent of substrate material and is a property of PTFE only.









Using the concept of dimensional analysis, Kar and Bahadur developed an equation

describing the wear of polymers in 1974. The Kar and Bahadur equation was a function

of pressure P, speed v, time T, modulus of elasticity E, surface energy y, thermal

conductivity K and specific heat Cp. The dimensional analysis yielded an equation for

volume loss represented by:


V = kPxvyz T3y+aE 3x+y (1)
K

where x, y, and z are exponents determined experimentally. In 1995 Viswanath and

Bellow [5] extended the Kar and Bahadur equation to include counterface surface

roughness. Viswanath and Bellow again used dimensional analysis to derive a

dimensionally homogeneous equation that included the specified variables. Viswanath

and Bellow included five dimensionless groups comprised of wear volume V, surface

energy y, modulus of elasticity E, specific heat Cp, thermal conductivity K, contact force

W, counterface roughness a, sliding speed v, and time T. The groups signify dependence

on interface contact and deformation (VE3/y3) normal load and strength characteristics

(WE/y2), speed and temperature (TECp/K), counterface roughness (oE/y), and thermal

contributions (yCp/vK) yielding the dimensionless function:

(VE3 WE vK TEC aEE
Y3 2 7C 0 (2)


Viswanath and Bellow then applied a non-linear relationship to map their experimental

results.


V = kWpvqTra'E 3+p+r+s 3-2p-q-s P (3)
K









where p, q, r, and s are exponents determined experimentally. They found, for a pin-on-

disk configuration with steel counterfaces, Equation (3) allows a predictive model for

volume losses based on operating and material factors know to affect wear.

In 1990, Yang et al. [6] investigated the effect of temperature on the formation of

transfer films. Yang and Hirvonen used a pin-on-plate configuration to produce sliding

of a PTFE specimen on a stainless steel counterface. They used y-ray yield curves to

verify the amount of PTFE transferred during sliding. They were able to observe that the

deposition rate was higher during the first traverse of unidirectional sliding, and that

prolonged sliding could be modeling using a linear equation:

x = x, + u(n- 1) (4)

where x is the amount of PTFE transferred, xo is the amount transferred after the first

traverse, v is the steady-state transfer rate, and n is the number of traverses. They also

observed that transfer rate depended on temperature. Their normalized data indicated

that thickness increased less rapidly with increasing temperature as the number of

traverses was increased. They model this using the following equation:

x(T, n)= x(T,,n)g(T) (5)

where x(T,n) and x(Tt,n) are the thicknesses of the transferred PTFE after n traverses at

room temperature and temperature T, and g(T) is a temperature enhancement factor.

Despite variations in transfer film appearance at certain temperatures and qualitative

disagreement with transfer and PTFE hardness, Yang and Hirvonen concluded that

transfer of PTFE during the first traverse increases non-linearly by approximately an

order of magnitude when increasing temperature from room to 2000C. However, the









steady-state transfer rate at room temperature and 1000C are virtually the same while

transfer at 2000C is increased significantly.

Using a pin-on-disk configuration, Blanchet and Peng [7] published an

investigation into irradiated PTFE wear on stainless steel. Irradiating PTFE causes

primary and secondary radicals to form by severing the C-C and C-F bonds. The radicals

then recombine to form a cross-linked network that has been shown to offer better wear

resistance than the original linear PTFE molecular structure. Blanchet and Peng showed

that the wear rate of PTFE decreased as irradiation dose increased from 0 to 30 Mrad then

increased somewhat as the irradiation dose increased from 30 to 100 Mrad. They also

observed that the hardness of PTFE increased from Rh = 43 to 53 over the same 0 to 30

Mrad dose. Friction was also found to increase as the irradiation dose increased from 0

to 5 Mrad. However, friction was then observed to decrease as to dose increased to 20

Mrad then settled to a constant value that was comparable to the coefficient of friction of

unirradiated PTFE. In addition to these findings, they also observed a change in the wear

debris morphology. Unirradiated PTFE is known to produce large plate like wear debris

where as the irradiated PTFE yielded a very fine debris. Blanchet and Peng concluded

that the difference in properties between unirradiated and irradiated PTFE was due to the

branched and cross-linked network preventing molecular orientation. Molecular

orientations of PTFE chains result in low shear strength and the subsequent formation of

transfer films and wear debris.

Briscoe proposed that polymer wear falls into two categories known as cohesive

wear and interfacial wear [8]. Briscoe describes cohesive wear as wear involving

dissipation of frictional work such as abrasion and fatigue. The level of damage due to









sliding is prescribed by surface asperities and surface traction largely governed by the

cohesive strength of the polymer. Briscoe describes interfacial wear as wear involving

dissipation of frictional work by transfer and is more damaging than cohesive. He also

points out that the current research indicated a correlation between wear rate and the

reciprocal of the work to rupture (1/pySy) for a given polymer. Briscoe goes on to state

that polymers considered ductile, such as PTFE, are far less affected by rough surface

anomalies than glassy polymers such as PMMA, and that PTFE can accommodate more

strain before rupture. He gives an equation for ductile abrasion based on a theory of

asperity resistance:


S= tan (6)
zr tan )

KW tan (7)
H

Where [t is the coefficient of friction, 4 is the slope of a conical asperity, W is normal

load, H is hardness, and K is probability constant. In regard to PTFE, Briscoe also points

out that molecular orientation affects transfer wear, and that the transfer is largely

unchanged polymer. Most polymers form a thick lumpy film during transfer. However,

PTFE transfer is inconsistent forming either a relatively thin film or a patchy uneven film

once steady sliding has occurred. There also appears to be an upper limit for sliding

velocity at a given temperature for which transfer can occur. Briscoe also states cracking

and delamination may be responsible for the transfer film often appearing "rumpled" and

having patches of plate-like polymer removed from the film. He also states that this

debris acts like a lubricant in the case of PTFE. To conclude, Briscoe points out the need










for continued research but suggests that wear rate is a function of the material and a

dominant wear mechanism.

Briscoe [9] also investigated the effects of combined rotating and linear sliding

motions on the wear and transfer of PTFE. Briscoe contends that the increasing wear of

PTFE with increasing number of traverses is dependent on kinematics of relative motion

between the sample and the substrate surface. Wear is dependent on sliding distance, and

normally increases with increasing distance. For the vertical pin-on-disc machine shown

in Figure 2.1 wear rate actually decreases with increasing number of traverses.





pin
disc





Omega disc



Figure 2.1. Vertical pin-on-disk configuration.

PTFE wear shows a behavioral change depending on the angular velocity of the pin.

Effective linear sliding is achieved when the pin is rotated with the same angular velocity

of the disc but in the opposite direction. Wear remains a function of sliding distance, but

decreases as the pin's angular velocity is varied to simulate pure linear sliding to

rotational sliding with a radius equal to the radius of the PTFE pin. Briscoe showed that

the combination of spins inputs energy into the system that results in wear rate

fluctuations. A relative rotation between the pin and disk inhibits molecular orientation









and results in increased wear and friction. Briscoe also states that motion aligned with

polymer fibers leads to fibers being pulled away from the bulk.

Wear and friction of PTFE are known to undergo a transition at approximately 0.1

km. Vijayan and Biswas [10] observed a slight expansion in the unit cell up to this

distance of sliding and unit cell shrinkage with increased sliding distance. They also

observed a steady increasing in wear rate and friction until a steady-state wear and

friction were reached around 3 km of sliding. They were also able to confirm that no

change in the atomic structure of the PTFE surface took place over this distance.

Although Vijayan and Biswas contend there is a strong correlation between wear rate and

unit cell volume, they were unable to correlate the change in crystallite sizes to wear rate.

Brainard and Buckley [11] discovered that PTFE would transfer to metals under

static conditions if a normal mass of 100 g were applied (pin dimensions not given). Due

to the increased area of contact under load, it is not clear whether or not the transfer is

load dependent. They also were able to conclude that PTFE transfer is possible with or

without an oxide layer covering the metal substrate. Therefore, the chemical behavior of

the substrate is not significant. The adhesion force was non-existent during the first two

minutes of loading. Therefore, Brainard believes that adhesion is strongly influenced by

creep.

2.2 Attempted Improvements in Wear Characteristics of PTFE

Cheng et al. [12] improved the wear properties of PTFE by preparing a composite

with 60% wt. lead and 5% wt. glass fiber. Xue and Xie recorded friction coefficients and

mass loss of PTFE composite samples sliding on steel using a pin-on-disc configuration

with an oil lubricant. They concluded that PTFE containing modified glass fiber

demonstrated better wear resistance and possessed a lower friction coefficient than virgin









PTFE. They credited the lead and glass fillers with decreasing wear by controlling the

size and shape of the wear debris. The glass fiber also provides load support at the

sample/substrate interface that decreases wear of the PTFE composite.

There have been several attempts to optimize the surface features of sliding

counterfaces. Optimizing these features would achieve the lowest possible polymer wear

and friction. Wieleba [13] studied the frictional and wear relationship between PTFE

composites and the surface texture of steel counterfaces. While maintaining sliding

velocity, pressure, and sliding distance constant throughout all tests, Wieleba carried out

his experiments under dry sliding conditions. Using regression functions, Wieleba

showed that wear rate is more dependent on surface features than friction. The regression

functions predict that friction is most greatly influenced by the mean spacing between

asperities. The larger the spacing between peaks the higher the friction value became.

Wieleba suggests that large spacing makes formation of a PTFE transfer film more

difficult. He showed friction was similarly influence by average asperity slope. The

smaller the average slope the larger area of contact, consequently, adhesion forces

friction to increase. Wear rate was most strongly influenced by asperity peak heights and

the peaks' average radius of curvature. The higher the asperity peaks as well as the

sharper the peaks the more severe the wear. This is explained by the increase in

mechanical interactions between the counterface and the PTFE under such conditions.

The combination of factors affecting wear and friction presented by Wieleba suggest that

optimal tribological conditions will yield low wear in conjunction with low friction.









Jintang [14] proposed that PTFE forms metal fluorides during sliding contact with

stainless steel. He claims extremely complex chemical reactions caused by compression,

tension, and shear may take place and allow PTFE its solid lubricant attributes.

PTFE transfer is initiated when a polymer particle is removed from the bulk and

adheres to the counterface. The particle then pulls the rest of the PTFE molecular chain

from the bulk and eventually forms a film. Pleskachevsky and Smurugov [15]

investigated the significance of thermal fluctuations on PTFE transfer film formation.

They were able to show that the friction upon restart of a paused sliding test is dependent

on the duration of the pause. After short pauses friction resumes the same as before the

pause. As the pause duration increases the restart friction increases until it reaches the

initial sliding value. Many researchers have suggested that the transfer film formation

causes a drop in the friction value, however, after long pauses the polymer behaves as if

there were no transfer film. Pleskachevsky and Smurugov suggest that the friction is

more dependent on the PTFE/substrate interface temperature. They believe this explains

the behavior of PTFE after pauses in sliding. They also observed that static and dynamic

friction decrease with increasing test temperature. This is significant because there is no

film transfer when static friction is occurring. Since both static and dynamic friction are

affected, temperature can not be given credit for reducing friction by improving PTFE's

self-lubricating properties. Pleskachevsky and Smurugov also show that test temperature

becomes less significant as normal load increases. Other tests have shown that increasing

load produces a thinner more efficient transfer layer.

After failing to produce an irradiated PTFE/unirradiated PTFE composite with

wear properties favorable to unirradiated PTFE, Blanchet and Peng [16] were able









manufacture a composite PTFE/irradiated Fluoropolymer (FEP) with improved wear

properties to PTFE. The composite contained 50% FEP irradiated to 30 Mrad and 50%

unirradiated PTFE. Their experiments were run using a three-pin-on-disk configuration

on a stainless steel counterface. They found the composite to wear considerably less and

have a lower friction coefficient than PTFE under the same sliding conditions. The large

plate-like wear debris commonly resulting from PTFE sliding is suspected to increase

friction by adding to surface ploughing and deformation. The composite sliding yielded

only thin oriented films. Blanchet and Peng suggest this is evidence that the wear debris

morphology is indicative of friction. Also, the irradiated/unirradiated composite was

non-abrasive to the steel counterface and possessed stronger creep resistance.

Sui [17] demonstrated that friction coefficient of a PTFE composite under sliding

has a non-linear dependence on speed. The composites friction was lower at slow speeds

than at high speeds. They also confirmed a strong dependence on contact stress. In order

to avoid severe wear they determined sliding must take place at speeds less than 20 m/s

along with contact stresses below 0.5 Mpa. Contact stress was shown to decrease while

sliding after an initial run in period. The stress is believed to drop with increasing contact

width.

PTFE wear rate data was collected from the aforementioned authors and is

displayed in Table 2A below. Wear rates along with the major test parameters known to

influence wear rate are displayed.









17




Table 2-1. Collection of PTFE wear rate data from previous authors.
Sliding in atmospheric conditions and room temperature
Paper Load (N) V (m/s) Distance (km) Pin Geometry Ra (um) Counterface K (mni/Nm) Avg. Publication


Pin on disk Circular wear path
*a Tanaka 147
et. al. 147
147
147
147

Pin on disk Circular wear path
Tanaka 98
et.al. 98
98
98
98

Pin on disk Circular wear path
Tanaka 98
et. al. 98
98
98

Circular wear path
Blanchet
et. al. 104

Pin on disk Circular wear path
104
Blanchet 104
et. al. 104
104
104
Pin on disk Linear wear path
Briscoe 79 6
et. al. 796
796
796
Pin on disk Circular wear path
b 170
Briscoe 170
et. al. 170
170
170
*Test run in vacuum


03
03
03
03
03


005
005
005
005
005


02
02
02
02



005


001
001
001
001
001


05
05
05
05


02


02
02
02
02
a Pre-rubbing


02
04
06
08
10


05
10
15
20
23


05
1 0
15
1 75



55


01
02
03
04
05


375
59
106
144


075
14
21
29
36


I of sample


3 mm diameter
3 mm diameter
3 mm diameter
3 mm diameter
3 mm diameter


3 mm diameter
3 mm diameter
3 mm diameter
3 mm diameter
3 mm diameter


3 mm diameter
3 mm diameter
3 mm diameter
3 mm diameter



4mmX4mm


4mm X 4mm
4mm X 4mm
4mm X 4mm
4mm X 4mm
4mm X 4mm


13 mm diameter
13 mm diameter
13 mm diameter
13 mm diameter


6 mm diameter
6 mm diameter
6 mm diameter
6 mm diameter
6 mm diameter


001


001
001
001
001
001


02
02
02
02


02
02
02
02
02


b Briscoe combined rotations


Glass
Glass
Glass
Glass
Glass


Glass
Glass
Glass
Glass
Glass


Glass
Glass
Glass
Glass



304 S Steel


304 S Steel
304 S Steel
304 S Steel
304 S Steel
304 S Steel


Mild Steel
Mild Steel
Mild Steel
Mild Steel


Mild Steel
Mild Steel
Mild Steel
Mild Steel
Mild Steel


6 6E-4
6 0E-4
5 6E-4
5 6E-4
5 7E-4


3 8E-4
2 2E-4
1 5E-4
1 2E-4
1 1E-4


4 9E-4
4 8E-4
4 9E-4
4 9E-4



1 1E-3


9 6E-4
8 7E-4
8 3E-4
8 9E-4
1 OE-3


7 6E-4
9 7E-4
7 5E-4
7 5E-4


1 2E-3
1 3E-3
1 3E-3
1 2E-3
1 3E-3


The Mechanism of
Wear of PTFE
Wear, 23 (1973)153-172
















Wear-Resistant PTFE
via Electron Irradiation
Lubrication Engineer 52,
6,489-495



Wear Resistant Irradiated
FEP I Unirradiated PTFE
Composites
Wear214(1998) 186-191


Transfer Wear of Polymers
During Combined Linear
Motion and Load Axis Spin
Wear, 104 (1985) 121-137


Briscoe, Evans, Pelillo, and Sinha [18] used a scratching technique to investigate



the energies related to surface deformation of polymers. Their research focused on the


characterizing wear and adhesion of polymers that produce thin transfer films including


PTFE. They indicate that the elastic properties of polymers such as PTFE impart a



restoring force to asperities that scratch the polymer surface, and that the nature of this


mechanism is a minimization of energy dissipated for material displacement. They also


indicate that PTFE experienced low levels of strain and responded elastically to a blunt


indenter.









2.3 Wear of UHMWPE

In regards to ultra-high molecular weight polyethylene (UHMWPE), Wang [19]

produced a theoretical wear model. Wang's theory originated from the observance of

fibrils composing most of the wear debris, and the importance of multi-directional sliding

on wear rate. Multi-directional sliding leads to shear and tensile stresses acting in

concordance. Any molecular orientation in the direction of one motion has weak

resistance to motion in another direction. Wang states that if enough energy is put into

the UHMPE surface a fibril may be pulled from neighboring fibrils. The work of friction

must act perpendicular to the fibril alignment in order for wear to take place. Wang also

states that there exists a critical value of coefficient of friction below which fibril removal

will not occur. There also exists a critical cross-link density above which fibril removal

will not occur.

While improving the wear resistance of UHMWPE, crosslinking has also resulted

in a decrease of material toughness. Muratoglu [20] investigated the crosslinking effects

on UHMWPE's wear behavior with a pin-on-disk tribometer. They demonstrated that

there was a linear relationship involving molecular weight between crosslinkings and

wear rate. Muratoglu's findings agree with previous findings that wear rate decreases

with increasing crosslinking density. They also state that crosslinking affects the

polymers ability to orient its molecular chains while sliding. Therefore, crosslinking

improves wear behavior under multi-directional sliding.

In their review of current concepts in wear of total hip and knee replacements,

Schmalzried and Callaghan [21] state that polyethylene wear is different from creep.

Creep deforms the polyethylene but does not produce any wear particles. Oxidation

reduces the ability of irradiated UHMWPE to form crosslinks between molecular chains.









Therefore, oxidized UHMWPE suffers from higher fatigue rates and delamination.

UHMWPE wear is also highly sensitive to scratches on the counterface surface. Wear

rate increases thirty to seventyfold when scratches of two micrometers in depth are

present [21].

Wear tests by Burroughs and Blanchet [22] in 2000 showed that shelf-aged

irradiated UHMWPE was less wear resistant than melt (2000C) vacuum post-irradiation

storage by approximately three-fold. The shelf-aged irradiated UHMWPE displayed

wear similar to unirradiated UHMWPE under multi-directional sliding against polished

surfaces. They also demonstrated that under multi-directional sliding UHMWPE does

not undergo the initial run-in wear period observed during unidirectional sliding. Given

the motion present in hip and knee joints, behavior during multi-directional sliding is of

more concern than unidirectional sliding. The tests run by Blanchet and Burroughs

consist of circular motion tests and rectangular motion tests. The wear rates for both

these forms of testing proved comparable. This suggests that multi-directional motion of

any kind is sufficient to produce higher wear than unidirectional motion.

Suh [23] presented the delamination theory of wear in 1973 for metals. The theory

is centralized around the assumption that subsurface material cold-works more than

material near the surface due to a greater dislocation density in the subsurface. Cold

working causes cracks to develop in the subsurface that eventually join together. When

the cracks reach a critical length the material between the crack and the surface will shear

and wear debris is thus produced. Suh also states that the same wear mechanisms are

responsible for adhesive and fatigue wear. Although this theory was developed for

metals, Briscoe [8] later indicated that delamination might play a role in polymer wear.






20


PTFE wear debris morphology appears in a plate-like form that is characteristic of

delamination. Delamination is driven by subsurface shear stresses that are not influenced

by the direction of sliding only that sliding exists.















CHAPTER 3
ENGINEERING APPROACH

3.1 Six-Station Pin-on-Flat Tribometer

The objective of this project was to elucidate some of the wear mechanisms

associated with load fluctuation and multi-directional motion of special polymers

(PTFE). Testing was performed using a pneumatically load controlled six-station pin-on-

flat tribometer developed at the University of Florida (UF) and presented in the thesis by

Aaron Ison. Several PTFE wear tests were performed using the six-station test rig prior

to commencement of this project to confirm the validity of this device as a tribometer.

The test rig was equipped with linear voltage differential transducers (LVDT) in order to

confirm translation and position of the movable stage. Load cells were used to verify that

the pneumatic pressure devices supplied the intended load to the samples. The results of

the preliminary testing indicate that the tribometer offers motion and load control

accurate enough for this project.

3.1.1 Table and Drive System

The six-station tribometer was designed to provide a means of creating multi-

directional motion while allowing in-situ variable load capabilities. The sample

counterfaces were translated using a multi-axis drive system. The system consists of a

stage set atop two linear tables stacked on top of one another and configured in a

perpendicular alignment. The top table is translated by a linear microstepper while the

stage is translated along the top table by another linear microstepper. This allows the









stage to travel anywhere in a horizontal x y plane. A controller manages the position,

motion, velocity, and acceleration of the tables. Programs describing the motion path are

sent from a computer to the controller's local memory. The controller then determines

the necessary signals needed to produce the desired motion and sends them to the

microsteppers.

3.1.2 Pneumatic Control

Six pneumatic cylinders are used to apply normal load to the polymer samples.

Pressure is supplied at 1.3 MPa by a 120 gallon compressor and is regulated by two 0 -

207 kPa pressure gauges. These pressure gauges can be seen in Figure 3.1 below.


Figure 3.1. Photograph of pressure gauges (top) backpressure to cylinders (bottom)
pressure to electro pneumatics.

The top mounted pressure gauge supplies backpressure to the cylinders as a means of

separating the polymer samples from the counterfaces. The bottom pressure gauge

supplies air to six electro-pneumatic pressure transducers. Each electro-pneumatic is









connected to one of the six pneumatic cylinders, and can supply pressure ranging from 0

- 138 kPa. A 0 5 Vdc signal sent to the electro-pneumatics controls the pressure

supplied to the cylinders. The electro pneumatic arrangement can be seen in Figure 3.2.



i"





















Figure 3.2. Arrangement of electro pneumatic gauges.

The output pressure is linearly proportional to the voltage supplied. Each electro-

pneumatic is controlled independently by separate signals sent from a computer that can

either be programmed to fluctuate or hold constant. For all the tests presented in this

paper, the pressure supplied to the electro-pneumatics was held constant at 207 kPa.

Voltage signals ranged from 0 5 V depending on desired test parameters, and were used

to produce output pressures ranging from 0 138 kPa sent to the cylinders. The signals

can be manipulated to model any desired loading pattern within the limits of the electro-

pneumatic pressure range and time response.









The pneumatic cylinders are all aligned perpendicular to the movable stage in two

rows of three. The arrangement can be seen in Figure 3.3.
























Figure 3.3. Photograph of pneumatic cylinder arrangement.

The cylinders possess a 2-inch inner diameter and have a threaded stem used to attach the

polymer sample holders. Each cylinder can apply a maximum load of 68 lbs (293 N)

directly to the polymer samples. An illustration showing the entire pin-on-flat tribometer

is displayed in Figure 3.4. Finally, the equipment log detailing the components needed to

operate the tribometer is displayed in Table 3-1.

3.1.3 Sample Holders

Six polymer sample holders were machined out of stainless steel to provide a

rigid brace for the polymer samples. The sample holders were designed specifically for

use with the pneumatic cylinders and can be viewed in Figure 3.5. A shop drawing of the

sample holder can be viewed in Appendix C.

















































Figure 3.4. Assembly drawing of entire pin-on-flat tribometer.











Table 3-1. Equipment register for pin-on-flat tribometer.
Equipment Manufacturer/Supplier Quantity
(Product Number)


Description


Linear Table


Amplifier


Pressure Gauge


Indexer/Drive Cable
Motors & Motor to


Load Cell


Connector Block


Conditioner


Power Supply


Daq Board Cable


Daq Board


Indexer


Pneumatic Cylinder



Connector Block



Daq Board


Electro Pneumatic
Transducer


Parker Automation
(406100XRMS)

Parker (Compumotor)
(OEMZL4)

Omega
(PRG501-30)

Parker (Compumotor)


Omegadyne
(LCKD-100)


National Instruments
(SCB-68)


Omega
(DP25-S)


Omegasnap
(DRN-PS-1000)

National Instruments
(SH68-68-EP)

National Instruments
(PCI-6034E)

Parker (Compumotor)
(6K8)


Bimba
(NR-311-D)


National Instruments
(CB-68LP)


National Instruments
(PCI-6713)


Omega
(IP413-020)


2 Square rail bearing with linear
screw

2 Amplifier for use with motors


2 Regulate supply pressure to
electro pneumatics

2 8 Amp drive microstepper used
to translate counterfaces

3 Signle-axis load cell for measuring
normal load


1 Allows computer to collect data
from tribometer

3 Condition voltage signal form
LCKD-100 to load

2 Convert wall voltage to 24 Vdc signal
@ 850 mA to power electropneumat

2 Communication between computer
and data aquistion boards

1 Analog Input board for data collection


1 8 axis indexer drive


6 Provide normal load ranging from
0 63 Ibs to the samples


1 Allows computer to send data
to tribometer


1 Analog output board for controlling
electro pneumatics

6 Provides active control of supply
pressure to cylinders



































Figure 3.5. Schematic of polymer sample holder.

The sample holders consist of a 12 inch inner diameter tapped to fit the cylinder stems

with a 14 inch thru hole cut through the centerline. The 12 inch diameter runs 3/ inch deep

and bottom shelf flattened by an end-mill. The 14 inch hole is cut through the shelf and is

14 inch deep. The large diameter end of the polymer sample fits snuggly inside the

sample holder with the polymer stem extending through the 14 inch hole. The 14 inch

hole allows the polymer to contact the counterface and lends side support to the polymer

stem during sliding. The sample holders are numbered along with the cylinders for

bookkeeping purposes.

3.2 Motion Paths and Loading Patterns

All sliding tests were performed in air under several motion paths and loading

patterns. The first motion path consisted of simple linear reciprocating through a stroke









length of 40.6 millimeters and an average sliding velocity of 46.5 mm/s. This motion

was used for several wear tests and in each case ran for 4 hours producing a sliding

distance of 670 meters. For the first series of tests all six cylinders were held at constant

pressure. The test pressures at the cylinders were 56, 84, and 112 kPa corresponding to

loads of 117, 176, and 235 N respectively. These loads equate to nominal pin pressures

of 3.69, 5.56, and 7.42 Mpa. The second series of tests consisted of the same motion

path, but applied loads oscillating at 6 second cycles. The range of load oscillation was

different for each test. The first test oscillated from 148 206 N the second 117 235 N

and the third 59 295 N. Each of these tests had an average load of 176 N. Finally a

third group of tests consisting of a random selection of load peaks and valleys cycling

once every 110 seconds was run with this motion path. The loading spectrum used in this

test was modeled after a rain-flow spectrum presented by J. A. Collins [26], but was

modified to achieve an average load of 176 N. A graph of the loading spectrum is shown

in Figure 3.6.

A second linear reciprocating motion path was used to provide the longest possible

stroke length that could be performed with the current counterfaces. The path ran

diagonal to the rectangular counterface with each stroke spanning a distance of 63.5

millimeters. Again, the test ran for 4 hours but was operating at a sliding velocity of 48

mm/s and a total sliding distance of 690 meters. The loading used for this test was held

constant at 176 N.

Three diamond patterns with varying degrees of crossing were also used as motion

paths. The first diamond path ran at 300 of crossing with lengths of 16.6 millimeters.











Spectrum Load Pattern
300


2 50 ---- .... -
200 ------------------------------------------
















0 20 40 60 80 100

Tim e (s)


Figure 3.6. Loading spectrum applied to samples for wear testing.

The second diamond path ran at 600 of crossing with lengths of 16.2 millimeters, and the

third diamond path ran at 900 of crossing with lengths of 15.8 millimeters. All of the

diamond pattern tests were run for 4 hours with all six cylinders at a constant load of 176

N. The path lengths were calculated to match the wear path areas for all diamond

patterns to the diagonal reciprocating test. The change in length of the diamond sides is

due to the increase in the polymer pin's wear path area contribution. Due to the pin's

circular geometry, an increase in degree of crossing causes an increase in wear path area

whenever a change in sliding direction takes place.

Finally, five circular motion paths were generated with different diameters. Each

test was run at a sliding velocity of 50 mm/s for 4 hours resulting in a sliding distance of

720 meters. For each of the five circular motion paths tested, all six cylinders were

functioning with constant loads being applied to the polymer samples. However, the load










was different from cylinder to cylinder. In groups of two, the cylinders applied loads of

117, 176, and 236 N respectively. The diameters for the five circle patterns were 6.35,

10.6, 15.0, 25.4, and 36.4 millimeters respectively. Figure 3.7 displays the motion paths

described in this paper.

Linear reciprocating Circular









Diamond



Three diamond paths used for comparison
inweartesting.



Figure 3.7. Motion paths used for wear testing.

The programs used to generate these motion paths are located in Appendix A.




3.3 Counterface Preparations and Handling

The counterfaces were cut from 440 stainless steel bar stock. Counterface material

was chosen to provide a significantly harder surface than that of the polymer sample.

Due to the high hardness of 440 stainless steel (Rockwell 54 C), wear of the counterface

could be neglected while sliding against the much softer polymer samples (PTFE 58 R,

UHMWPE 63 Shore D). The bar stock was cut into 6 rectangular plates 3 X 2.75 X

0.125 inches. The counterfaces were then prepared by a series of polishing steps as

described in Table 3B., beginning with 220 grit sandpaper and ending with a 0.3 |tm

particle slurry.









Table 3-2. Polishing steps for raw counterface samples.
Step Materials Method
1. 220 grit SiC sandpaper Polishing wheel
2. 440 grit SiC sandpaper Polishing wheel
3. 600 grit SiC sandpaper Polishing wheel
4. 15 m particle slurry with billard cloth Polishing wheel
5. 5 pm particle slurry with billard cloth Polishing wheel
6. 1 pm particle slurry with embry cloth Polishing wheel
7. 0.3 pm particle slurry with embry cloth Polishing wheel

After polishing, the counterfaces were washed using water and a mild detergent then

rinsed with methanol. Finally the surfaces of the counterfaces were characterized using

an optical profilometer. Based on data collected from the profilometer the polishing

technique described above yields an average surface roughness of approximately 0.02

jtm. Following each wear test, the counterfaces were washed with water and detergent

then polished, cleaned, and characterized using the steps described in Table 3C.

Table 3-3. Polishing steps for used counterface samples.
Step Materials Method
1. 5 gm particle slurry with billard cloth Polishing wheel
2. 1 m particle slurry with embry cloth Polishing wheel
3. 0.3 pm particle slurry with embry cloth Polishing wheel

Each of the six stainless steel counterfaces are constrained to the movable stage by four

4-40 facets, and are numbered to correspond with the cylinders and sample holders.



3.4 Polymer Sample Preparation

The PTFE samples were cut from molded V2 inch rod stock of virgin Teflon.

Receiving coordinate information from the G code presented in the appendix of this

paper, the samples were cut using a CNC mini milling machine. A schematic of the









polymer sample can be viewed below in Figure 3.8 or a shop drawing can be viewed in

Appendix C.

















Figure 3.8. Schematic of polymer sample.

The polymer sample has a large diameter base of /2 inches and is approximately /4 inches

in thickness. The stem has a smaller diameter of 14 inches and extends 2/5 inches from

the base. Allowing the stem to extend 2/10 of an inch from the sample holder. The

milling machine ensures that the base shelf and the stem's top surface are parallel to each

other. The sample's base shelf sits flat against the inner surface of the sample holder.

The sample's stem extends through the hole cut in the sample holder to contact the

counterface surface. The sample is locked into position when the sample holder is

threaded onto the cylinder. The UHMWPE samples were cut in the same manner and

with the same dimensions as the PTFE samples. The unirradiated UHMWPE was cut

from 1-inch stock samples 3/8 inch in diameter. The irradiated UHMWPE was cut from

the interior of a shelf aged puck shaped sample. Prior to each wear test, the samples were

placed in their respective sample holders and weighed individually. Once the wear test

was completed, loose debris was removed from the polymer stem. The samples would









remain inside the sample holders as they were weighed. Mass loss during the wear test

could then be calculated and used to determine wear rates.
















CHAPTER 4
EXPERIMENTAL RESULTS

4.1 Electro-Pneumatic Performance Data

Load cell data collected to evaluate the performance of the electro-pneumatics is

displayed in Figure 4.1.

Input Signal
---- Load Cell Output

Load vs Time


10 15 20 25 30 35 40 45
Time (s)
Figure 4.1. Electro-pneumatic performance data output from load cells.

The electro-pneumatics demonstrate the ability to follow a sinusoidal loading cycle that

ranges from 0 to 80 % of maximum pressure and cycles once every 6 seconds. The









conditioners used to translate data received from the load cells were not capable of

handling frequencies over 0.333 Hz. Therefore, there was inconclusive evidence

regarding the electro-pneumatics ability to follow a 3 second cycle that spanned 0 to 80

% of the max pressure range. To ensure reliable testing parameters, all dynamic loading

tests performed were constrained to frequencies equal to or below 0.167 Hz.




4.2 Variations in Wear Rate and Sliding Conditions

Along with the respective sliding conditions, the wear rates from every PTFE

wear test performed are shown in Table 4-1. The raw data collected for these

calculations can be viewed in Appendix B.

4.2.1 Transfer Film Formation

Although the tests involving PTFE were fairly consistent, data outliers occasionally

appeared in the results. Out of the 93 tests run, where mass loss was the measured

quantity, only 5 data points varied by more than 1 standard deviation from the mean for

that test. The outlier appeared as an unexpectedly high wear rate 2 times, and appeared

as an unexpectedly low wear rate 3 times. Unwaveringly, the appearance of outliers

corresponded with two very distinct transfer films. Under reciprocating motion, high and

normal wear always corresponded with the transfer film shown in Figure 4.2. The film

appeared patchy and uneven with portions of the counterface still exposed. The patches

appeared to be drawn out in the direction of sliding, but vary in width and length. Some

patches even appeared to be deposited on top of a previous patch. In contrast, low wear

always occurred in conjunction with the appearance of the smooth transfer film shown in

Figure 4.3.











Table 4-1. Raw data with calculated wear rates.

Wear Path Load Wear Rate (mm3/Nm)
Reciprocating Constant
Test Area (in2) Newtons 1 2 3 4 5 6
1 0.4491 117 3.47E-4 2.08E-4 3.64E-4 1.56E-4 2.49E-4 1.79E-4
2 0.4491 176 5.32E-4 5.05E-4 4.74E-4 4.85E-4 4.43E-4 3.93E-4
3 0.4491 235 5.43E-4 3.21E-4 5.43E-4 5.49E-4 5.61E-4 5.11E-4
Oscillating (Avg. 176 N)
4 0.4491 147-205 4.93E-4 4.32E-4 5.74E-4 5.24E-4 5.09E-4 4.74E-4


S0.4491 117-235 3.7
0.4491 59-293
Reciprocating Spectrum (Avg. 176 N)
Area (in2)
6 0.4491 6.2


8E-4 4.05E-4 4.35E-4


4.32E-4 4.74E-4 4.66E-4


4E-4 6.36E-4 6.05E-4 6.86E-4 6.70E-4 5.90E-4


Circular
Area (in2) Constant (N)
0.1963 117
7 0.1963 176
0.1963 235

0.3267 117
8 0.3267 176

0.3267 235

0.4625 117
9 0.4625 176
0.4625 235

0.7854 117
10 0.7854 176
0.7854 235

1.125 117
11 1.125 176
1.125 235


Semi-Circular
Area (in2)
0.4491
12 0.4491
0.4491


1 2
1.51E-4 1.02E-4


3 4 5 6


1.86E-4 2.33E-4


2.69E-4 2.55E-4


1.34E-4 1.51E-4


3.08E-4 2.08E-4


2.58E-4 3.60E-4


2.42E-4 2.26E-4


3.76E-4 2.94E-4


3.95E-4 3.60E-4


4.25E-4 3.87E-4


4.66E-4 4.59E-4


4.33E-4 5.89E-4


3.71E-4 3.01E-4


4.77E-4 5.31E-4


4.52E-4 5.59E-4


3.60E-4


4.10E-4


4.39E-4










Wear Path Load 176 Newtons Wear Rate (mm3/Nm)
Diamond Included
Test Area (in2) Angle 1 2 3 4 5 6
13 0.6741 0 5.39E-4 5.50E-4 6.02E-4 5.39E-4 5.65E-4 5.91E-4
14 0.6741 15 7.03E-4 6.73E-4 6.82E-4 7.07E-4 6.82E-4 6.44E-4
15 0.6741 30 6.18E-4 5.97E-4 5.97E-4 6.22E-4 6.09E-4 6.05E-4
16 0.6741 45 4.66E-4 4.44E-4 4.44E-4 4.78E-4 4.40E-4 4.32E-4


Although these transfer films displayed regions of light and dark patches that appeared to

run in the direction of sliding, there is no gross exposure of the underlying counterface.

Unlike the high and normal wear films, the low wear films appeared very smooth and

even. However, the regions of light and dark patches most likely indicate regions of

varying thickness within the transfer film.

The wear debris associated with both kinds of transfer films are plate-like in

geometry with most of the longer debris strips folded in an accordion fashion. However,

the wear debris appeared slightly smaller under low wear than it did under high wear.

Another glairing contrast between the two films is the extension of transfer film

over the wear path. For all reciprocating motion tests, the wear path consists of a long

portion of constant width capped by semi-circular portions at both ends. In cases of high

and normal wear, the transfer film extended only over the long portion of the wear path

and large amounts of wear debris were deposited at the end points of the wear path.

Whereas low wear transfer films cover the entire wear path.

Transfer films would also appear for circular sliding motion. However, the long

drawn out patches that appeared during high and normal wear under linear reciprocating

motion were now assembled into circular patches with diameters approximately equal to

the width of the wear path. The transfer films were all consistent in appearance and no
















(a) b)










(c) (d)
Figure 4.2. Optical micrograph of PTFE transfer film characteristic of high wear rates (a)
wear debris (b) end of wear path (c) top middle (d) bottom middle.


c) c)
Figure 4.3. Optical micrograph of PTFE transfer film characteristic of low wear rates (a)
end of wear path (b) top middle (c) bottom middle (d) middle.









major difference could be observed in the wear debris. Several images of circular motion

transfer films can be seen in Figure 4.4.





















c) (d)

Figure 4.4. Optical micrograph of PTFE transfer film deposited by 14.9 mm diameter
circular wear path (a) top (b) bottom (c) left (d) right.

Finally, images of the diamond motion transfer films are displayed in Figure 4.5.

These transfer films possessed characteristics resembling that of linear reciprocating

along the sides, meaning that the transfer film is consistent and appears drawn out in the

direction of sliding. At the corners, motion comes to a stop then restarts in a different

direction. This change in sliding direction is similar to what the polymer pin experienced

at the end points of the linear sliding tests. At all the covers the transfer film appeared to

breakup and deposit itself as small irregular shaped patches with sections of the

counterface exposed. Large amounts of the plate-like wear debris observed in previous

tests were deposited at the corners. Although the morphology of the transfer film









appeared different from the circular sliding tests, the wear debris showed no appreciable

difference from either the circular or linear sliding tests.











Figure 4.5. Optical micrograph of PTFE transfer film deposited by diamond pattern (a)
soft corner (b) sharp corner.

4.2.2 Wear Rate Comparisons

Comparisons between several wear tests are shown in Figures 4.6, 4.7, 4.8, 4.9, and

4.10 respectively. The volume of material lost during a wear test, v, was calculated by

dividing the sample's mass loss by its density. The wear rate, k, was then calculated by

dividing the volume loss by the total sliding distance and the average normal load applied

during the test as shown in equation 8.

k (8)
psd

Figure 4.6 shows the variation in wear rate as load increases under linear

reciprocating sliding motion. Each data point represents a sample slid for 670 meters

under constant load. The data shows an obvious trend that wear rate is proportional to

load, but variations in wear rate for each load do exist. At 117 N, the wear rate varied

from 1.56E-4 to 3.64E-4 mm3/Nm with a standard deviation of 35 percent. The average

of all six samples was 2.5E-4 mm3/Nm. With the exception of a single outlier at 235 N,

the wear behavior is more consistent at 176 and 235 N with standard deviations of 10 and

18 percent, and average wear rates of 4.72E-4 and 5.05E-4 mm3/Nm respectively.










0 6
X 5.1
E 5
z (0 4.7
-.
E4

E 3
0 Avg. = 2.5
2 2



0
0 50 100 150 200 250
Load (N)
Figure 4.6. Wear rate as a function of load for 670 meters of linear reciprocating sliding.

The proportionality of increase in wear to increase in load for identical motion paths is

similar to data presented by Tanaka [1].

For the same reciprocating wear path and sliding distance, data presented in Figure

4.7 indicates that oscillating load does not produce any change in wear behavior from that

of constant load. Regardless of the magnitude of oscillation, the wear rates remained

comparatively close to previously reported wear rates produced under a constant load

equal to that of the oscillating load's average value. The average wear rates were 5.01E-

4 for 58 N of oscillation, 4.06E-4 for 118 N of oscillation, and 4.57E-4 mm3/Nm for 234

N of oscillation. However, wear rates for the loading spectrum shown in Figure 3.6 of

the Engineering Approach section of this paper were significantly higher than those

produced with a constant load averaging 176 N. The average wear rate of all six samples

was 6.35E-4 mm3/Nm. The load fluctuations in Figure 3.6 act at nearly the same






42

frequency as those applied in the oscillating load test, however, the loading is more

heavily weighted at the beginning of the cycle.

SOscillating at 0.167 Hz
l Loading Spectrum

Wear Rate as a Function of Range

,T Q 6.4
S6
XF
^ 5 s.o
E 0 4.6
z
4 04.1
E o

.3


Range Avg. 176 N


0
0 50 100 150 200 250 300
Range of Load (N)
Figure 4.7. Effects of varying load on wear rate compared with effect of loading
spectrum on wear rate.

Following the progression of wear rate and load shown in Figure 4.6, the wear results for

spectrum loading more closely resemble those expected for a constant load of 293 N. A

load of 293 N was applied for brief moments during the spectrum test. For the samples

used in this test, 293 N of load borders on exceeding the yield strength of the material and

may have altered the material's wear behavior.

Figure 4.8 shows the effect of diameter on the wear rate for samples slid 720

meters in a circular motion path. As with reciprocating motion, the three constant load

values were applied. Each test indicated that wear again increased with increasing load.










6 Circular Wear Path Sliding


X

z 4

E








0
53 n
3 [6o .35 mm Diameter
Sa 1io.eB mm Diameter
So 15.0 mm Diameter
254 imm Diameter
+ 3814 mm Diameter




0 0.2 0.4 0.6 0.8 1 1.2
Area (in2)
Figure 4.8. Effects of load and diameter on wear rate for circular motion.

An overall view of the circular motion tests indicated that wear increased at a fairly linear

rate with increasing diameter. This behavior agrees with data presented by Briscoe [9]

and what was expected given the increase in wear path area. Although Tanaka [1]

provided data showing higher wear rates for smaller diameter circular wear paths, the

tests loading and sliding speed were not identical. The only exception to the increasing

wear with increasing diameter trend presented in this paper appears for the 36.4 mm

diameter test when the wear rate decreased for 117 and 235 N loads from the 25.4 mm

diameter test. All circular motion tests were preformed at the same sliding velocity.

Therefore, the number of cycles increased with decreasing diameter. For tests with the

same wear path area, wear rate proved to be higher under linear reciprocating motion

than it did for a circular motion path. As can be seen in Figure 4.9, the wear rate for









circular sliding motion does not become equal to that of linear sliding motion until the

circular wear path area is nearly 1.75 times greater than the linear wear path.


2 Ratio of Wear Rates


VI

=> 0) x


S .5












Figure 4.9. Ratio of circular motion wear rates over linear reciprocating motion wear
3+









0 0.5 1 1.5 2 2.5 3




rates.

The wear rate ratios were calculated by taking the average wear rate from a circular

motion path of given diameter and dividing that value by the average wear rate from

linear reciprocating motion of the same load. The area ratios were calculated by

determining the wear path area of a given diameter test and dividing that value by the

wear path area created from a single pass under linear reciprocating motion. Although

this data is consistent with findings by Briscoe [9], increasing wear with increasing

diameter is in contrast to what was expected given the current theories regarding

polymers such as UHMWPE and multi-directional sliding. However, this trend may be

explained by the increasing number of cycles incurred by the polymer pin with

decreasing wear path diameter and assuming a directional independent wear mechanism.









Figure 4.10 shows the change in wear rate as a function of the angle of inclusion

for several diamond shaped wear paths. All of the diamond pattern tests were run with a

constant load of 176 N for a total of 4 hours. Due to changes in the motion path from test

to test, the total sliding distance for 00 angle of inclusion differed from the other tests.

The 0 angle of inclusion test slid for 695 meters while all other tests slid for 621 meters.

The distances correspond to sliding speeds of 48.3 and 43.1 mm/s. Despite the small

difference in speed, it is important to note that the higher sliding speed may result in an

increase in wear relative to lower sliding speed tests. However, it is safe to assume that

whatever change in wear rate resulted from different sliding speeds, it is not significant

enough to produce a change in any trend observable in Figure 4.10. For 0 angle of

inclusion the six samples had an average wear rate of 5.64E-4 mm3/Nm. The average

wear rate at 15, 30, and 450 of inclusion were 6.82E-4, 6.08E-4, and 4.51E-4 mm3/Nm

respectively. Although there was excellent consistency in the data from each set of tests,

standard deviations of 4.78, 3.29, 1.74, and 3.88 % respectively, there was no consistency

in the wear behavior from test to test. Wear rate increased from 0 to 15 of inclusion

then decreased slightly at 300 of inclusion and finally decreased sharply at 450 of

inclusion. With the exception of the 450 angle of inclusion, the wear rates collected

under diamond shaped sliding motion were higher than any other wear test run with the

same load. The wear path area was 4.35 cm2 for all four tests. This wear path was 1.5

times greater than the wear path created under the previous linear reciprocating motion

tests. However, wear rate was only shown to increase by this much when 150 of

inclusion was incurred, and at 00 of inclusion wear rate increased by only 1.2 times that

of the previous linear reciprocating tests.






46


0 0 Degree Inclusion
a 15 Degreetnclusion
S30 Degree Inclusion
45 Degree Inclusion

8 Daimond Wear Path Sliding

6, 7
o 7 6.8
X 6/ 6.1
S5.6
z 5
4.5
E
E 4







0
0 10 20 30 40 50
Angle of Inclusion (Degrees)
Figure 4.10. Wear rates for diamond pattern sliding motion as a function of inclusion.

4.3 Cycle dependence on Wear

Wear rate shows an inverse dependence when plotted against the number of cycles

incurred during sliding as shown in Figure 4.11 below. Regarding circular path wear

tests, the least number of cycles run was approximately 8000 and corresponded to the

highest wear rate for any such motion. As the number of cycles increased, the wear rate

asymptotically approached some value around 0.18 x 10 -3 mm3/Nm. A similar trend

appeared with the linear reciprocating and diamond path tests. Although the wear rates

for this group were generally shifted up from the circular path wear curve, the final data

point at 52000 cycles fell below what was expected from the circular path wear curve.






47


experimental data from tests with constant
load and equal sliding distance
0.8

'- 0.7- 15

i 0.6- 300

E 0.65 3 00 reciprocating


E 450
T 0.3 bO full-circle

S0.2- O O curve-fit

0.1 -

I I I III
0 10,000 20,000 30,000 40,000 50,000 60,000
n
Figure 4.11. Wear rate as a function of number of cycles.

4.4 Images of Wear

An image of the polymer samples before and after wear testing can be seen in

Figure 4.12.


Figure 4.12. Photograph of polymer samples before 4 hour linear reciprocating wear test
at 176 N (left), and after (right).









The amount of PTFE wear resulting from different loads after being ran for 720 meters at

50 mm/s in a circular sliding motion of diameter 36.4 mm can be seen in Figures 4.13

and 4.13. Figure 4.13 shows the sample holders marked 0 and 3 run at 117 N, holders

marked 2 and 5 run at 176 N, and holders 1 and 4 run at 235 N.


Figure 4.13. PTFE wear post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235
N, right 176 N.

Figure 4.14 shows the PTFE wear paths resulting from the same test.


Figure 4.14. Wear paths post 720 meters slid testing at 50 mm/s. Left 117 N, middle 235
N, right 176 N.















CHAPTER 5
SURFACE CHARACTERIZATION AND SUBSURFACE STRESS MODELING



Based on data collected from optical stylus scans of the steel counterface prior to

wear testing, the surface was assumed to be sinusoidal in nature with a 40 nm amplitude

and a 30 |tm period. The optical scan information supporting these numbers is shown in

the appendix of this report. The scan filtered out data with frequencies higher than 100

cycles per millimeter. This frequency was iteratively chosen because it gave a roughness

value approximately equal to the average roughness of the entire counterface while

eliminating any sharp asperities that would not actually support a significant amount of

the normal load. The asperity density was modeled by assuming the pit-centered

configuration shown in Figure 5.1.

The area per asperity peak was calculated using equations 9 and 10.

A P(9)


S= 2 (10)

Where x is the counterface area per asperity peak and is the length of one side of the

square shown in Figure 5.1 above. The number of asperity peaks in contact with the

polymer pin at any time was calculated using equation 11.

A
0= (11)


































Figure 5.1. Asperity peak configuration at the surface of the steel counterface.

Where Q is the number of asperity peaks in contact with the polymer pin and Ap is the

area of the pin face. Once the number of asperity peaks in contact with the pin is known

the load per peak, fasp, can be calculated by dividing the normal load by the number of

peaks Q. The peak radius, Rcp, shown in Figure 5.2 can be calculated before

determining the Hertz elastic contact patch by using equation 12.


R, = 2 -L +h (12)


Assuming that the polymer surface is identical to the counterface surface, the composite

radius becomes


R =RP (13)
2

R can now be used with the Hertz contact patch equations to determine the size of the









Asperity Peak


h



( C -









Figure 5.2. Radius of asperity peak at surface of steel counterface.

contact patch between the polymer and counterface.


a 3x f,, xR 3 (14)
a 4E (14)
4xE )

Where a represents the radius of a circular contact patch and E' is the composite modulus

of elasticity for steel and PTFE. R was shown to be orders of magnitude larger than the

contact patch therefore confirming the Hertzian contact assumptions. The maximum

pressure is then calculated using equation 15.

C 3 x f p
Pmax = 2a2 (15)


The pressure profile along the surface of the polymer pin is shown in Figure 5.3. Given

the low friction value between PTFE and steel, distortion of the pressure profile can be

neglected once sliding has occurred [27]. The equation describing the pressure profile

shown in Figure 5.3 is substituted in place of P(s) in the stress equations for an elastic

half-space 16,17,18.









s = a Polymer Pin


Pma 1 -

,S





Direction of Sliding Steel Counterface Incomplete Contact at
Pin/Counterface Junction
Figure 5.3. Pressure profile applied to PTFE surface when in contact with counterface.



S{( X-S)2 +Z2} a {(XS)2 +Z2}

2z3 j P(s)x- s 2z 2 j/P(s)(x- s)
s= -z 2ds--- ds (17)



2 2 a d Pd s)( x _1
j =s)ds 2 P( ds (18)
-a (x- 2+z a {(x- S) +z2}

Where ox and yz describe the stress in the x and z directions while cxz describes the shear

stress in the xz plane. When the two materials in contact have modulus of elasticity an

order of magnitude different, the presence of traction may cause a distortion of the

pressure profile. However, the pressure profile here is assumed to be unaffected since the

friction coefficient, [t, of PTFE sliding against steel is not significant enough to distort

the pressure profile from that shown in Figure 5.3.

To analyze the subsurface affects of the asperities, a section of polymer 150 |tm

long by 200 |tm deep was chosen to represent the entire polymer pin. Given the spacing

of asperities along the surface, six asperities are in contact with this section of the pin.









The section of polymer was discretized into regions 6 |tm long by 10 |tm deep. Using the

stress equations above, the stress resulting from one of the six asperities was calculated

for each of these regions within the main section of polymer. This process was repeated

for all six asperities at their respective locations along the surface. The location of the

regions remained the same while the distance to the corresponding surface asperity

changed. Each asperity's increase in stress on a specific region was summed together to

attain the combined stress of all six asperities on that region. Once the total stress of

every region was calculated the results were plotted as shown in Figure 5.4, 5.5, and 5.6.

Once the original stress and shear state was calculated, the Mohr's circle technique was

implemented to find the maximum shear stress within the subsurface. The maximum

shear stress at a given depth is given by


A =v (19)


R = ( A2 (20)


R, = (r)2 (21)


=x (R +R) (22)

Where Cmax is the shear stress of the maximum shear stress element. Once Cmax is

calculated for every location the results were plotted against the Z-axis to reveal the depth

at which maximum shear stress occurs. These results can be seen in Figure 5.7.































,Ho














-50 Sigma X compressive stress in subsurface of PTFE.

H orizontal x-axis (Rlm)
Figure 5.4. Sigma X compressive stress in subsurface of PTFE.






























311-


N













Ii




-350 0 3E

Horizontal x-axis (pom)
Figure 5.5. Sigma Z compressive stress in subsurface of PTFE.







56














S201







10
Cl








-50 0 50

Horizontal x-axis (.m)
Figure 5.6. Tau XZ shear stress in subsurface ofPTFE.









2.5 106


2 106


1.5 106

taumax
1 106
(Pa) 1
Max tau @ 8.328 pm along X =0
5 1approx. 0.78 x a
5 10


01 --- L -- L -- L ------ 1 _
0 0.5 10 1 10 1.5 10 2 10 2.5 10
Z
Subsurface Depth (m)
Figure 5.7. Plot of subsurface shear stress along x = 0 indicating shear max.















CHAPTER 6
DISCUSSION

6.1 Delamination

Unlike UHMWPE, PTFE appears to wear at approximately the same rate

regardless of any multi-directionality in the sliding motion. Results presented in tests 1-3

and 9 of Table 4A and Figure 4.8 indicate that there is considerable data overlap between

wear rates resulting from unidirectional sliding and those resulting from multidirectional

sliding. Images from the optical microscopes show that as PTFE wears it tends to form

debris as thin plate shaped flakes of material. Wear of UHMWPE is often described as a

surface wear process that is dependent on the orientation of the molecular chains relative

to the direction of sliding. UHMWPE exhibits low wear under unidirectional sliding

because the molecular chains eventually align themselves with the direction of motion,

making the chains more difficult to remove from the bulk. The chains are more easily

removed when experiencing shear, as is the case under multidirectional sliding. PTFE

appears to wear as a result of a subsurface process known as delamination.

Delamination occurs when a subsurface crack propagates long enough to linkup

with other subsurface cracks until eventually one crack large enough to break from the

bulk is present. The delamination process is depicted in Figures 6.1, 6.2, and 6.3 below.

Once sliding begins stresses stemming from the combination of normal load and traction

develop within the polymer pin. The stress equations used to model the subsurface

stresses experienced by the polymer pin were presented in chapter 5 and have no







59


Normal Load





Polymer Pin





Subsurface Crac





Direction of Motion
SSteel Counterface

Figure 6.1. Presences of subsurface cracks within polymer pin under stress.


Normal Load


4r


Polymer Pin




Subsurface Cracks Propagating






Direction of Motion
Steel Counterface

Figure 6.2. Subsurface cracks begin to propagate and link up.










Normal Load




Polymer Pin



Subsurface Cracks Propagating
Delamination Delamination Wear Flakes



Direction of Motion
0 Steel Counterface

Figure 6.3. Ejections of polymer wear debris resulting from large subsurface crack.

dependence on direction of motion. Therefore, the delamination process has no

dependence on direction of sliding. This accounts for the lack of any directional

dependence in PTFE sliding wear. Delamination may not influence UHMWPE's wear

process because of the extremely low friction values incurred under lubricated sliding.

The pressure profile applied to the polymer surface is the main factor influencing

the subsurface stress. The pressure profile depends on both loading and surface

topography. For the experiments pertaining to this paper, the only external load applied

to the polymer was the normal load. Therefore, modeling the surface topography became

the key factor when describing the pressure profile. Depending on the state of stress

present in the polymer bulk compression, tension, and shear may be exuded on the

imperfections present near the polymer surface. The imperfections include cracks that

begin to propagate under either tension or shear. The state of stress depends on the

loading and surface features at the polymer/counterface junction. Based on the surface

model describing the counterfaces used during wear testing, the stress equations indicate









that both compression and shear are present within the subsurface. Crack propagation

may take place by either one or both of the two modes shown in Figure 6.4a, and b.
Tension

tttt hear
Y I I



X T a) x b)
Figure 6.4 Modes of crack propagation a) mode I b) mode II.

Tension, associated with mode I failure, will force the crack to pull apart and thus

propagate. Shear will force the top and bottom halves of the crack to move in opposite

directions. Mode II is characterized by motion perpendicular to the leading edge of the

crack, whereas Mode III the motion is parallel. For the case describing both the

counterface and polymer surface topography as sinusoidal with 40 nm amplitude and 30

|tm spacing, the subsurface stress model indicates the presence of shear and compression.

Given a unidirectional sliding path, mode II propagation best agrees with the subsurface

stress model, and indicates that the maximum shear takes place at about 10 |tm deep.

The correlation between experimental data and theoretical modeling lend support to the

theory of delamination as an explanation for the directionally independent wear behavior

of PTFE sliding against a polished steel counterface. The subsurface shear present within

the subsurface of the polymer results in mode II crack propagation. As the propagation

begins to link up several cracks with one another it becomes easier to break a plate of

material free from the bulk than to break the polymer/counterface junction. Once a

section breaks from the bulk it is either ejected from the wear path as debris or deposited

on to the counterface as a transfer film. Although delamination theory accounts for much

of what was observed during PTFE wear, it is still unclear as to what mechanism causes









the material to become part of a transfer film instead of ejected as debris. However, it is

clear that a full description of PTFE sliding wear behavior must include an explanation

for any third body interactions resulting from the transfer film.

6.2 Reversal Zones

An inconsistency between wear tests that employ unidirectional sliding and those

that employ multidirectional sliding involves reversal points in the motion path and the

subsequent appearance of static friction. A reversal point is any location along the

motion path where the polymer pin comes to a stop then restarts its motion in another

direction. Such as the endpoints of a linear reciprocating wear path. These locations are

of particular interest because they introduce static friction to the sliding process. Motion

paths such as circles contain no reversal points since it is not necessary to stop the motion

in order to change directions. As can be seen by comparing tests 1-3, 9, and 12 of Table

4A, under identical test conditions, tests that include reversal points have higher wear

than tests with no reversal points. Tests 14-16 of Table 4A each include 4 reversal points

and possess even higher wear rates than the tests with only 2 reversal points 1-3, and 12.

However, the wear path area for tests 14-16 is greater than the area for tests 1-3, and 12.

This is significant because an overall view of the data in Table 4A indicates that wear rate

increases with increasing wear path area. A comparison could be made between test 13

and tests 14-16. Test 13 has the same wear path area as 14-16, but has only 2 reversal

locations. On the average, the wear rate of 14-16 was higher than 13, but some data

overlap did occur.

Reversal points are identifiable on the transfer films as interruption areas of

broken, patchy film. As stated in the results section of this report, the appearance of

broken patchy transfer films was coincident with relatively high wear rates. For smooth






63

transfer films with reversal points, the same could be said about wear rates over those

areas of film interruption. Hence, wear tests have higher wear rates under smooth

transfer film formation when reversal points are present in the sliding motion.

6.3 Cycle Dependence in Wear Rate

Figure 4.10 of the results section shows that wear rate has a strong dependence on

the number of cycles incurred during sliding. A simple model for this dependence is

shown in Figure 6.5.

Transition of Wear Rates



-A /
0 n c K
~ n c ^- K2
E /


> //K1



number of cycles
Figure 6.5. Model showing wear rate transition at some critical number of cycles nc.

This model depicts a transition from the initial high wear rate K1 at some critical number

of cycles no to a lower steady state wear rate K2. Therefore, any test that runs beyond no

will begin to appear more and more like the steady state wear rate. Using linear rules of

mixing, a prediction of wear in terms of volume loss for single point measurements can

be made using equation 23.


V,. = ncKF,d + K2,d(n n)


(23)






64


Where d is the total sliding distance divided by the total number of cycles n. An

expression based on wear rate of a single point measurement is given by equation 24.


K = = K2 + nc(K -K2) (24)
SFD n

Where Ksp is the average wear rate over the entire test. Wear at any point during the test

can be predicted if the number of cycles at that point is known.
















CHAPTER 7
CONCLUSIONS



Changes in wear rate have been observed when changes in test parameters have

been implemented. Factors influencing wear rate are load, wear path area, and number of

cycles incurred during testing. The influence of number of cycles incurred indicates that

wear is driven by the transfer film setup on the counterfaces.

The calculated wear rate values from the pin-on-flat tribometer agreed closely

with wear rate values calculated by previous authors. Preliminary testing of this

tribometer indicated that the results are repeatable and that changes in wear rate can be

attributed to changes in the testing parameters. Therefore, it is believed the tribometer

used in this report is functioning properly and can be used to identify factors influencing

wear.

1. The initial experiments with cyclic loading suggest that slowly varying cyclic loads
have similar wear rates as produced by a constant load equal to the cyclic mean
load. This has only been tested for a small range of load that has peak amplitudes
that are within the same order of magnitude as the mean load.
2. The qualitative competitive rate models previously proposed for PTFE appear
appropriate for explaining the dependence on the number of cycles on wear, but not
for the development of the transfer film.
3. The number of reversals in a given wear path is related to wear rate.
4. The directionality of sliding shows significant differences in transfer film
morphology within the reversal zones, but does not show significant differences in
overall wear rate as compared to linear reciprocating sliding.

There is no satisfactory explanation for how the transfer film develops. Questions

concerning transfer film development as well as wear debris growth and expulsion still






66


remain. A system designed specifically to observe the transfer films development under

similar test conditions to those described in this paper is needed to help reveal useful

information regarding wear behavior and would serve well as a point of future research.

Such a system could be implemented on an existing tribometer with the addition of video

imaging equipment and a proper scope setup.




















APPENDIX A
MOTION PATH PROGRAMS


constant ;Program for linear reciprocating motion path


del constant
def constant

drivel0000
ma00000
a200,200
ad200,200
v9.8425

1
d-200000
gol
d200000
gol
In
end


circle


;Clear controller memory of previous program
;Define new program

;Activate tables
;Set table to absolute coordinates
;Set accelerations and decelerations

;Set velocities

;Initiate loop
;Command table number of units to move
;Initiate table motion


;End loop
;End program


;Program for circular motion path


del circle
def circle

Drive11000
pv9.8425,9.8425

pa200
pad200

11000
parcop0,0,0,62500

In



end


;Clear controller memory of previous program
;Define new program

;Turn tables on
;Set path velocities and accelerations and
decelerations



;Initiate loop
;Define circle end points (x,y) and center
points (x,y)
;End loop
;Before this program will run type pcomp circle
into the terminal and enter
;Then type prun circle and enter
;End program

















del cir
def cir


pcomp circle
1200
prun circle
In


end


;Program for loading and running circular motion
path

;Clear controller memory of previous program
;Define new program

;Compile circle program into controller memory
;Initiate loop for 200 cycles
;Run circle program
;End loop


;End program


arc2


del arc2
def arc2
drive11000


1
pv9.5
pal000
padl000
prtol5
parcp-127328,0,63664

In



end


;Program for semi-circular motion path

;Clear controller memory of previous program
;Define new program
;Turn tables on

;Initiate loop
;Set path velocities and accelerations
;Set path accelerations and decelerations


;Define arc startpoints (x,y) and end points
(x,y)
;End loop
;Before this program will run type pcomp arc2
into the terminal and enter
;Then type prun arc2 and enter
;End program


constantly ;Program for 0 diamond motion path


del constantly
def constantly


;Clear controller memory of previous program
;Define new program


drive11000
ma00000
a200,200
ad200,200
v7.3159,6.5843,1,1,1

1
d-232279,209051
goll
d232279,-209051
goll
In


;Activate tables
;Set tables in absolute/
;Set acceleration to 200
;Set deceleration to 200


;Initiate loop
;Command table number of
;Initiate table motion


;ends loop


incremental mode
revs/sec^2
revs/sec^2



units to move


;end main


end
















diamond


del diamond
def diamond

drive11000
ma00000
a200,200
ad200,200

1
v5.3625,8.2534
d-44793,68942
goll
v8.7707,4.4665
d-73263,37309
goll
v5.3625,8.2534
d44793,-68942
goll
v8.7707,4.4665
d73263,-37309
goll


end


;Program for 15 diamond motion path

;Clear controller memory of previous program
;Define new program

;Activate tables
;Set tables in absolute/ incremental mode
;Set acceleration to 200 revs/sec^2
;Set deceleration to 200 revs/sec^2

;Initiate loop
;Set velocity to 9.8425 revs/sec
;Command table number of units to move
;Initiate table motion










;ends loop


;end main


diamond ;Program for 30 diamond motion path


del diamond
def diamond

drive11000
ma00000
a200,200
ad200,200

1
v3.0436,9.3601
d-24791,76241
goll
v9.6279,2.0442
d-78422,16651
goll
v3.0436,9.3601
d24791,-76241
goll
v9.6279,2.0442
d78422,-16651
goll


;Clear controller memory of previous program
;Define new program

;Activate tables
;Set tables in absolute/ incremental mode
;Set acceleration to 200 revs/sec^2
;Set deceleration to 200 revs/sec^2

;Initiate loop
;Set velocity to 9.8425 revs/sec
;Command table number of units to move
;Initiate table motion










;ends loop


;end main


end
















diamond


del diamond
def diamond

drive11000
ma00000
a200,200
ad200,200


v.51732,9.8289
d4106,78017
goll
v9.8289,.51732
d78017,4106
goll
v.51732,9.8289
d-4106,-78017
goll
v9.8289,.51732
d-78017,-4106
goll
n ;ends loop


;Program for 45 diamond motion path



;Clear controller memory of previous program
;Define new program

;Activate tables
;Set tables in absolute/ incremental mode
;Set acceleration to 200 revs/sec^2
;Set deceleration to 200 revs/sec^2

;Initiate loop
;Set velocity to 9.8425 revs/sec
;Command table number of units to move
;Initiate table motion


end ;end main


















APPENDIX B
RAW DATA


Cylinder Mass Loss
(g)
1 60
2 36
3 63
4 27
5 43
6 31


Normal Load
(N)
117
117
117
117
117
117


Sliding Distance
(m)


Speed Path Length


(mm/s)

47
47
47
47
47
47


(mm)

45.7
45.7
45.7
45.7
45.7
45.7


Linear 1 138 176 670 47 45.7
Recip 2 131 176 670 47 45
2 131 176 670 47 45.7
3 123 176 670 47 45.7
4 126 176 670 47 45.7
5 115 176 670 47 45.7
6 102 176 670 47 45.7

Linear 1 188 235 670 47 45.7
Recip 2 111 235 670 47 45.7

3 188 235 670 47 45.7
4 190 235 670 47 45.7
5 194 235 670 47 45.7
6 177 235 670 47 45.7

Linear 1 25 176 670 47 12.7
Recip 4 30 176 670 47 12.7


Linear 1 128 147-205 670 47 45.7
Recipw/0.167Hz 2 112 147-205 670 47 45.7
Sinusoidal Load
3 149 147-205 670 47 45.7
4 136 147-205 670 47 45.7
5 132 147-205 670 47 45.7
6 123 147-205 670 47 45.7


Linear
Recip w/ 0.167 Hz
Sinusoidal Load


117-235
117-235
117-235

59 293
59 293
59 293


45.7
45.7
45.7

45.7
45.7
45.7


Test

Linear
Recip














Circular 1 28 117 700 49 19.9
6.35 mm
35m 2 19 117 700 49 19.9
diameter
3 52 176 700 49 19.9
4 65 176 700 49 19.9
5 100 235 700 49 19.9
6 95 235 700 49 19.9

Circular 1 25 117 700 49 33.2
10.6 mm 2 28 117 700 49 33.2
diameter 3 86 176 700 49 33.2

4 58 176 700 49 33.2
5 96 235 700 49 33.2
6 134 235 700 49 33.2

Circular 1 45 117 700 49 47.0
15.0 mm 2 42 117 700 49 47.0
diameter 3 105 176 700 49 47.0

4 82 176 700 49 47.0
5 147 235 700 49 47.0
6 134 235 700 49 47.0


Circular 1 79 117 700 49 79.8
25.4 mm 2 72 117 700 49 79.8
diameter 3 130 176 700 49 79.8
4 128 176 700 49 79.8
5 161 235 700 49 79.8
6 219 235 700 49 79.8

Circular 1 69 117 700 49 114.3
36.4 mm 2 56 117 700 49 114.3
diameter 3 133 176 700 49 114.3

4 148 176 700 49 114.3
5 168 235 700 49 114.3
6 208 235 700 49 114.3












Spectrum 1 162 176 (avg) 670 47 45.7
Load
2 165 176 (avg) 670 47 45.7
3 157 176 (avg) 670 47 45.7
4 178 176 (avg) 670 47 45.7
5 174 176 (avg) 670 47 45.7
6 153 176 (avg) 670 47 45.7

Linear 1 145 176 690 48 68.5
Recip 2 148 176 690 48 68.5
3 162 176 690 48 68.5
4 145 176 690 48 68.5
5 152 176 690 48 68.5
6 159 176 690 48 68.5

Diamond 1 169 176 620 43 68.5
w/15 Inclusion 2 162 176 620 43 68.5
Angle
3 164 176 620 43 68.5
4 170 176 620 43 68.5
5 164 176 620 43 68.5
6 155 176 620 43 68.5

Diamond 1 146 176 620 43 68.5
w/ 300 Inclusion 2 141 176 620 43 68.5
Angle 3 141 176 620 43 68.5

4 147 176 620 43 68.5
5 144 176 620 43 68.5
6 143 176 620 43 68.5

Diamond 1 110 176 620 43 68.5
w/450 Inclusion 2 105 176 620 43 68.5
Angle
3 105 176 620 43 68.5

4 113 176 620 43 68.5
5 104 176 620 43 68.5
6 102 176 620 43 68.5

Semi-Circular 1 63 117 670 47 45.7
























APPENDIX C

SHOP DRAWINGS


Shop drawing of aluminum base mounted to tribometer stage.
'7




E 0



Ic

I 1| .





E
o |V




o __ o -



-----------







co V) L--- 10 ----




















U,
-o 6 '

















Shop drawing of counterface mount.




^--i-----------

















I)l
10
00





o






76




Shop drawing of polymer pin sample.













Shop drawing of polymer sample holder.


(9wo)


LO
L


'(o.o 0













APPENDIX D
SURFACE METROLOGY

Surface profiles of steel counterfaces.


X Profile


m 0.00 Y Profile
0.04
0.04 -


-0.06


220 lm


um


300 pm


0.00
















LIST OF REFERENCES


1. Tanaka K., Uchiyama Y., and Toyooka S., The Mechanism of Wear of
Polytetrafluoroethylene, Wear 23, pp 153-172 (1973)

2. Tanaka K.,and Miyata T., Studies on the Friction and Transfer of Semi-Crystalline
Polymers, Wear 41, pp 383-398 (1977)

3. Schott M., Preparation and Properties of Highly Oriented Polytetrafluoroethylene
Films, Synthetic Metals 67, pp 55-61 (1994)

4. Bodo P., and Schott M., Highly Oriented Polytetrafluoroethylene Films: A Force
Microscopy Study, Thin Solid Films 286, pp 98-106 (1996)

5. Viswanath N., and Bellow D. G., Development of an Equation for the Wear of
Polymers, Wear 181-183, pp 42-49 (1995)

6. Yang E-L., Hirvonen J. P., and Toivanen R. O., Effect of Temperature on the
Transfer Film formation in sliding Contact of PTFE With Stainless Steel, Wear
146, pp 367-376 (1991)

7. Blanchet T. A., and Peng Y-L., Wear-Resistant Polytetrafluoroethylene via
Electron Irradiation, Lubr. Eng. 52, pp 489-495 (1995)

8. Briscoe B., Wear of Polymers: An Essay on Fundamental Aspects, Tribol. Int. 14,
pp 231-243 (1981)

9. Briscoe B. J., and Stolarski T. A., Transfer Wear of Polymers During Combined
Linear Motion and Load Axis Spin, Wear 104, pp 121-137 (1985)

10. Vijayan K., and Biswas S. K., Wear of Polytetrafluoroethylene: Some
Crystallographic Observations, Wear 150, pp 267-273 (1991)

11. Brainard W. A., and Buckley D. H., Adhesion and Friction of PTFE in Contact
With Metals as Studied by Auger Spectroscopy, Field Ion and Scanning Electron
Microscopy, Wear 26, pp 75-93 (1973)

12. Cheng X., Xue Y., and Xie C., Tribological Investigation of PTFE Composite
Filled With Lead and Rare Earths-Modified Glass Fiber, Materials Letters 4198, pp
1-5(2002)









13. Wieleba W., The Statistical Correlation of the Coefficient of Friction and Wear
Rate of PTFE Composites With Steel Counterface Roughness and Hardness, Wear
252, pp 719-729 (2002)

14. Jintang G., Tribochemical Effects in Formation of Polymer Transfer Film, Wear
245, pp 100-106 (2000)

15. Pleskachevsky Y. M., and Smurugov V. A., Thermal Fluctuations at PTFE Friction
and Transfer, Wear 209, pp 123-127 (1997)

16. Blanchet T. A., and Peng Y-L., Wear Resistant Irradiated FEP/Unirradiated PTFE
composites, Wear 214, pp 186-191 (1998)

17. Sui H., Pohl H., Schomburg U., Upper G., and Heine S., Wear and Friction of
PTFE Seals, Wear 224, pp 175-182 (1999)

18. Briscoe B. J., Evans P. D., Pelillo E., and Sinha S. K., Scratching Maps for
Polymers, Wear 200, pp 137-147 (1996)

19. Wang A., A Unified Theory of Wear for Ultra-High Molecular Weight
Polyethylene in Multi-Directional Sliding, Wear 248 (2001), pp 38-47 (2001)

20. Muratoglu O. K., Bragdon C. R., O'Connor D. O., Jasty M., H. Harris W. H., Gul
R., and McGarry F., Unified Wear Model for Highly Crosslinked Ultra-High
Molecular Weight Polyethylenes, Biomaterials 20, pp 1463-1470 (1999)

21. Schmalzried T. P., and Callaghan J. J., Current Concepts Review Wear in Total Hip
and Knee Replacements, The Journal of Bone and Joint Surgery, Incorporated 81-
A, pp 115-136 (1999)

22. Burroughs B. R.,and Blanchet T. A., Factors Affecting the Wear of Irradiated
UHMWPE, Society of Tribology and Lubrication Engineers Presentation NY
(2000)

23. Suh N. P., The Delamination Theory of Wear, Wear 25, pp 111-124 (1973)

24. Steijn R. P., The Sliding Surface of Polytetrafluoroethlene: An Investigation with
the Electron Microscope, Wear 12, pp 193-212 (1968).

25. Gong D., Xue Q., and Wang H., Physical Models of Adhesive Wear of
Polytetrafluoroethylene and its Composites, Wear 147, pp 9-24 (1991)

26. Collins J. A., Failure of Materials in Mechanical Design, 2nd Edition, A Wiley-
Interscience Publication NY, NY (1993)

27. Johnson K. L., Contact Mechanics, Cambridge University Press Cambridge, UK
(1985)


















BIOGRAPHICAL SKETCH

Darren McGuire was born the second son of Joseph and Louise McGuire

September 2nd, 1976. Darren lived in Wappingger Falls, New York until the age of 11

when he and his parents moved to beautiful Flagler Beach, Florida. Darren began his

engineering career at Santa Fe Community College and then transferred to the University

of Florida where he received his Bachelor of Science degree. After receiving the

University of Florida alumni fellowship award he went on to pursue his Master of

Science in the field of tribiology.