The Effect of Grease Composition on Fretting Wear

© 2019 Alireza Saatchi ALL RIGHTS RESERVED THE EFFECT OF GREASE COMPOSITION ON FRETTING WEAR

A Dissertation Presented to The Graduate Faculty of The University of Akron

In Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

Alireza Saatchi May 2019 THE EFFECT OF GREASE COMPOSITION ON FRETTING WEAR

Alireza Saatchi Dissertation

Approved: Accepted:

______Advisor Department Chair Dr. Gary Doll Dr. H. Michael Cheung

______Committee Member Dean of the College Dr. Qixin Zhou Dr. Donald P. Visco, Jr.

______Committee Member Dean of the Graduate School Dr. Rajeev Gupta Dr. Chand Midha

______Committee Member Date Dr. Curtis Clemons

______Committee Member Dr. Gopal Nadkarni

ii ABSTRACT

This research studies the effects of composition and oil release mechanisms of greases on its performance properties, especially against a certain type of tribological damage called fretting. Grease is a complex lubricant that consists mainly of a base-oil and a thickener.

Whereas the base oil is the primary lubricous component of the grease, the thickener gives the grease its consistency, or its ability to “stay put” wherever it is applied. Most thickeners are in the form of fibers dispersed colloidally in oil, entangled and connected together to form a three dimensional structure that traps the oil and prevents it from flowing freely. The oil needs to separate from the mixture and insert itself into the tribological contact in order to perform its function of preventing friction and wear.

Therefore, the separation of oil from the grease, which is loosely referred to as the grease bleed, is an essential step in the protection mechanism of the grease. The “bleed rate” is a standard specification of the grease determined by a test in which the grease is put in a cone sieve under high temperature and the relative amount of oil which is “bled” out is measured. Despite the widespread application of greases, their oil release mechanism is not well understood. The theories that have been developed for oil lubrication have been applied to grease lubricated contacts, which fail to accurately account for the experimental observations. Therefore the grease bleed mechanism is the major topic in this study. We will focus on the oil release mechanism of the grease under fretting contact. Fretting is defined as a small oscillatory displacement of bodies in sliding or rolling contact and is often observed when a source of vibration or cyclic stress is present, causing excessive wear or premature fatigue failures in different applications and it is considered as a plague of the modern industry. In many of these applications, grease is

iii employed as the lubricant. Another key motivation for this research is the fact that the term fretting in the literature and industrial standards fail to provide distinction between fretting under rolling and sliding contact. This is despite the fact that rolling and sliding fretting have been seen in numerous occasions to be two entirely different phenomena, especially as far as the performance of the grease is concerned, causing a great deal of confusion for both grease manufacturers and end-users. Finally, another motivation for this research is the investigation of non-soap thickened greases that unlike the more traditional soap thickened greases are not so well studied or understood especially their oil release mechanism which is for the most part, somewhat of a mystery. A good example of that is the calcium sulfonate grease that is among the newer grease types in the market and has exceptional properties. In grease making, this grease acts very differently than others in the way that it requires a lot more thickener to achieve the desired consistency. This is the reason for the extremely low bleed rate of the calcium sulfonate grease, however to the best of our knowledge, no one in the past had been able to explain the exceptionally good performance of this grease despite its nominal bleed rate being practically zero. This brings us back to the discussion of the grease bleed and points out the discrepancies in its definition. How can we define bleed as the oil release mechanism of the grease being responsible for its protection properties for the case of calcium sulfonate grease, where there is exceptional protective properties but zero bleed rate? Therefore, it seems to be necessary to redefine the bleed itself. To this end, three different grease types with soap and non-soap thickeners have been tested for their fretting wear performance. The soap thickener chosen was the lithium complex, which is one the most common grease thickener types of all times. We also tested non-soap,

iv calcium sulfonate and polyurea thickened greases for comparison. Each grease was mixed with its own base oil to obtain a range of bleed rates and then the mixtures were tested for their fretting performance under different rolling and sliding conditions to find a correlation between the grease bleed rate and its protective properties.

The results showed that the fretting performance of the grease under rolling contact strongly depends on the grease bleed rate whereas it did not affect the sliding fretting performance of the grease. Furthermore, there had been instances of scuffing which is the result of metal-to-metal contact due to failure in lubrication, which gave us a chance to study the starvation and scuffing resistance of the greases as well. There were some observations that were rather unexpected or somewhat hard to explain at first glance.

Firstly, the polyurea grease, which showed a comparatively high bleed rate, performed worse than others and especially showed the lowest resistance against scuffing.

On the other hand, the calcium sulfonate grease showed the best performance on all fronts, even though for most of the calcium sulfonate mixtures, the standard bleed rate was zero. The calcium sulfonate grease raised more questions as it showed an extremely non-linear bleed rate trend with the addition of oil.

The bleed rate results were interpreted within a model that correlated the bleed with the grease thickener geometries. To obtain the geometry of the actual thickeners, the greases were tested using Dynamic/Static Light Scattering (DLS/SLS) techniques. In the proposed model the grease thickeners were treated as particles suspended in a matrix of oil instead of the traditional view, in which the thickener structure was assumed to be a sponge that holds the oil and releases it into the tribological contact. In the present model, the “sponge” view is reserved only for the “static bleed” which simulates the oil flow

v behavior of the standard bleed test. Within the tribological contact, however, the

“dynamic bleed” should be viewed as the separation of oil from the loose and mobile thickener particles. In other words, the thickener particles are connected with each other and hold the oil, resulting in the consistency of the grease. However, at the same time, the thickener particles do not constitute a rigid body and can move independently in the oil matrix and are also capable of separating from each other if their attraction forces are overcome or in instances where they are diluted. Another significant difference between this model than those of the past is that a region surrounding each thickener particle, called the effective media, is assumed to have wholly immobilized the oil. In older models, the bleed was inaccurately assumed to occur from the viscous flow of oil through thickeners. In the current model, it was shown that the static bleed is indeed the viscous flow of only the unbound oil in a porous structure made of not only the thickeners but the entire effective media.

Further to clarifying the differences between rolling fretting and sliding fretting, in conclusion of this work, three different types of bleed mechanisms are discussed, static bleed, dynamic bleed under sliding and dynamic bleed under rolling contact. The difference between the two dynamic bleed mechanisms was attributed to the direction of the motion of the particles with respect to the sliding/rolling direction. In the light of this model, many aspects of the observed phenomena were also explained such as the unusual protection and bleed behavior of the calcium sulfonate grease which was related to the spherical geometry of its thickener micelles.

vi ACKNOWLEDGEMENTS

This dissertation is a summary of my Ph.D work which would not have been possible without the help, guidance, encouragements and well wishes of great mentors, colleagues, friends and family members.

At this moment of accomplishment, I wish to express my sincere gratitude to my advisor, Professor Gary Doll for his constant support, guidance, mentorship and encouragement throughout the five years of my research work. I would also like to thank

Dr. Paul Shiller for his invaluable suggestions, support and mentorship in every step of the development of this project.

I thank Professors Qixin Zhou, Rajeev Gupta, Curtis Clemons and Gopal

Nadkarni for serving as my committee members and providing valuable insights to improve this research.

I would like to thank Dr. S. Ali Eghtesadi for his guidance and contributions in microstructural characterizations and his advisor Professor Tianbo Liu for the support of this project. I thank Dr. Soroush Heidari Pahlavian for most of the artwork in this dissertation and his guidance regarding fluid dynamic problems. I sincerely thank Dr.

Gareth Fish from the Lubrizol Corporation for his mentorship and guidance as well as providing most of the baseline grease samples for this study. I thank Dr. Barbara Fowler and Richard Fowler for their help and guidance in this project and William Wenzel and

Brett Bell for technical support. I, along with our entire research team, acknowledge the

Timken Company for providing funding for this work and Drs. Kuldeep Mistry, Carl

Hager and Ryan Evans from Timken for their help and support of this project.

vii I also acknowledge the support of the faculty and staff of the University of Akron as well as my fellow graduate students at the Timken Engineered Surface Labs, and in the

Chemical Engineering Department for their generous support throughout this journey.

Finally, it is my fortune to acknowledge the people who mean a lot to me, my friends, near or far, and my family. I specially thank Diane Negi and her family for welcoming us to the United States and being nothing less than family to us. I thank my

Mother, Mansoureh, my Sisters Sara and Venus, my Brothers-in-law Arash, Sadid and

Sam and my Parents-in-law Maryam and Behzad for their constant love and support.

Finally, I gratefully dedicate this work to three people, my Father, Ahmad, my best mentor, my Wife, Elmira, my greatest companion, and my Son, Liam, my sweetest joy.

viii TABLE OF CONTENTS

Page

LIST OF TABLES ...... xii LIST OF FIGURES ...... xiii CHAPTERS I. INTRODUCTION...... 1 1.1 Background and Literature Survey ...... 1 1.1.1 Lubrication and Tribology ...... 1 1.1.2 Fretting ...... 2 1.1.3 Grease ...... 5 1.1.4 Grease Bleed and Application in Roller Bearings ...... 6 1.1.5 Saponification and Production of Grease ...... 7 1.1.6 The Microstructure of Thickeners in Greases and Characterization Methods ...... 8 1.1.7 Testing Lubricants for Their Fretting Wear Performances ...... 11 1.1.8 Quantification of The Fretting Damage ...... 13 1.1.9 Bleed and the “Sponge” Analogy ...... 16 1.1.10 Occurrence of Scuffing Under Sliding Condition ...... 18 1.2 Summary of Unknowns and Motivations for this Study ...... 20 1.2.1 Sliding vs. Rolling Fretting ...... 20 1.2.2 Bleed and Oil Release Mechanism of Grease ...... 21 1.2.3 Bleed in Non-Soap Thickened Greases ...... 21 1.2.4 The Effects of Grease Composition on its Protection Mechanism ...... 22 1.2.5 The Effect of Grease Composition on Scuffing ...... 22

II. EXPERIMENTS ...... 24 2.1 Overview ...... 24 2.2 Materials ...... 25 2.2.1 Greases and Base Oils ...... 25 2.2.2 Boron-Based Additives ...... 26

ix 2.3 Experimental methods and procedures...... 28 2.3.1 Overview ...... 28 2.3.2 Lubricant Testing ...... 29 2.3.3 Tribological Testing ...... 31

III RESULTS ...... 37 3.1 Standard Grease Parameters ...... 37 3.1.1 Bleed Rate ...... 37 3.1.2 Penetration (Grease Consistency) ...... 40 3.2 Morphological Characterization of Thickener Particles Using Laser Light Scattering ...... 42 3.3 Proof of Concept and Preliminary Experiments ...... 45 3.3.1 Overview ...... 45 3.3.2 Reproducing Reciprocating Sliding Fretting Test, SRV, with HFRR ...... 46 3.3.3 Reproducing Rolling Contact Fretting-False Brinelling, The Fafnir Test .. 49 3.4 High Amplitude Reciprocating Sliding Wear (H-1)...... 52 3.4.1 Lithium Complex Grease ...... 52 3.4.2 Calcium Sulfonate Grease ...... 54 3.4.3 Polyurea Grease ...... 55 3.4.4 Comparison Between Grease Types and Wear Scar Analysis ...... 56 3.5 Evaluation of Fretting Wear Performance in Sliding Contact (H-2) ...... 63 3.5.1 Lithium Complex and Calcium Sulfonate Greases ...... 63 3.5.2 Polyurea Grease and Occurrence of Scuffing ...... 65 3.5.3 Overall Comparison Between Grease Types ...... 67 3.6 Low Amplitude Fretting Tests (H-3) and Scuffing Analysis ...... 69 3.7 Rolling and Sliding Wear with Standard Tests ...... 73 3.7.1 Sliding Contact (SRV) ...... 73 3.7.2 Rolling Contact Fretting- False Brinelling (Fafnir) ...... 75 3.8 Boron Based Additives and WC/a-C:H Coating ...... 78 3.8.1 Overview ...... 78 3.8.2 The Effect of Boron Additives on Steel-on-Steel Contact ...... 79 3.8.3 The Effect of Boron Additives on Steel-on-WC/a-C:H Contact ...... 81

x 3.8.4 Boron Additives and DLC Coatings in Rolling Contact ...... 82

IV MODELING...... 84 4.1 Static Bleed Based on Permeability of the Thickener Structure...... 84 4.1.1 Background ...... 84 4.1.2 Permeability Model ...... 84 4.2 Dynamic Bleed ...... 94 4.2.1 Overview ...... 94 4.2.2 Interactions of the Thickener Particles with the Advancing Body ...... 96 4.2.3 Dynamic Bleed in Rolling vs. Sliding Contact ...... 103

V DISCUSSION ...... 107 5.1 Different Bleed Mechanisms ...... 107 5.2 Experimental and Model Fit for Static Bleed ...... 107 5.3 Overall Comparison of the Static Bleed of Different Thickener Types ...... 111 5.4 Rolling vs. Sliding Contact Fretting and their Relationship with Bleed Rate 112

5.5 Different Grease Thickeners Under Rolling Contact ...... 113 5.6 Grease Thickeners Under Sliding Contact and their Effect on Scuffing ...... 115

VI CONCLUSION ...... 117

VII SUMMARY ...... 122 7.1 Summary of the Most Important Findings of this Study ...... 122 7.1.1 Lubricant Test Results Summary ...... 124 7.1.2 Wear in Rolling and Sliding Fretting Results Summary ...... 124 7.1.3 Scuffing, B-based Additives & WC/a-C:H Coating: ...... 125

VIII FUTURE WORK ...... 126

REFERENCES ...... 128

xi LIST OF TABLES

Table Page

Table 2.1 Specification of baseline grease samples...... 26

Table 2.2 Specification of fully formulated base oils used for mixing in greases...... 27

Table 2.3 Specification of boron-based additives used for mixing in the ISO 460 grease. .... 28

Table 2.4 HFRR and TE-77 test parameters...... 33

Table 2.5 Fafnir Friction Oxidation test parameters...... 34

Table 3.1 The hydrodynamic radius (Rh(0)) and radius of gyration (Rg) of lithium complex and calcium sulfonate thickened greases measured by DLS/SLS techniques along with comparison of experimental and theoretical (Rg/Rh(0)) values...... 44

xii LIST OF FIGURES

Figure Page

Figure 1.1: The fretting contact [8]...... 4

Figure 1.2: The chemical structure of lithium stearate[45]...... 7

Figure 1.3: Self-assembly of thickener molecules and formation of micelles during saponification[48]...... 8

Figure 1.4: SEM image of a lithium hydroxystearate soap network after washing out the base oil [41]...... 9

Figure 1.5: AFM image of lithium complex (a), calcium sulfonate (b) and polyuria (c) thickened greases [41]...... 11

Figure 1.6: The SRV test apparatus [59]...... 12

Figure 1.7: The Standard Fafnir Fretting Oxidation Test apparatus [65]...... 13

Figure 1.8 : Energy partitioning of the dissipated energy and the activation of various damage mechanisms [10]...... 15

Figure 2.1 : Chemical structure of the used boron-based additives for the LiX-460 grease. 28

Figure 2.2: The standard cone penetration test for measuring the consistency of the grease [104]...... 30

Figure 2.3: The cylinder and mesh assembly for working the grease [104]...... 30

Figure 2.4. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10]...... 34

Figure 2.5. Smoothing the surface profiles by removing the ball or cylinder to be able to measure the wear volume...... 36

Figure 3.1. Bleed rate vs. percent added oil in the grease for different thickener types, lithium complex (LiX-100), calcium sulfonate (CaS-100) and polyurea greases (PU)...... 38

Figure 3.2. Bleed rate vs. percent added oil in the grease for different calcium sulfonate greases CaS-100 and CaS-Mix...... 39

xiii Figure 3.3. Bleed rate vs. percent added oil in the grease for different lithium complex greases LiX-100, LiX-220 and LiX-460...... 39

Figure 3.4. Pentration vs. percent added oil in the grease for different thickener types, lithium complex, calcium sulfonate and polyurea...... 40

Figure 3.5. Pentration vs. percent added oil in the grease for different lithium complex greases LiX-220 and LiX-460 and calcium sulfonate CaS-Mix...... 41

Figure 3.6. Bleed rate vs. Pentration for different grease thickener types, lithium complex, calcium sulfonate and polyurea...... 41

Figure 3.7. CONTIN analysis of DLS results obtained from diluted calcium sulfonate grease thickener particles...... 43

Figure 3.8. CONTIN analysis of DLS results obtained from diluted lithium complex grease thickener particles...... 44

Figure 3.9. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10]...... 46

Figure 3.10. Sliding contact fretting wear performance of the SHC 220 vs. grease bleed rate obtained from both SRV and HFRR...... 47

Figure 3.11. Sliding contact fretting wear performance of the SHC 460 vs. grease bleed rate obtained from both SRV and HFRR...... 47

Figure 3.12. Bearing raceway wear test of the SHC 220 false brinelling performance with 10% (left) and 20% (right) added oil, the latter showing close to zero wear...... 50

Figure 3.13. Rolling contact fretting wear performance of the SHC 220 vs. grease bleed rate obtained from Fafnir and TESL Modified Fafnir...... 51

Figure 3.14 Rolling contact fretting wear performance of the SHC 460 vs. grease bleed rate obtained from Fafnir and TESL Modified Fafnir...... 51

Figure 3.15 Load and displacement during the fretting test obtained from modified Fafnir.. 52

Figure 3.16. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10]...... 53

xiv Figure 3.17. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under grease lubrication of 220-Lithium Complex grease (LiC-220)...... 53

Figure 3.18. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under grease lubrication of 460-Lithium Complex grease (LiC-460)...... 54

Figure 3.19. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under grease lubrication of the Calcium Sulfonate Grease (CaS-Mix)...... 55

Figure 3.20. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under lubrication of the Polyurea (PU) Grease...... 56

Figure 3.21. Disk wear volume after reciprocating sliding (e=3.5) with HFRR under grease lubrication with different grease types. Polyurea disk wear volume is over 10000(µm^3). .. 57

Figure 3.22. Wear scar analysis of a typical ball and disc wear scar on ball and disc of HFRR experiment lubricated with lithium complex grease. The figure includes microscopy image (a and b) 3D depth map (c and d) and cross sectional profile (e and f) for disk and ball...... 59

Figure.3.23. Wear scar analysis of ball and disc wear features on ball and disc of HFRR experiment lubricated with baseline calcium sulfonate (CaS-Mix) grease which is more or less similar to the same for the baseline with added oil up to 40%. The figure includes microscopy image (a and b) depth map (c and d) 3D depth map (c and d) and cross sectional profile (e and f) for disk and ball...... 60

Figure 3.24. Wear scar analysis of a typical ball and disc wear scar on ball and disc of HFRR experiment lubricated with calcium sulfonate (CaS-Mix) grease with 50% added oil. The figure includes microscopy image (a and b) 3D depth map (c and d) and cross sectional profile (e and f) for disk and ball...... 61

Figure 3.25. Wear scar analysis of a typical ball and disc wear scar on ball and disc of HFRR experiment lubricated with polyurea grease. The figure includes microscopy image (a and b) 3D depth map (c and d) and cross sectional profile (e and f) for disk and ball...... 62

Figure 3.26. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10]...... 63

Figure 3.27. Wear volume of ball and disk after fretting wear (e=0.8) with HFRR under lubrication with 220-Lithium Complex grease...... 64

xv Figure 3.28. Wear volume of ball and disk after fretting wear (e=0.8) with HFRR under lubrication with 460-Lithium Complex grease...... 64

Figure 3.29. Wear volume of ball and disk after fretting wear (e=0.8) with HFRR under lubrication with calcium sulfonate (CaS-Mix) greases...... 65

Figure 3.30. Typical HFRR friction patterns of a normal tests (non-PU)...... 66

Figure 3.31. A HFRR friction patterns of tests with signs of scuffing (PU)...... 66

Figure 3.32 Ball wear scar analysis after testing of lithium complex (a), calcium sulfonate (b) and polyurea (c) for their fretting performance (e=0.8) with HFRR...... 68

Figure 3.33. Disk wear volume after fretting (e=0.8) with HFRR under grease lubrication with different grease types. Polyurea disk wear volume is over 7000(µm^3)...... 69

Figure 3.34. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10]...... 70

Figure 3.35. Typical friction pattern (left) and wear scar (right) of unscuffed fretting test. .. 72

Figure 3.36. Typical friction pattern (left) and wear scar (right) of scuffed fretting test...... 72

Figure 3.37. Occurance of scuffing vs. bleed for different greases...... 73

Figure 3.38. Standard SRV (TE-77) sliding wear test results of three grease types chosen with close base oil viscosities (LiX-100, CaS-100 and PU) drawn against the oil content for each grease mixture...... 74

Figure 3.39. Standard SRV (TE-77) sliding wear test results of three grease types chosen with close base oil viscosities (LiX-100, CaS-100 and PU) drawn against the oil bleed rate for each grease mixture...... 75

Figure 3.40. Standard rolling contact Fafnir Oxidation test results of three grease types chosen with close base oil viscosities (LiX-100, CaS-100 and PU) drawn against the oil content for each grease mix...... 77

Figure 3.41. Standard rolling contact Fafnir Oxidation test results of three grease types chosen with close base oil viscosities (LiX-100, CaS-100 and PU) drawn against the oil bleed rate for each grease mixture...... 77

xvi Figure 3.42 Intensity of scuffing on HFRR ball (a) and disk (b) in testing LiC#3 and that mixed with three different boron based additives in 1, 3 and 5 wt% concentrations...... 80

Figure 3.43 Intensity of scuffing on HFRR of 30K cycles with contact composed of steel ball (a) on WC/a-C:H disk (b) in testing LiC#3 mixed with TIPB in 1, 3 and 5 wt% concentrations...... 81

Figure 3.44 Intensity of scuffing on HFRR of 120K cycles with contact composed of steel ball (a) on WC/a-C:H disk (b) in testing LiC#3 mixed with TIPB in 1, 3 and 5 wt% concentrations...... 82

Figure 3.45 Mass loss of bearing raceways after Fafnir test with baseline LiC#3 grease and that mixed with 5% TIPB...... 83

Figure 4.1 The thickener particle lattice with particles of actual radius r, effective radius r’ and spacing a...... 86

Figure 4.2 Isolated pores for ...... 87

Figure 4.3 Point of transition to a permeable structure at ...... 88

Figure 4.4 Formation of channels for ...... 89

Figure 4.5 The shape and geometry of the pores...... 89

Figure 4.6 The tortuous path of fluid inside the pores...... 92

Figure 4.7 Interactions of the advancing body with a layer of thickener particles causing squeeze flow towards the contact (yellow arrow)...... 98

Figure 4.8 Analyzing particle geometry effects by estimating flow surrounding a single effective thickener particle...... 99

Figure 4.9 Rotation of a single thickener particle (blue) stuck between the advancing contact (gray) and another thickener particle or the remaining thickener structure (green) opposite of the direction of the motion (purple arrow) in response to the forces (yellow arrows) applied by the advancing body and the remaining thickener structure...... 100

Figure 4.10 Total clogging of the equivalent oil flow duct by the rotation of a plate-like effective thickener particle...... 100

xvii Figure 4.11 Cross-section thickness of a particle (tx) located at the center of the equivalent oil flow duct showing increase with rotation due to motion of the advancing body...... 102

Figure 4.12 The change in cross-sectional area of the equivalent oil flow duct with displacement normalized with the length of each particle...... 103

Figure 4.13 The direction of the motion of the thicker particles (purple arrows) and flow of oil (yellow arrows) with respect to the direction of the motion of the ball (gray arrows) for both rolling (top picture) and sliding (bottom picture)...... 105

Figure 4.14 The two-dimensional schematic drawing, showing the magnitude displacement of the thickener particle under rolling contact...... 106

Figure 5.1 Permeability of calcium sulfonate grease model vs. experimental result...... 108

Figure 5.2. The model and experimental results for the permeability of the thickener structure vs. the ratio of added oil to the CaS-Mix grease...... 109

Figure 5.3. Permeability of lithium complex (LiX-100) grease model vs. experimental result...... 110

Figure 5.4. Bleed vs. grease oil content for different thickener types...... 112

xviii CHAPTER I

INTRODUCTION

1.1 Background and Literature Survey

1.1.1 Lubrication and Tribology

To overcome or at least reduce the friction to a level that enables the machines to function without turning into scraps of metal, the first and most important action is lubrication [1]. Lubrication is basically interposing a lubricious substance between two surfaces under relative force and motion. The science of friction, wear, lubrication and the design of bearings, gears or any other industrial part subject to friction and wear is called tribology [2,3].

Friction causes inefficiency of the machinery and reduction in fuel economy. It is estimated that about one-third of the world’s energy resources are being dissipated as a response to friction [4]. The surfaces subject to pressure, affected by their relative motion, is of particular focus in tribology. The lubricant interacts with each surface forming a protective layer called the tribo-film which separates the surfaces in motion, lessens friction and thereby mitigates the damage and heat generation. Therefore, the tribo-film formation is essential in lubrication and plays a crucial role in the function of any tribological machine part [5,6].

1 With increasing performance demands for tribological systems, engineers have attempted to improve both lubricants and machinery parts. These improvements should be made based on the type of application and the type of damage that is encountered. This is where the importance of tribological tests or tribo-tests lies. The current research is about the function of the grease particularly for a particular damage called fretting.

1.1.2 Fretting

Fretting is distinguished from other types of wear by the existence of a small oscillatory displacement of bodies in contact [7]. The extent of the oscillatory movement should be less than the width of the contact area for the damage to be considered as fretting and is often observed when a source of vibration or cyclic stress is present. Fretting concerns many different industrial branches and is considered as the plague of modern industry [8].

In some applications, fretting may result in excessive wear that disrupts the function of the surface (fretting wear). In others, it may drastically drop the fatigue limit of the material resulting in premature catastrophic fatigue failures (fretting fatigue) [9]. In earlier studies on this type of wear, it was most commonly recognized and referred to as

“fretting corrosion” due to the reddish, muddy oxidization product that was observed in ferrous engineering components involved. The word “corrosion” was later eliminated for the general description of the fretting phenomena. This change in terminology happened when fretting was also observed between nonmetals and noble metals or metals under non-oxidizing conditions [7], and the term “fretting corrosion” is now reserved mostly for specific metallic applications.

The key difference between fretting and other types of wear lies under the fact that when the displacement is less than the width of the contact, there will be an area in the contact

2 that is never exposed to the surrounding environment (see Figure 1.1). This may result in the entrapment of wear debris within the contact during the tribological process [7,10].

Therefore many experts have linked cracking and wear in fretting to debris formation

[10–14].

Due to the widespread occurrence and the complex nature of the fretting phenomena, it has been extensively studied in different industrial and engineering branches [8]. One of the earliest theoretical approaches for modeling the fretting contact was based on a Hertz-

Mindlin analysis for dry condition, which is illustrated in Figure 1.1, for a conventional ball on flat sliding contact. The model, followed by experimental evidence, suggested the existence of a “stick zone” in the center of the fretting contact surrounded by the annulus domain where slip occurs [8,11].

Different sliding parameters such as displacement and frequency lead to change in the size of the stick and slip domains resulting in different fretting regimes with different damage modes. Therefore, many studies have been devoted to identifying these different fretting regimes [15–17] and determining transition criteria between them from the mechanical point of view [18–20]. Since the fretting regime determines the susceptibility of a fretting loading to a specific type of failure (i.e., wear or fatigue) the concept of fretting maps was later introduced [21–23]. In more recent studies, Fouvry and collaborators [8,9,24] summarized many of the prior works, assuming more general transition criteria for the fretting regimes, analyzed sliding behavior of different uncoated and coated surfaces in fretting contacts.

3 Figure 1.1 The fretting contact [8]. It should be mentioned that almost all of the cited works [7-24] experimented and developed theories and models for dry fretting, i.e., fretting in unlubricated conditions, even though many applications that are subject to fretting damage are lubricated [6,25–

36].

Another clarification that should be made is that when bearings undergo vibration while under loading, fretting, meaning the small oscillatory movement of the rolling ball, occurs on the raceways resulting in a similar damage as sliding fretting. This damage will cause imprints on the location of the center of the balls, comparable to the brinelling phenomena. However, the cause and mechanism of the damage is oscillation rather than indentation of the ball as it is in the case of brinelling. This type of fretting wear under rolling condition is therefore often referred to as false brinelling. Despite the apparent difference in rolling and sliding fretting, in current scientific and industrial terminology, such as in ASTM standards [64,65], both rolling and sliding oscillation is defined as

4 fretting when the amplitude is small enough. This results in a great deal of confusion when the greases are tested for their so-called “fretting” wear performance but could show contradictory results because one fretting test is under sliding and the other is under rolling condition. It will be seen in this study that the performance of grease under rolling fretting and is vastly different than and often opposite of that under fretting caused by oscillation under sliding contact. One of the primary goals of this study will, therefore, be to distinguish between these two types of fretting, namely, sliding fretting and rolling fretting or false brinelling and clarifying their difference and underlying mechanism as far as the function of the grease under such conditions is concerned.

1.1.3 Grease

Grease is a complex lubricant that consists of a base oil as the lubricious component and a thickener which provides consistency to the grease. Grease has rheological properties of a solid at rest. In other words, the thickener turns the viscous oil into a viscoelastic material that would “stay put” where it is applied. The consistency of the grease gives it the capacity to supply lubricant to the contact for long intervals. It also helps to prevent lubricant loss under operation and also avoiding penetration of contaminants such as solid particles and water; a comparatively simple sealing mechanism [39]. Grease is the preferred lubricant for hard-to-reach places and in any situation where lubricant leakage of the lubricated joints is a problem due to either gravitational, capillary or centrifugal forces. Therefore, in any application where grease is used rather than oil, consistency of the grease may be just as important as its lubricity for having proper and successful lubrication [40]. Some thickeners may also impart the grease with enhanced protection properties, such as calcium sulfonate that as a thickener in the grease, would also improve

5 protection for extreme pressure (EP) conditions. As there is usually a need for such performance improvements or some other particular property or function, most greases also contain some modifiers called additives. For example, there are EP additives for greases that function mainly by facilitating the interactions of the oil molecules with the surfaces of the contacting bodies and the formation of the tribo-film [39–41]. With all the above being said, the fact is that the oil remains as the main lubricous component of the grease responsible for separating tribological surfaces. Therefore, for the grease to function, it should be able to release its oil into the tribological contact. The oil release mechanism of the grease is often referred to as the grease bleed which will be explained as follows.

1.1.4 Grease Bleed and Application in Roller Bearings

While fretting happens in many different kinds of applications this research has been mainly directed towards solving fretting problems in roller bearing applications (e.g., in the wheel bearing for automotive applications or in wind turbines). Therefore, among different grease properties, the oil release capability of the grease, which is of enormous importance in bearing applications, is a major topic in this study. The oil release capability of the grease manifests itself in a certain standard characteristic of the grease called the “bleed rate.”

As mentioned before, grease is a multiphase system consisted of oil, thickener, and additives. Grease, as the reservoir for the oil, releases the oil and provides lubricity into the contact in a process often referred to as “oil bleeding.” The ability of the grease to bleed the base oil may be measured with standard tests [44]. To measure the bleed rate of the grease, emphasizing here that it is, in fact, the “static” bleed rate of the grease, it is

6 put into a sieve cone and held at a high temperature to let the oil bleed out. The amount of the oil that is bled out is then measured and reported in percent weight that will be identified as the bleed rate of the grease. This characteristic of the grease is found to have a significant correlation with the protective properties of the grease in rolling element bearing applications and has become one of grease selection criteria for such applications

[25]. However, it is noted here that the condition under which the oil is separated from the thickener in the tribological contact could be very different from the bleeding of oil from a static grease in a cone sieve. Therefore, in this study, the separation of oil from grease into the contact will be referred to as dynamic bleed.

1.1.5 Saponification and Production of Grease

The greases can be classified based on their thickener type as soap and non-soap thickened greases. As for the soap-thickened grease, which is the most common type, the thickener or the thickening agent is a metallic soap. Fatty acid soaps of lithium, calcium, sodium, aluminum, and barium are examples of soap thickeners [39]. Figure 1.2 shows the chemical structure of a lithium stearate, a fatty acid which consists of an ionic “head group” and a hydrocarbon chain as the “tail group”.

– H3C C C C C C C C C O C C C C C C C C C Li O

Figure 1.2: The chemical structure of lithium stearate[45].

7

The soap-thickened greases are prepared from fatty acids in a process called saponification. In this gelling process, the thickener molecules assemble into crystalline nuclei that grow into worm-like supramolecules or micelles in the form of fibers [46].

The fibers, colloidally suspended in oil, come into contact and entangle together with van der Waals forces [47]. The entangled and connected fibers form a continuous three- dimensional structure that traps the oil and prevents it from flowing freely. Therefore, most common physical models for the flow properties of the grease are based on the assumption that grease is a porous medium or a “sponge impregnated by oil” [40]. The explained process of the formation of micelles and their entanglement to form a three- dimensional structure are depicted in Figure1.3. In the following section, the structure of fibrous soap thickened greases will be discussed in more detail.

Figure1.3: Self-assembly of thickener molecules and formation of micelles during saponification[48].

1.1.6 The Microstructure of Thickeners in Greases and Characterization Methods

Since the microstructure of the grease contributes to most of its functional properties, it has been extensively studied by many groups, mostly using Scanning Electron

Microscopy (SEM) [38–44], [46–52]. Figure 1.4 is an SEM micrograph of a pre-treated lithium hydroxystearate thickener structure. The pretreatment is the removal or washing of the oil out of the structure using a solvent. As can be seen, the thickener is made up of

8 worm-like fibers that in the oil-washed-out state have formed a dense network still containing a considerable amount of inter-related and accessible pores. The reason for treatment of the grease before imaging is that the base oil that makes up most of the grease is volatile which means that vacuum imaging techniques such as SEM are not possible unless the base oil is removed from the grease [40].

Figure 1.4: SEM image of a lithium hydroxystearate soap network after washing out the base oil [41].

It should be mentioned, though, that the structure of the thickener may not be the same in the grease where the oil is also present, as it is in the oil-washed-out state. Many of the researchers felt that the structures they obtained are likely to be a consequence of the washing technique. Some have considered the structure observed by the SEM to be a

“collapsed” structure of the thickener fibers and “meaningless” as far as drawing conclusions about the actual thickener structure in the grease is concerned [37–40]. Even regardless of the washing technique, the fibers in air, in the absence of the viscous resistance of the oil may be more entangled, driven by their surface tensions and mutual interactions. Due to the importance of the grease microstructure and the fact that it

9 determines most essential properties of the grease, some studies have tried different imaging methods such as cryo-TEM and AFM that would be non-invasive or a little less invasive and did not require rigorous pretreatment of the grease [37–40,53]. Figure 1.5 shows AFM images of lithium complex (a), calcium sulfonate (b) and polyuria (c) thickened greases. An important note here is that the calcium sulfonate and polyuria greases do not have fibrous structures as lithium complex greases do. The calcium sulfonate thickener particles are spheres, and the polyuria grease is made of plate-like thickener particles. Another observation comes from the comparison between the AFM images of the lithium complex grease and the SEM image of a similar type of grease shown in Figure 1.4. This image confirms the worm-like structure that is twisted and connected, but the densely bonded structure as seen in the SEM image is not observed.

To summarize, the SEM which is the most commonly used technique for characterization of the thickener microstructure, images the thickener particles in an oil-washed-out state which measures a collapsed structure that is not a precise representation of the actual thickener structure as it is present inside of oil. Other methods that are less invasive or destructive such as cryo-TEM and AFM are far more accurate in describing the thickener structure but are less common due to their difficulties such as being timely and expensive. This calls for a cheaper and quicker method for characterizing the grease as it interacts with oil. Therefore, in this work, we will attempt to develop a new testing methodology for characterizing the grease microstructure.

10

Figure 1.5: AFM image of lithium complex (a), calcium sulfonate (b) and polyuria (c) thickened greases [41].

1.1.7 Testing Lubricants for Their Fretting Wear Performances

There are a few different standard tests for fretting wear performance of lubricants [57].

SRV and Fafnir Fretting Oxidation Test are two examples [64,65]. The SRV which is a high-frequency, linear oscillation test, uses a ball on disc reciprocating sliding contact under starved lubrication condition, whereas the Fafnir test measures fretting effects of a rolling element bearing with cyclic angular motion of small amplitudes in flooded condition. Figure 1.6 and Figure 1.7 show the SRV and Fafnir fretting test apparatus respectively. The SRV, is a widely used standard test methods for determining the fretting wear resistance of lubricating greases (ASTM D 7594), measuring friction and wear properties of lubricating greases (ASTM D 5707), and determining extreme pressure properties of lubricating oils (ASTM D 4721) and solid bonded films (ASTM D

7217). The Fafnir test, also considered a standard test method for fretting wear protection of lubricating greases (ASTM D 4170) under rolling condition resulting to false brinelling, has particular application for bearings which have been reported to fail prior to service due to the vibrations that may inevitably occur during shipment of vehicles. It is reminded here that in these standards the term “fretting wear” has been loosely used for both sliding and rolling contact. As will be also seen in the present study, the Fafnir test

11 results for commercial greases which is basically a false brinelling test, often contradicts results from SRV testing of same greases despite the fact that according to the standards they are supposed to be testing the same phenomena [58]. In this study, the difference between the two tests will be explained and distinguished as measurement of grease performance for two separate damage mechanisms. It is hypothesized that the SRV-type test results might be related to additive interactions while the Fafnir-type test results is contributed to the grease “bleed rate”. This subject will be discussed exclusively in later sections.

Figure 1.6: The SRV test apparatus [59].

12

Figure 1.7: The Standard Fafnir Fretting Oxidation Test apparatus [65].

1.1.8 Quantification of The Fretting Damage

As previously discussed, friction is the principal cause of wear and dissipation of energy in tribological systems [2]. Wear, progressive loss of material, constitutes a complex relationship with several processes including mechanical or chemical reactions depending on distinct physical parameters. Quantification of wear is necessary for predicting life of tribological parts in service. The most common wear model proposed in tribology is the

Archard model written as:

where V is the wear volume, P is the normal force, S is the sliding distance, H is hardness and K is the proportionality constant. In the Archard model, wear volume is related to the product of the sliding distance with the normal force applied [3]. Since both K and H are related to the material response, they can be combined to a single parameter k=V/PS, which is called the specific wear rate. As established by the Coulomb friction model

(F=µP), frictional force (F) is proportional to the normal load applied. If the friction

13 coefficient is taken as constant, wear volume would have a proportional relationship with the frictional force as:

In the case of unidirectional sliding FS would be equal to the work done by friction force

[3]. It has been shown in different studies that the value of k, the specific wear rate, depends strongly on the wear mode and parameters such as displacement amplitude, contact geometry and etc. Many of these parameters may undergo changes during the course of a tribotest, which may consequently change the wear mode beyond a critical point where the friction coefficient and the Archard wear coefficient changes drastically.

This points out the limitation in Archard’s formulation that would not integrate such variations [60]. To obtain a more accurate correlation between friction and wear, the so- called global energy-wear models was developed [10]. The energy description of wear relates material removal to the cumulated energy dissipation (Ed) through the interface of rubbing surfaces as:

where the constant values c1 and c2 are energy-wear coefficient and residual volume respectively. Extensive amount of experimental work confirmed energy dissipation as the primary parameter needed to quantify damage [3,8,10,24]. It should be noted that the frictional energy activates different processes and therefore dissipates through at least

14 three processes: rise in temperature, wear particle generation and material transformation at the interface such as mechanical/microstructural transformations and tribochemical phenomena [3]. Such partitioning of the system input energy is shown schematically in

Figure 1.8 which is presented by Fouvry et. al [10].

Figure 1.8 : Energy partitioning of the dissipated energy and the activation of various damage mechanisms [10].

Fouvry et al. proposed a question concerning the distribution of the accumulated energy in the different processes involved. To answer this question, they employed numerical simulations concluding that the particle generation and oxidation were the contact processes that consumed more energy. Their model is based on a postulate that if the wear mechanism remains the same the relative amounts of the distribution of the input energy to different mechanism remains constant. This postulate was not theoretically justified in their work.

15

1.1.9 Bleed and the “Sponge” Analogy

Oil in the grease is assumed to flow into the structure of the solid thickener microstructure, which is regarded as a porous matrix. In other words, the thickener is viewed as a “sponge” that holds and releases the oil [40]. Given the relationship between the bleed rate and protection properties of the grease, it is believed that the oil is replenished into the contact by bleeding, although the mechanism by which the oil release in the contact is activated is not clearly understood. Some researchers have assumed the centripetal force to be responsible for bleed in fast rotating applications [41] which is cannot possibly be the general case.

Based on the “sponge” assumption, the oil bleed has been modeled as a viscous flow through a porous media and has been correlated to the permeability of the porous structure involved [61]. These models have been based originally on works done by

Darcy (1856), who tested water flow through a fully saturated column of sand. His findings could be described as:

Where U is superficial fluid velocity through the porous media and is the frictional pressure drop across the length of . Fluid superficial velocity equals fluid volumetric flow rate divided by cross sectional area A or:

16

The Darcy model in its complete form is written as:

where is the fluid viscosity and is the permeability, which is a proportionality constant specific to the porous media, describing its ability to pass a fluid through its structure. According to Darcy’s law, the permeability of a porous structure (e.g. grease) is proportional to the flow rate of the liquid when frictional pressure drop, length, and viscosity can be assumed to be constant. Darcy’s law applies only when viscous shear forces dominate the friction so that inertia effects can be neglected. Baart et. al. [41] modeled the porous structure as orthogonally arranged fibers which would distort during the bleed process. The model was compared with the thickener structure observed by

AFM technique. An important part of their assumptions was based on the flexibility of the thickener structure.

In this type of models, the thickener is assumed to be the matrix holding the base oil.

There are a few reasons behind this assumption. The first reason is the fact that the thickener provides consistency to the grease by formation of a three dimensional structure throughout the grease. Although, the notion that this structure is one similar to a porous sponge comes mostly from the SEM images that were also proven to be inaccurate

17 [38,39] (see section 2.4). Nevertheless, it is generally intuitive to consider the grease as a sponge-like medium in different instances where oil is seen to bleed out of the grease.

Examples of such instances include the standard bleed test itself and also when the oil has been separated from the grease, commonly seen after long storage times. In this work a few arguments will be put forward against the sponge interpretation of the grease, and it will be argued that such analogy would only be useful for static bleed i.e. the standard bleed rate experiment. In the actual contact and where the oil release mechanism will be referred to as the dynamic bleed, the models based on the traditional interpretation fail to explain the driving force for the bleed or provide an accurate account of the phenomenon.

Even in the case of static bleed, the assumption that grease holds the oil in a three dimensional structure made of thickeners, and the static bleed is simply the viscous flow of oil through the thickener structure, is questionable considering the fact that the grease mostly consists of the base oil (80-95%). Even worse, a three dimensional structure of spherical thickeners of calcium sulfonate which should be separated because of dilution is virtually impossible. In this work the sponge interpretation of the grease in its static state will be fixed with a new model and in the light of that model, theories will be provided for explaining the function of the grease in dynamic mode under the tribological contact.

1.1.10 Occurrence of Scuffing Under Sliding Condition

Different machine elements and components involved in sliding under lubrication undergo various types of surface damage or wear that are usually gradual and well enough understood to be able to predict the rate of their progression and a reliable assessment of the longevity of the affected components exists. However, sliding surfaces under lubrication may undergo severe, unexpected or poorly anticipated damage called

18 scuffing [62,63]. Scuffing may happen due to insufficient lubrication. Following the starvation of the contact, scuffing results in rapid loss of the integrity and functionality of the surface of tribological components. It constitutes severe plastic deformation of the near-surface material accompanied by surface roughening which has sometimes resulted from local welding and material transfer [64,65]. Scuffing is best signified and characterized by a sudden and significant rise in friction, contact temperature and possibly, vibration and noise [64,66]. This type of failure may also be referred to as scoring, galling/frictional seizure and adhesive wear in the literature [63,66–68].

Due to the catastrophic nature and the significance of scuffing as a major reliability concern, it has been extensively studied for several decades with different approaches to predict or mitigate its occurrence. However, scuffing still remains one of the least understood tribological failure processes as it involves a complex combination of mechanical, chemical, metallurgical and thermal processes [69]. In situ observations and

X-ray diffraction of the contact during scuffing have shed some light on phase transformations [67,70–72] and the role of wear debris in the process [73,74]. Numerous theoretical and empirical approaches have been applied to model and predict the initiation of scuffing. In earlier models, the scuffing phenomenon was attributed to the failure of boundary film [75–77] or the elastohydrodynamic fluid lubricant film [78][79] triggered by the excessive local temperature at the contact. In these models, any metal-to-metal contact results in scuffing and a critical temperature for scuffing initiation was assumed.

Although, it has also been reported that sliding may continue without scuffing with minimal or no fluid films [80]. In other words, the failure of the fluid film or surface chemical films is an essential step in the initiation of scuffing but not necessarily

19 sufficient. Addressing many of the shortcomings of previously proposed mechanisms,

Ajayi et. al. [63,66] considered the surface material as the last element resisting catastrophic failure and related scuffing to the adiabatic shear plastic instability of the near-surface material. The sudden increase of contact temperature is assumed to be a result of the plastic instability and not the cause of it. They proposed that in severe contact conditions, plastic deformation at highest asperities causes two competitive processes, work hardening and thermal softening. According to the proposed mechanism, scuffing happens when the latter exceeds the former, resulting in a sudden and rapid burst of dislocation motion and severe plastic deformation accompanied by an increase in friction. This deformation consequently results in a sudden increase in temperature of the material in the contact. While the work of Ajayi et. al. has been by far the most successful in explaining many observations regarding the scuffing phenomena, there is still some ambiguity when it comes to incorporating the failure of fluid and boundary film into the mechanism.

1.2 Summary of Unknowns and Motivations for this Study

1.2.1 Sliding vs. Rolling Fretting

It was mentioned in section 1.1.7 that the term fretting has been loosely defined and used, describing different phenomena that appear to be substantially dissimilar to one another.

Specifically, fretting under sliding contact happens differently than that under rolling contact which results in false brinelling in applications such as in bearings. This becomes considerably problematic when greases are tested for their fretting performances. As was discussed before, for testing greases for their fretting wear performances, different

20 standard tests, all referring to “fretting wear” could give results in complete contradiction.

So the important question here is how the rolling and sliding contact fretting differ and why? In this work we will attempt to answer this question.

1.2.2 Bleed and Oil Release Mechanism of Grease

It has been previously reported that the bleed rate of the grease seems to have a strong effect on its false brinelling performance but not so much for its sliding contact fretting performance. While this research aims to verify this finding, it also attempts to answer many questions that arises from such observation. The main question being proposed here is, what is the oil release mechanism which brings about the main protection function of the grease and how and why it could change in different contact conditions.

The answer to this question should provide fundamental understanding to the function of the grease and its underlying protection mechanisms. As was mentioned in sections 1.1.4 and 1.1.9 the oil release mechanism of the grease is also poorly understood and to the best of our knowledge any model proposed so far, lack a few essential logical steps (e.g. failing to determine the driving force of the bleed) and may not justify or show an agreement with certain experimental observations.

1.2.3 Bleed in Non-Soap Thickened Greases

It was mentioned before that unlike soap thickened greases, non-soap thickened greases are not so well studied and there are many questions unanswered regarding their oil release mechanism. For example, the calcium sulfonate greases are seen to have exceptional performances in different applications while their bleed behavior has been

21 different. In this work we will try to first establish what these differences are and then attempt to explain them.

1.2.4 The Effects of Grease Composition on its Protection Mechanism

Another important question that, in a way, combines many of the mentioned questions is that how does different components and parameters in the grease play a role in the function of the grease. For example, what is the role of the grease thickeners in the grease protection mechanism and why do different grease types provide substantially different protection properties in different applications. When does the grease thickener becomes more important, when does the bleed rate play a more important role and when do the oil additives become prevalent? Following the problem of rolling and sliding fretting, this research also seeks to evaluate the hypothesis that while the grease bleed seems to have a more pronounced effect on rolling contact fretting, the sliding contact fretting may be more related to interactions of the lubricant and the tribological surfaces being more dependent on the effect of components such as additives in the lubricant or coatings on the tribological surfaces. While seeking a general understanding of the effects of these components could be helpful, a more important goal would be finding the underlying mechanism of these effects.

1.2.5 The Effect of Grease Composition on Scuffing

As mentioned before, starvation and breakdown of the boundary film occurs before the occurrence of scuffing. This explains the importance of additives which are responsible for facilitating the interactions of the surface and lubricants. A variety of lubricants and additive chemistries have been tested for their anti-scuffing performances [81–91]. It is

22 also seen that additives are not the only component of the lubricant that plays a role in the prevention of scuffing. In greases, the thickener type may also have an effect in prevention of scuffing. Greases with similar base oils and additive packages but different thickener types have shown to perform differently against scuffing [92–94]. However, to the best of our knowledge, prior to this work, no proper explanation had been provided for these reports.

23 CHAPTER II

2 EXPERIMENTS

2.1 Overview

The main goal of this research is to investigate the effects of grease composition on fretting phenomena. As explained before, among different properties of grease, its oil release mechanism under rolling and sliding conditions is of particular interest.

Therefore, the grease thickeners, their bleed rates and their effects on rolling and sliding fretting wear performance of the grease will be the main focuses of the experiments. For this purpose, three different grease types with lithium complex, calcium sulfonate and polyurea thickeners have been tested for their fretting wear performance. Each grease was mixed with its own base oil to obtain a range of bleed rates and then the mixtures where tested for their fretting performance under different rolling and sliding conditions to find a correlation between the grease bleed rate and its protection properties. A variety of stroke lengths were used for the sliding tests to evaluate the effects of general reciprocating sliding damage versus those specific to fretting with smaller stroke lengths.

In addition to the main series of experiments focused on the oil release of the grease, the effect of boron based additives on one of the grease types have also been studies. In one case the effect of a Diamond-like Carbon coating is also investigated. The following sections describe the details of the materials and methods used in this study.

24 2.2 Materials

2.2.1 Greases and Base Oils

The grease mixtures are prepared from both soap thickened and non-soap thickened greases. To study the effect of the thickener on bleed behavior of the grease, three different thickener types have been studied. Further to three soap thickened lithium complex greases used in this study, three non-soap thickened greases have also been used, two of which are calcium sulfonate thickened and one is polyurea thickened. The grease mixtures are formulated in a way to have different bleed rates. The bleed rate of the grease may be tuned with its oil content. Considering the grease as a porous structure of the thickener, the greater the oil content, the more separated the thickener particles will be and presumably, the higher the bleed rate. Such tuning of the grease bleed is usually limited by a loss in the grease consistency. To obtain different bleed rates, greases are mixed with their base oils, and in one case, a different base oil, in ten weight percent increment from 10% to 70%, which would potentially result in a wide range of bleed rates for each grease. The reason that no more than 70% oil was added to the grease was that further dilution of the grease would compromise its structural integrity and the mixture seizes to be a visco-elastic grease as has also been reported before [95]. The sample mixtures were blended by hand until a homogeneous mixture was obtained. The specifications of the greases and their base oils are brought in more detail in Table 2.1 and Table 2.2 respectively.

25 Table ‎2.1 Specification of baseline grease samples.

Code Thickener Type NLGI Thickener Base oil Density

Grade concentration Visc. at at 15.6

40 °C cSt °C

LiX-100 lithium (sebacate/12- NA 17% 112 0.92

hydroxystaerate 1:2)

LiX-220 lithium (sebacate/12- 2 NA 220 0.90

hydroxystaerate 1:2)

LiX-460 lithium (sebacate/12- 1.5 NA 460 0.93

hydroxystaerate 1:2)

CaS- 100 Calcium Sulfonate/ NA 16% / 23% 103 1.05

Calcium Carbonate

CaS- Mix Calcium Sulfonate/ 2 NA 75 1.00

Calcium Carbonate

PU Polyurea 3 20% 147 0.92

2.2.2 Boron-Based Additives

To study the effect of additives, especially to be examined for sliding contact, three boron-based additives were tested; namely Triisopropyl Borate (TIPB), 5-Indolylboronic

Acid (I-5) and Boric Acid (BA). Each additive was mixed into ISO 460 grease in three different concentrations of 1, 3 and 5wt%. As can be seen in Table 2.3 and Figure 2.1,

26 where the specification of the additives used are brought in more detail, the TIPB is liquid whereas the BA and I-5 are in powder form. Naturally, the latter two additives were mixed more aggressively and for a longer period of time to maximize dissolution.

However, it is noted here that the resulting mixture for both the BA and I-5 added grease samples were still visibly heterogeneous. The grease samples were put under mild vacuum for 10-15 minutes prior to experimentation to remove possible air bubbles that form as a result of mixing. To avoid the reaction of the additives with moisture, all of the mentioned mixing was done in a nitrogen box. The grease samples was used fresh once removed from their container.

Table ‎2.2 Specification of fully formulated base oils used for mixing in greases.

The grease Base oil Base oil Density at Commercial Base Oil mixed with Viscosity at Viscosity at 15.6 °C Name Similar to

40 °C cSt 100 °C cSt Grease

LiX-100 112 12.0 0.88 Yes

LiX-220 220 ISO VG 220 PAO Yes

LiX-460 460 ISO VG 460 PAO Yes

CaS-100 103 11.71 0.89 Yes

CaS-Mix 485 No

PU 147 0.92 Yes

`

27 Table ‎2.3 Specification of boron-based additives used for mixing in the ISO 460 grease.

Additive Triisopropyl Borate 5-Indolylboronic Acid Boric Acid

Code TIPB I-5 BA

Formula C9H21BO3 (C8H6N)B(OH)2 H3BO3

Form Liquid Powder Powder

Triisopropyl Borate 5-Indolylboronic Acid Boric Acid

Figure ‎2.1 : Chemical structure of the used boron-based additives for the LiX-460 grease.

2.3 Experimental methods and procedures

2.3.1 Overview

Different greases have different properties and not every grease is suitable for a given type of application. Many of these properties are evaluated in a number of standard tests that would determine the functional specifications of the grease [54–65]. Two of the standard grease testing used for the as-obtained grease mixtures were the Cone

Penetration and Cone Bleed measurements that were used to determine the Consistency and Bleed Rate of the greases respectively. Another test that was done on the grease, which was not among the common standard lube tests, was the characterization of 28 thickener particles using Dynamic/Static Light Scattering (DLS/SLS). This technique, to the best of our knowledge, had not been used for characterizing grease before. The grease mixtures were then tested for their sliding and rolling contact fretting performances under different tribotests. The details of the so called “lube tests” and “tribotests” are explained as follows.

2.3.2 Lubricant Testing

Grease Consistency - Standard Cone Penetration Test

The consistency of the grease was measured using standard cone penetration test ASTM

D-217 and ASTM D-1403 [62,63]. Figure 2.2 is a drawing of the cone penetration test apparatus. In this test a solid weight in the shape of a cone is dropped into a worked grease sample and the penetrated length is a measure of the consistency of grease. The cones are in full-size, half-size and quarter-size each pertaining to a specific standard.

The dropped length for the half or quarter-sized should be converted to that equivalent to the full-size which is defined as the consistency. Prior to this measurement, depending on the size of the cone and standard method used, the grease is worked to a certain number of strokes which basically means being smashed and squeezed through a coarse mesh, up and down a cylinder such as can be seen in Figure 2.3. This is expected to enable measuring the consistency of the grease closer to its operation condition. The consistency of the grease is among the most important characteristics of the grease as a selection criteria and also for grease classification. The consistency of the grease is also the main parameter for determining its NLGI grade which is one of the most commonly used classifications for greases.

29

Figure ‎2.2: The standard cone penetration test for measuring the consistency of the grease [104].

Figure ‎2.3: The cylinder and mesh assembly for working the grease [104].

Grease Bleed Rate - Standard Cone Bleed Measurement

The static bleed rate of the grease mixtures which were obtained as explained in section 2.2.1 were measured using the standard cone bleed test method ASTM D1742. In this test, the grease is put into a sieve cone with specifications detailed in the standard, weighed, and held at 100°C temperature for 30 hours to allow the oil to bleed out. The amount of the oil that is bled out by the end of the experiment and is gathered in the

30 container is then weighed and reported in percent weight, which will be identified as the standard bleed rate of the grease.

Grease Thickener Characterization - Dynamic and Static Light scattering DLS/SLS

The geometry and dimensions of the colloidal thickener particles have been evaluated using Dynamic and Static Light scattering techniques (DLS/SLS). Light scattering techniques have proved to be useful for characterization of micellar colloids [5]. No report is found in the literature on the use of such techniques for lubricating greases. This could be due to the fact that commercial greases are mostly opaque and do not allow the laser beam to pass through them. In the present work, we have managed to conceive of a methodology for assessing the geometries of lubricating greases using DLS/SLS. This was done by diluting the grease sample with its own base oil to a homogenous transparent solution which involved performing a few cycles of dilution sometimes with mild heating and stirring followed by separation and/or filtering.

2.3.3 Tribological Testing

Sliding Contact Fretting – SRV (TE-77)

Some of the grease mixtures were tested with a TE-77 which is a high frequency, linear- oscillation (SRV) testing machine to determine their fretting wear resistance according to the standard method detailed in ASTM D 7594. The test has been explained in section 1.1.7. The parameters and materials used for this test are detailed in the Table 2.4 in a column labeled as “SRV”.

31

Sliding Contact Fretting - High Frequency Reciprocating Rig (HFRR)

The grease samples of the three thickener types and various bleed rates have been tested for their fretting performance and scuffing occurrence using a High Frequency

Reciprocating Rig (PCS Instruments HFRR). The parameters used in these tests and the sections where the results are presented (last row) are shown in the Table 2.4. It can be seen that different stroke lengths, sliding ratios frequencies, temperatures, number of cycles and balls sizes have been used. The H-0 parameters were set to imitate the SRV test to assess the reproducibility of the experiments. The stroke lengths 20 and 222 are the equivalent of approximately 0.4 and 4 sliding ratios respectively. The sliding ratio is the ratio of stroke length over the contact diameter as can be seen in Figure 2.4 taken after

Fouvry [10]. H-1 tests were also done in similar sliding ratios but some other test parameters were changed for different reasons. The H-2, H-3, and H-4 reduced the sliding ratios to gross slip and partial slip sliding ratios respectively. The map in

Figure 2.4 shows the regions of different fretting/sliding regimes with different loads and sliding ratios. The marks on this map show the regimes for the parameters used for the tribotests in this study.

Rolling Contact Fretting - Fafnir Friction Oxidation Test

The Fafnir Friction Oxidation test was done according to the standard methods, materials and instruments detailed in the ASTM D 4170. As had been explained in section 1.1.7 and depicted in Figure 1.7 the grease is tested in a set of two ball bearings with oscillatory displacement and load applied by a spring. The parameters of the experiment are detailed in Table 2.5.

32

Table ‎2.4 HFRR and TE-77 test parameters.

TEST Code SRV H-0 H-1 H-2 H-3 H-4

Instrument used TE-77 HFRR HFRR HFRR HFRR HFRR

Normal load (N) 200 10 5 3 10 5

Max Hertz. P (GPa) 2.7 2.7 2.1 1.8 1.4 2.1

Stroke Length (µm) 1000 222 222 50 40 20

Sliding Ratio (e) 3.5 3.5 3.5 0.8 0.6 0.4

Frequency (Hz) 20 20 25 25 25 60

Temperature (C) 50 50 80 80 80 50

Number of cycles 360K 360K 45K 45K 45K 30K

Upper Specimen 52100 52100 52100 52100 52100 52100

Mat.

Ball Diameter mm 10 2.4 2.4 2.4 6 2.4

Lower Specimen 52100 52100 52100 52100 52100 WCC*

Lubrication 0.1 ml 4 µl 4 µl 13 µl 13 µl 4 µl

Results Section 3.3.2, 3.7.1 3.3.2 3.4 3.5 3.6 3.8.3

*Either the ordinary 52100 disk or otherwise 52100 coated with WC/a-C:H thin film was

used.

33

H-4 3 2 H-1

Figure ‎2.4. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10].

Table ‎2.5 Fafnir Friction Oxidation test parameters.

Parameter Value

Spring Load 2450 N (550 lb)

Oscillation Frequency 30 Hz

Stroke 0.21 rad (12)

Duration 22 hours

Lube – 1 g +/- 0.05 g

Rolling Contact Fretting – Modified Fafnir

The Modified Fafnir (MF) was basically an imitation of the Fafnir test with an MTS axial torsion load frame. An up and bottom housing for one set of bearing was designed and built for developing the MF. The other parameters were taken the same as the standard

Fafnir test, except the oscillation frequency that the load frame was unable to sustain.

Instead of 30 Hz, the frequency of 15 Hz was used and therefore the duration of the test

34 was extended to 44 hours. Aside from the frequency limitation, the hydraulic load frame showed a promising performance. It was capable of maintaining a steady load, resulting into increased reproducibility for the experiments. Furthermore, the load frame was also capable of measurement of torque and displacement which proved to be useful for analyzing the results. However due to technical problems only a limited number of experiments was successfully performed on the MF and unfortunately, the instrument ceased to function properly and therefore no more experiments was done using the MF.

Wear scar analysis – 3D Optical Profiler

Wear scars on the balls and discs from the HFRR experiments and those on the bearing raceways from the Modified Fafnir experiments were observed using a Zygo 3D profiler which uses light interferometry to obtain high resolution depth measurements which enables us to quantitatively measure the wear volumes from the tribological experiments.

In one of the steps in the data analysis of the surface profile measurements, the curvature of the surface in which the wear scar had been carved, is taken into account and removed from the results to smoothen the surface so that the volume of the void or wear can be compared to a leveled surface. As can be seen in Figure 2.5, the curvature of the spherical surface was removed before the wear volume on the balls wear calculated. In the case of the bearing raceways, a cylindrical surface was removed from the data to smoothen the surface. The wear volume for fretting experiments would then be defined as the volume bellow the leveled surface of the specimens which would be equal to the removed material from the surface. When scuffing occurred, a different convention was used and a new terminology, “Scuff Intensity”, was introduced for that which was reported instead of the wear volume. The scuff intensity was defined as the entire volume of the wear scar

35 meaning the sum of the volume bellow and the volume above the leveled surface of the specimen. The reason for defining and using this term instead of the standard wear volume will be explained in the results section.

Figure ‎2.5. Smoothing the surface profiles by removing the ball or cylinder to be able to measure the wear volume.

36

CHAPTER III

3 RESULTS

3.1 Standard Grease Parameters

3.1.1 Bleed Rate

The results of bleed tests on grease mixtures that were prepared as explained in section 2.2.1 can be seen in Figure 3.1, Figure 3.2 and Figure 3.3. In these figures the bleed rates of the grease mixtures are shown vs. the percent oil added to the baseline grease. In Figure 3.1 the bleed rates of the mixtures of lithium complex (LiX-100), calcium sulfonate (CaS-100) and polyurea greases (PU) that have close base oil viscosities, have been compared. The bleed rate of lithium complex and polyurea thickened greases increases more or less linearly with the oil content as it is expected.

Supposedly, by adding oil the porous structure made by the thickeners would expand, separating the thickener particles form one another resulting to an increase in pore size and permeability of the structure. However, contrary to what was expected and seen for the other two grease types, the bleed rate of calcium sulfonate grease does not show a linear relationship with oil content. The baseline calcium sulfonate grease has zero bleed.

The bleed rate stays zero up to 30% added oil. Beyond that, the bleed rate of the mixtures starts to increase with a slight slope up to 40% added oil. With further addition of oil, the bleed rate of the calcium sulfonate grease starts to increase very sharply. Such characteristic, specifically for the calcium sulfonate grease, has been recognized by grease manufacturers. This sudden jump in bleed rate with the addition of oil to the calcium sulfonate grease gives less room for flexibility in adjusting parameters for

37 making the grease and therefore makes its production more challenging than other grease types. This unusual behavior of calcium sulfonate that has not been properly explained before, will be described in light of the new model developed in this research and will be further discussed in section 4. This unusual behavior of calcium sulfonate that has not been properly explained before, will be described in light of the new model developed in this research and will be further discussed in section 4. To be able to make further comparisons in the mentioned modeling another calcium sulfonate grease (Cas-Mix) was mixed with a much higher viscosity oil than that of the original base oil. This gives a profile of viscosity in the mixtures, which would allow us to observe additional bleed regimes in the modeling section. At this point, it is useful to compare the overall behavior of the mentioned grease mix with that of the other calcium sulfonate grease as can be seen in Figure 3.2. The bleed rates of the three lithium complex grease mixtures are also shown and compared in Figure 3.3.

Figure ‎3.1. Bleed rate vs. percent added oil in the grease for different thickener types, lithium complex (LiX-100),

calcium sulfonate (CaS-100) and polyurea greases (PU). 38

Figure ‎3.2. Bleed rate vs. percent added oil in the grease for different calcium sulfonate greases CaS-100 and CaS-Mix.

Figure ‎3.3. Bleed rate vs. percent added oil in the grease for different lithium complex greases LiX-100, LiX-220 and

LiX-460.

39

3.1.2 Penetration (Grease Consistency)

The result of the penetration tests can be seen in Figure 3.4 and Figure 3.5. As can be seen in these figures, the consistencies of all three grease types change more or less linearly with increasing the oil content, which is a behavior that would be expected.

Looking at Figure 3.4 to make a comparison between the three grease types, it can be seen that this time, the calcium sulfonate grease acts in a similar way than the other two grease types. The change of the bleed rate with the change in the consistency of the lithium complex (LiX-100), calcium sulfonate (CaS-100) and polyurea greases (PU) is shown in Figure 3.6. One important point seen in this figure is that with similar consistencies, the highest bleed rates are achieved with the lithium complex greases and calcium sulfonate greases show almost zero bleed in the initial region. These behaviors will also be explained to some extent in the modeling and discussion sections.

Figure ‎3.4. Pentration vs. percent added oil in the grease for different thickener types, lithium complex, calcium

sulfonate and polyurea.

40

Figure ‎3.5. Pentration vs. percent added oil in the grease for different lithium complex greases LiX-220 and LiX-460

and calcium sulfonate CaS-Mix..

Figure ‎3.6. Bleed rate vs. Pentration for different grease thickener types, lithium complex, calcium sulfonate and

polyurea.

41

3.2 Morphological Characterization of Thickener Particles Using Laser Light

Scattering

The size distributions and their scattering angular dependency of calcium sulfonate and lithium complex grease thickener particles are shown in Figure 3.7 and Figure 3.8 respectively. The graphs in these figures show the intensity distribution of the colloidal particles at different hydrodynamic radiuses (Rh). As seen in Figure 3.7 the dynamic light scattering at different angles revealed that the size distribution of calcium sulfonate thickeners are angular independent suggesting the spherical geometry for these particles.

On the contrary, it can be seen in Figure 3.8 that the size of the lithium complex particles showed dependency to the scattering angle, which is due to its asymmetrical geometry.

To further analyze the morphology of particles in both samples, the hydrodynamic radius of particles at zero degree, Rh(0), was calculated by extrapolating Rh at different angles.

Rh(0) values were later compared with radius of gyration (Rg) of the same particles obtained from the static light scattering (SLS) between 50ο to 100ο angles using Zimm plot. The morphology of particles could then be assessed by calculating the ratio of the hydrodynamic radius at zero degree and the radius of gyration (Rg/Rh(0)). These results are shown in Table 3.1 It was found that this ratio for calcium sulfonate and lithium complex particles were around 0.8 and 1.7 respectively which was in good agreement with theoretical calculation over the structure of solid spheres and rods respectively. This finding is consistent with the reports found in literature about the same thickener types using different characterization techniques such as SEM and AFM that determined the lithium complex grease thickener particles to have wormlike structure and calcium sulfonate grease thickener structure to have spherical geometry [42]. In one of these

42 studies the geometry of the polyurea thickener particles had been determined as platelets.

In this study the geometry of the polyurea thickener particles could not be determined using the light scattering techniques due to the constituents of its base oil that gave way to different optical properties as the other two grease types that would make the determination of the size and geometry of the particles impossible.

Figure ‎3.7. CONTIN analysis of DLS results obtained from diluted calcium sulfonate grease thickener particles.

43 Figure ‎3.8. CONTIN analysis of DLS results obtained from diluted lithium complex grease thickener particles.

Table ‎3.1. The hydrodynamic radius (Rh(0)) and radius of gyration (Rg) of lithium complex and calcium sulfonate thickened greases measured by DLS/SLS techniques along with comparison of experimental and theoretical (Rg/Rh(0)) values.

Particle Lithium Complex Rod Calcium Sulfonate Solid Sphere

(experimental) (theoretical) (experimental) (theoretical)

Rh(0)* 176 nm - 63 nm -

Rg** 299 nm - 50 nm -

(Rg/Rh(0)) 1.70 1.73 0.79 0.77

*Hydrodynamic Radius

**Radius of Gyration

44 3.3 Proof of Concept and Preliminary Experiments

3.3.1 Overview

There were two objectives for these series of experiment which will serve as proof of concept for the proceeding experimental procedure. The first objective was to verify the hypothesis made for the effect of grease composition on sliding and rolling fretting. In other words, the intention was to verify if the change in bleed rate of the grease affects the grease performance in sliding contact fretting differently than that in rolling contact fretting and to see if a more pronounced effect is seen in the latter case. Secondly, the objective for these tests was to verify the validity of the results from the experimental setup developed in the University of Akron labs and facilities, for both sliding and rolling contact by comparing them to the results obtained from ASTM standard testing rigs that are commonly used in the industry. By a basic statistical analysis, the precision of the lab-scale measurements was also verified and compared to that of the industrial standard test. As explained in section 0, one of the ASTM standard testing for the measuring the grease performance in sliding contact fretting was done on the so-called SRV or TE-77 rig located in the Timken Company labs. A testing rig with the same essential ball-on- disc setup was designed with the HFRR in the TESL and the test results were compared to that of the TE-77. For rolling contact fretting, the standard Fafnir test results from

Timken labs were compared to those of the similar scale Modified Fafnir in the UA. The results from these experiments are discussed in the following sections.

45 3.3.2 Reproducing Reciprocating Sliding Fretting Test, SRV, with HFRR

The results of the testing of the grease sliding fretting performance of the SHC 220 and

SHC 460 are shown respectively in Figure 3.10 and Figure 3.11. The wear volumes were obtained by both the SRV and HFRR (H-0) and have been compared with one another in both figures. The details of the experimental parameters are listed in Table 2.4 and the location of the test conditions on the fretting map is shown in Figure 3.9.

Figure ‎3.9. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10].

Regarding the first objective of these experiments as explained in the previous section, it can be seen that the sliding contact fretting under both SRV and HFRR shows no dependence on the grease bleed rate for either of the two greases. As it regards to the second objective, an overall similarity can be seen between the lab scaled Modified

HFRR and the TE-77 test results. Actually, judging by the variability of the HFRR data as compared to that of the TE-77, it can be said that precision of the test was improved

(i.e. much less scatter in the data).

46 Figure ‎3.10. Sliding contact fretting wear performance of the SHC 220 vs. grease bleed rate obtained from both SRV

and HFRR.

Figure ‎3.11. Sliding contact fretting wear performance of the SHC 460 vs. grease bleed rate obtained from both SRV

and HFRR.

47 Furthermore, with the use of the optical surface profiler, the wear volume can be quantitatively measured which is a great improvement in the accuracy of the measurement over the normalized wear area. According to the literature [8], the dissipated energy should relate to the wear volume that is expected to be proportional to the “normalized wear area” which is used in the standard method. As will be seen in other sections, the normalized wear area is not necessarily a good representative of the wear volume whereas with the optical surface profiler the wear volume is directly measured. It should be noted that, generally, the TE-77 is considered to be a more reliable testing rig than smaller rigs such as the HFRR because it simulates conditions that are closer to those occurring in real industrial machinery. The smaller scale used in HFRR would specially affect the rate of heat dissipation through conduction which makes a difference in smaller sliding amplitudes as will be seen in future sections. However, as far as the oil release performance of the LiX grease is concerned, the rigs show similar results. Moreover, the improved precision and accuracy in wear volume measurement that is brought about by the optical profiler could be advantageous for these experiments.

The higher sensitivity will allow reproducible and meaningful measurements in much smaller number of testing cycles. This will be shown and explained in the following sections.

To summarize, the above results showed, first and foremost, that the sliding contact fretting does not correlate to the bleed rate of the lithium complex greases. This behavior will be compared to that under the rolling sliding in the next section. Additionally, the experimental setup developed in this project was shown to produce similar results as the

48 industrial-scale standard test with improved accuracy and precisionwhich would also allow the assessment of the grease performance in fewer cycles for future experiments.

3.3.3 Reproducing Rolling Contact Fretting-False Brinelling, The Fafnir Test

The performance of greases under rolling contact fretting with standard Fafnir and

Modified Fafnir (MF) are compared and shown against bleed rate for the lithium complex, 220 and 460 grease in Figure 3.13 and Figure 3.14 respectively. It can be seen, that unlike the case of sliding wear, there is a strong dependence of rolling contact fretting wear on bleed rate. The measured wear starts as a high number at low bleed rates and drops sharply with increase in bleed rate, reaching zero at about 5% bleed rate. This is true for both grease types and on both Fafnir-type testing rigs. Again, it can be seen that the precision of the experiments has been significantly reduced in most of the cases.

Another observation is that although the testing parameters and the bearing used in the

MF are set to be the same as those in the standard Fafnir rig, on bleed rates lower than

5% where fretting wear occurs, it is more sever for the case of MF. This could be due to the fact that the load frame used in the MF supposedly holds the nominal load more steadily than standard Fafnir in which a spring is responsible for the same job.

Figure 3.15 shows the load and displacement during the fretting test cycle of the MF.

This result is useful in ensuring the integrity and quality of the fretting tests which provides a considerable advantage over the standard Fafnir which does not record either of these parameters.

Going back to general implications of these results, namely, the significant reduction of wear with bleed rate and the 5% drop-to-zero point, an important note here is that the standard limit of allowable bleed rate is 5%. Therefore, for most industrial applications,

49 including those prone to rolling contact fretting, or false brinelling, due to other consideration, such as life of the grease, the target bleed rate for grease manufacturing is within the 5% range which is susceptible to false brinelling or fretting. Therefore, merely choosing a higher bleed rate to reduce rolling contact fretting is not a practical solution to this problem. However, based on these results the effect of oil-thickener interaction on rolling contact fretting performance of the grease seems to be evident, which could be an initial step towards finding a more concrete solution to the problem, possibly with a different approach, such as the choice of the grease thickener.

Figure ‎3.12. Bearing raceway wear test of the SHC 220 false brinelling performance with 10% (left) and 20% (right)

added oil, the latter showing close to zero wear.

50 Figure ‎3.13. Rolling contact fretting wear performance of the SHC 220 vs. grease bleed rate obtained from Fafnir and

TESL Modified Fafnir.

Figure ‎3.14 Rolling contact fretting wear performance of the SHC 460 vs. grease bleed rate obtained from Fafnir and

TESL Modified Fafnir.

51

Displacement(mm)

mm)

-

Torque(N

Cycles

Figure ‎3.15 Load and displacement during the fretting test obtained from modified Fafnir.

3.4 High Amplitude Reciprocating Sliding Wear (H-1)

3.4.1 Lithium Complex Grease

In this section the performance of different grease types was tested under sliding contact with a displacement amplitude well above the limit of gross slip fretting regime and in the reciprocating sliding regime with the sliding ratio of e=3.5. The maximum Hertzian contact pressure for these test were 2.1 GPa and the tests were run for 45,000 cycles. The details of the experimental parameters are listed in Table 2.4 and the location of the test conditions on the fretting map is shown in Figure 3.16. The arrow shows how the test parameters changed from the location in the map with respect to the last HFRR experimental results presented.

Grease mixtures with increasing oil content and bleed rates of the LiC- 220 and 460 were used and wear volume under reciprocating sliding on the ball and the disk was obtained

52 and drawn against the bleed rate of each grease sample shown in Figure 3.17 and

Figure 3.18 respectively. Expectedly there is not a significant dependence of the grease performance under sliding condition on its bleed rate. This is true for both greases and the overall performance of the two greases were generally in a close range for these tests.

Figure ‎3.16. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10].

2500 Disk

2000 Ball

) 3 1500

1000

Wear Vol. (µm Vol. Wear 500

0 0 5 10 15 20 25 Bleed rate (%)

Figure ‎3.17. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under grease lubrication

of 220-Lithium Complex grease (LiC-220).

53

2500 Disk

2000 Ball

) 3

1500

1000

Wear Vol. (µm Vol. Wear

500

0 0 5 10 15 20 25 Bleed rate (%)

Figure ‎3.18. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under grease lubrication

of 460-Lithium Complex grease (LiC-460).

3.4.2 Calcium Sulfonate Grease

The performance of the calcium sulfonate grease mixtures with varying bleed rates under similar conditions are shown in Figure 3.19. Although for the case of calcium sulfonate there is a slight reduction after about 1 per cent bleed rate, it can be said that the overall change in the ball and disk wear volume with bleed rate is insignificant. It should be noted that unlike the other grease types, the baseline grease and the first few mixtures show zero bleed. As can be seen in the figure, the wear volume shows a slight decrease as the grease starts to bleed although the reduction could be considered within the uncertainty of the measurement for the most part. After 1 percent of bleed rate and for the last two data points which there is more than about 2 percent increase in the bleed rate, the wear volume stays constant.

54

2000 1800 Disk

1600 Ball

)

3 1400 1200 1000 800

600

Wear Vol. (µm Vol. Wear 400 200 0 0 1 2 3 4 5 Bleed Rate %

Figure ‎3.19. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under grease lubrication

of the Calcium Sulfonate Grease (CaS-Mix).

3.4.3 Polyurea Grease

The ball and disk wear volume versus bleed rate of polyurea grease is shown in

Figure 3.20. While the polyurea grease seems to performs quite differently than the other two grease types in terms of overall performance, with much higher ball and disk wear, but regarding dependency of performance on bleed rate, the polyurea grease mixtures act similarly to other greases, showing no significant change with the increase of the bleed rate. Another point to note is that for the case of the polyurea, the base line grease has the highest bleed rate among the tested greases.

55

30000 Disk 25000

Ball

) 3 20000

15000

10000

Wear Vol. (µm Vol. Wear 5000

0 3.385 4.27 7.13 8.53 13.57 24.9 Bleed rate (%)

Figure ‎3.20. Wear volume of ball and disk after reciprocating sliding wear (e=3.5) with HFRR under lubrication of the

Polyurea (PU) Grease.

3.4.4 Comparison Between Grease Types and Wear Scar Analysis

To be able to draw an overall conclusion on the experiment results of this section, the performance of the greases have been compared. Figure 3.21 shows the comparison between volumes of reciprocating sliding (e=3.5) wear on disks after testing with HFRR under lubrication with different grease types. It can be seen that while wear volume for all the grease types show little to no dependence on the grease bleed rate, the type of grease used could have significant impact on the performance of the grease. The calcium sulfonate grease shows a slight improvement over the lithium complex greases and the polyurea shows significantly higher wear rates than the other two grease types. The effects of grease type show no relationship to their measured bleed rate either. It can be seen that polyurea, which is the grease type with highest bleed rates, shows significantly lower protection under reciprocating sliding condition than the other two grease types.

56

This result is somewhat against what would be expected. If anything, higher bleed rate is expected to improve the performance of the grease, whereas the opposite is seen here, which implies that the way different grease types are performing does not relate with their measured bleed rate.

2500

2000

) 3 1500

1000

Wear Vol. (µm Vol. Wear Calcium Sulfonate 500 220 460 0 0 2 4 6 Bleed Rate (%)

Figure ‎3.21. Disk wear volume after reciprocating sliding (e=3.5) with HFRR under grease lubrication with different

grease types. Polyurea disk wear volume is over 10000(µm^3).

To further understand the observed behavior, the wear scars have been analyzed using the optical surface profiler. Figure 3.22-Figure 3.25 shows typical wear scars for each of the grease types. It can be seen that the wear scars under lithium complex grease lubrication have symmetrical boat-like wear scar. Such wear scars are exactly what is expected of the reciprocating sliding wear[105]. Surprisingly, the samples lubricated with the other two grease types did not wear in similar fashion. For the case of polyurea grease, some raised material is also observed on the wear scar, further to cavities from removed material.

Such features are usually accompanied by some special phenomenon such as micro

57 welding or scuffing. While the friction pattern for the reciprocating sliding tests done on polyurea grease does not show sudden and significant rise in friction coefficient, such that would have been indicative of occurrence of scuffing, metal-to-metal contact and microwelding remains as the cause for the special features on the wear scars from tests done on polyurea samples. According to the literature[66] such features could also be attributed to occurrence of micro-scuffing in confined “quenched” regions.

The most unusual and unexpected wear scars belong to calcium sulfonate samples. As can be seen in the figure, the wear scars for calcium sulfonate grease is composed of shallow or unworn regions next to isolated deep scars. As explained before, this results in an overall lower wear volume than that for the lithium complex grease. The origin of the deep wear scars is unknown. One of the few cases where similar features appear is under certain fretting conditions. Especially, in higher oil concentration which is also accompanied by increase in bleed rate and slight drop in wear, the unworn region in the center of wear scar grows in size (Figure 3.24) forming a shape quite similar to wear features in partial slip/mixed fretting regimes in dry contact[106].

58

(a) (b)

(c) (d)

(e) (f)

Figure ‎3.22. Wear scar analysis of a typical ball and disc wear scar on ball and disc of HFRR experiment lubricated

with lithium complex grease. The figure includes microscopy image (a and b) 3D depth map (c and d) and cross

sectional profile (e and f) for disk and ball.

59

(a) (b)

(c) (d)

(e) (f)

Figure.‎3.23. Wear scar analysis of ball and disc wear features on ball and disc of HFRR experiment lubricated with baseline calcium sulfonate (CaS-Mix) grease which is more or less similar to the same for the baseline with added oil

up to 40%. The figure includes microscopy image (a and b) depth map (c and d) 3D depth map (c and d) and cross

sectional profile (e and f) for disk and ball.

60

(a) (b)

(c) (d)

(e) (f)

Figure ‎3.24. Wear scar analysis of a typical ball and disc wear scar on ball and disc of HFRR experiment lubricated

with calcium sulfonate (CaS-Mix) grease with 50% added oil. The figure includes microscopy image (a and b) 3D

depth map (c and d) and cross sectional profile (e and f) for disk and ball.

61

(a) (b)

(c) (d)

(e) (f)

Figure ‎3.25. Wear scar analysis of a typical ball and disc wear scar on ball and disc of HFRR experiment lubricated

with polyurea grease. The figure includes microscopy image (a and b) 3D depth map (c and d) and cross sectional

profile (e and f) for disk and ball.

62

3.5 Evaluation of Fretting Wear Performance in Sliding Contact (H-2)

3.5.1 Lithium Complex and Calcium Sulfonate Greases

In this section, the results from testing the effects of grease type and bleed rate in gross slip fretting regime is presented. The details of the experimental parameters are listed in

Table 2.4 and the location of the test conditions on the fretting map is shown in

Figure 3.26. The arrow shows how the test parameters changed from the location in the map with respect to the last HFRR experimental results presented.

Figure ‎3.26. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10].

The sliding ratio for these tests are just below one (e=0.8), meaning the sliding distance is less than the width of the contact diameter and therefore the synergic effects of debris retention and mechanical effects of fretting are expected to be incorporated into the damage. Four greases from the three general grease types under study, the LiC-220 and

460 and the CaS-Mix and the PU were tested for their fretting performances. The wear results from these tests are shown in Figure 3.27, Figure 3.28 and Figure 3.29 against the bleed rate of the grease mixtures. Similar to what was seen before for wear in higher sliding ratios, fretting wear does not greatly depend on the grease bleed rate. For the LiC-

460 and CaS-Mix a slight decreasing trend in fretting wear can be seen with increasing 63 bleed rate but it’s not significant, especially considering the uncertainty of the measurements or in other words, the variability of the results for each data point.

3000 Disk 220

2500 Ball 220

) 3 2000

1500

1000

Wear Vol. (µm Vol. Wear

500

0 0 5 10 15 20 25 Bleed Rate (%)

Figure ‎3.27. Wear volume of ball and disk after fretting wear (e=0.8) with HFRR under lubrication with 220-Lithium

Complex grease.

3500 Disk 460 3000

Ball 460

) 3 2500

2000

1500

Wear Vol. (µm Vol. Wear 1000

500

0 0 5 10 15 20 25 Bleed Rate (%)

Figure ‎3.28. Wear volume of ball and disk after fretting wear (e=0.8) with HFRR under lubrication with 460-Lithium

Complex grease.

64

2000 1800 Ball

1600

) Disk 3 1400 1200 1000

800

Wear Vol. (µm Vol. Wear 600 400 200 0 0 1 2 3 4 5 Bleed Rate (%)

Figure ‎3.29. Wear volume of ball and disk after fretting wear (e=0.8) with HFRR under lubrication with calcium

sulfonate (CaS-Mix) greases.

3.5.2 Polyurea Grease and Occurrence of Scuffing

For the PU grease, the general behavior of the materials was entirely different which seemed to be resulting from an entirely different phenomenon. Signs of scuffing was seen before, for reciprocating sliding wear under PU lubrication. In the case of lower amplitude fretting condition (e=0.8) the occurrence of scuffing is more evident. As can be seen by comparing Figure 3.30 and Figure 3.31, unlike all other grease types in these experiments, the tests done on PU resulted in friction patterns with sudden and significant jumps in the friction coefficient, which is indicative of the occurrence of scuffing.

65

Figure ‎3.30. Typical HFRR friction patterns of a normal tests (non-PU).

Figure ‎3.31. A HFRR friction patterns of tests with signs of scuffing (PU).

66

As can be seen in Figure 3.32 (c), the damage analysis using the optical profiler also confirms significant amount of material transfer with large amount of raised materials present both on the ball and the disks. The volume of damage for all the PU greases are several times larger than that of other greases. The reason that the term “damage” was used hear rather than referencing to wear, is that the damage here, does not only consist of material removal (i.e. wear) but rather a significant amount of the damage is resulted from material transfer and plastic deformation. Indeed, even in measuring the damage, if the wear, meaning the removed material from the surface is recorded the same way as in other experiments, no meaningful and reproducible result will be obtained. In such a case the data will have huge scatter with no consistency or repeatability. Therefore, to be able to accurately measure the effects of the scuffing damage with acceptable precision, the raised material is also taken into account, and is introduced as a new parameter called the

“scuff intensity”. The discussion on scuffing is done in further detail in following sections where lower amplitude test results are discussed and scuffing becomes the prevalent damage mechanism.

3.5.3 Overall Comparison Between Grease Types

In Figure 3.32 the wear scars created under lubrication with lithium complex and polyurea thickened greases can also be seen. Once again for the case of lithium complex greases the wear scars follow the expected typical features of fretting wear, with deeper wear at the periphery of the scar. The calcium sulfonate however, produces distinctively different wear scars with points of shallower wear surrounded by pit-like damage over the surface of the contact area.

67 The fretting wear volume under lithium complex and calcium sulfonate greases are compared together in Figure 3.33. The 220 and 460 lithium complex greases, once again, perform similarly, showing fretting performances in the same range. However, this time the protection properties of the calcium sulfonate grease under fretting conditions show significant improvement than that of the lithium complex grease. The polyurea grease performs poorly under fretting condition with occurrence of scuffing and damage that measures three to five times as much as that for the other grease types (>7000µm^3).

(a) (b) (c)

Figure ‎3.32 Ball wear scar analysis after testing of lithium complex (a), calcium sulfonate (b) and polyurea (c) for their

fretting performance (e=0.8) with HFRR.

68

3500

3000

) 3 2500

2000

1500

Wear Vol. (µm Vol. Wear 1000 460 500 220 Calcium Sulfonate 0 0 1 2 3 4 5 6 Bleed Rate (%)

Figure ‎3.33. Disk wear volume after fretting (e=0.8) with HFRR under grease lubrication with different grease types.

Polyurea disk wear volume is over 7000(µm^3).

3.6 Low Amplitude Fretting Tests (H-3) and Scuffing Analysis

As seen in the previous section, another damage phenomenon called scuffing started to happen for one of the grease types when reducing the sliding ratio bellow one, into the gross slip fretting regime. Further reduction in the sliding ratio will result into scuffing becoming the prevalent damage mode. The details of the experimental parameters are listed in Table 2.4 and the location of the test conditions on the fretting map is shown in

Figure 3.34. Same as before, the arrow shows how the test parameters changed from the location in the map with respect to the last HFRR experimental results presented.

69

Figure ‎3.34. HFRR test parameters shown in terms of the sliding ratio in a fretting map [10].

As explained in the previous section, the reciprocating tests were performed at different sliding ratios. Similar to Figure 3.34, in Figure 2.4 shown earlier, the marks on the map contain all of the fretting/sliding regimes for the parameters used for the tribotests in this study. The H-0 and H-1 tests with highest stroke lengths and sliding ratios are in reciprocating sliding regime. As explained in previous sections in those tests, the friction coefficient remained more or less constant after it plateaued during the test and the wear scars looked normal with no sign of raised material or scuffing. Figure 3.35 shows a typical friction pattern of the tribotest along with the topography of a typical wear scar in this regime.

On the other hand, the H-4 test has the lowest stroke length relating to a sliding ratio in the partial slip-fretting regime. With little exception, a sudden and intense increase in the friction coefficient was observed in these tests, which is characteristic of the occurrence of scuffing[107]. This was confirmed by the observation of the unusual shape of wear scars with raised material; signs of micro-welding and material transfer. Figure 3.36 shows a typical friction pattern and wear scar of the tests in which scuffing has occurred.

70

Scuffing happens due to excessive heat generation in the contact, usually as a result of lubricant starvation[107]. In fact, the HFRR with a 6 mm ball, was originally designed for testing the lubricity of diesel fuels [108] and without modification, may not be suited for testing of grease for its fretting wear performance in such low sliding amplitudes where the generated heat does not get a chance to dissipate. In middle range stroke lengths which correlates with sliding ratios in gross slip fretting regimes (H-2 and H-3) the testing condition is not as bad as H-4 but is still not perfectly suited for fretting tests since some samples still suffer scuffing. However this situation could be ideal for evaluating the occurrence of scuffing due to the fact that the heat generation seems to be at the borderline critical amount for occurrence of scuffing. This assumption comes from the fact that in these sliding ratios, scuffing occurred randomly with different variability in occurrence from one grease sample to another. The randomness in the occurrence of scuffing is likely to be a result of the randomness in surface asperities and roughness, which for a given contact pair, depending on the lubrication/starvation condition, may or may not result in scuffing. This gave us the ability to analyze the scuffing performance of the grease samples. Figure 3.37 shows the percent occurrence of scuffing in reciprocating test for the grease samples with different thickener types and bleed rates. As can be seen for lithium complex greases the occurrence of scuffing reduced with increase in bleed rate. The relevance of scuffing to bleed cannot be immediately concluded due to the contradictory observations made for the other two grease types. As can be seen in

Figure 3.37 the calcium sulfonate grease samples, most having zero bleed, gave no scuffing. The polyurea greases, some with very high bleed rates scuffed without exception. These conflicting observations as regards to the effect of bleed rate to scuffing

71 protection will be discussed further in following sections where a model will be developed for describing the bleed which will shed some light on the observed phenomenon, making it more understandable.

0.7

0.6 0.5 0.4 0.3 0.2

Friction Coefficient Friction 0.1 0 0 100 200 300 400 500 Time

Figure ‎3.35. Typical friction pattern (left) and wear scar (right) of unscuffed fretting test.

0.7

0.6 0.5 0.4 0.3 0.2

Friction Coefficient Friction 0.1 0 0 200 400 Time

Figure ‎3.36. Typical friction pattern (left) and wear scar (right) of scuffed fretting test.

72 100

80 Polyurea 60 Lithium Complex Calcuim Sulfonate 40

20 Scuffing Occurance (%)Occurance Scuffing

0 0 10 20 30 40 50 Cone Bleed (%)

Figure ‎3.37. Occurance of scuffing vs. bleed for different greases.

3.7 Rolling and Sliding Wear with Standard Tests

3.7.1 Sliding Contact (SRV)

As a general understanding of the behavior of the greases and the effect of their composition on their protection properties under sliding contact with high to low amplitudes was obtained, the same was not achievable for rolling contact wear because of problems that occurred with the Modified Fafnir. Due to limited access to the SRV (TE-

77) and standard Fafnir, grease samples were selectively chosen for testing with these rigs. LiX-100 CaS-100 and PU greases that have close viscosities were chosen to be able to make comparisons and to have bleed rate in a practically acceptable ranges the greases with 10% and 30% added oil to the baseline were selected for the experiments.

73 Figure 3.38 and Figure 3.39 show the normalized wear area of the SRV sliding contact tests done on the mentioned grease samples drawn against oil content and bleed rate respectively.

It can be seen that increase in bleed rate for polyurea, calcium sulfonate and lithium complex grease had minimal change in fretting wear. Furthermore, CaS and LiX grease showed superior fretting wear resistance in comparison with PU grease.

4

3.5 LiX-SRV CS-SRV 3 PU-SRV 2.5

2

1.5

1

Normalized Wear Area Wear Normalized 0.5

0 60 65 70 75 80 85 90 Oil Content (%)

Figure ‎3.38. Standard SRV (TE-77) sliding wear test results of three grease types chosen with close base oil viscosities

(LiX-100, CaS-100 and PU) drawn against the oil content for each grease mixture.

74

4

3.5

3

2.5

2 LiX-SRV 1.5 CS-SRV 1

Normalized Wear Area Wear Normalized PU-SRV 0.5

0 0 1 2 3 4 5 6 7 8 Bleed Rate

Figure ‎3.39. Standard SRV (TE-77) sliding wear test results of three grease types chosen with close base oil viscosities

(LiX-100, CaS-100 and PU) drawn against the oil bleed rate for each grease mixture.

Once again it is confirmed that the sliding wear is not affected by the bleed rate or the oil content of the grease. However, the effect of the thickener can be easily seen. Unlike the

HFRR that was equipped with quantitative wear volume measurement, the SRV measures the normalized wear area which may not be as accurate. While the tests do not differentiate between the performance of lithium complex and calcium sulfonate thickened greases, the polyurea grease shows significantly lower protection properties in sliding.

3.7.2 Rolling Contact Fretting- False Brinelling (Fafnir)

The results of the rolling fretting performance of the grease samples done on the Fafnir test are shown in Figure 3.40 and Figure 3.41. As for the rolling contact, also similar to

75 what was seen before in the preliminary experiments with Fafnir and Modified Fafnir in section 3.3.3, the dependence of rolling contact fretting on bleed rate is considerable.

However, now that the effect of thickener type is also being shown, it can be seen that the thickener type has an even more pronounced effect than the bleed rate. This is suggested by the fact that the CaS grease with zero bleed, once again performs significantly superior than the other grease types, showing almost no wear, whereas other grease types would perform as well as CaS only in much higher bleed rates.

It can be seen that the increase in bleed rate for polyurea grease has had significant reduction (>25 times) in false brinelling wear. Calcium sulfonate grease showed excellent false brinelling resistance and showed further improvement (>20 times) with increase in oil content. Lithium complex grease showed some reduction (<2 times) in false brinelling wear with increase in bleed rate, however, the critical 5% bleed rate threshold was not met within 30wt% added oil to the mixture. Based in these observation so far, it can be said that the bleed properties are critical to false brinelling performance.

Another interesting point here is, by comparing the LiX and PU greases in both graphs, it is seen that in a given oil content of about 87%, the PU brings about zero rolling wear while LiX sample with similar oil content shows a significant amount of wear/oxidation.

This can be immediately explained considering the next graph which shows the much higher bleed rate of the PU grease. In a given bleed rate, however, it can be seen that the

LiX would provide much better protection. These effects are described more clearly following the revelation of the grease structure and their unique features through the proposed mode in next section.

76

30

LiX-Fafnir 25 CS-Fafnir

20 PU-Fafnir

15

10

5

False Brinelling Bearing Wear (mg) Wear Bearing BrinellingFalse 0 60 65 70 75 80 85 90 Oil Content (%)

Figure ‎3.40. Standard rolling contact Fafnir Oxidation test results of three grease types chosen with close base oil

viscosities (LiX-100, CaS-100 and PU) drawn against the oil content for each grease mix.

30

LiX-Fafnir 25 CS-Fafnir PU-Fafnir 20

15

10

5 False Brinelling BearingWear (mg) BearingWear BrinellingFalse 0 0 1 2 3 4 5 6 7 8 Bleed Rate (%)

Figure ‎3.41. Standard rolling contact Fafnir Oxidation test results of three grease types chosen with close base oil

viscosities (LiX-100, CaS-100 and PU) drawn against the oil bleed rate for each grease mixture.

77

3.8 Boron Based Additives and WC/a-C:H Coating

3.8.1 Overview

A thorough discussion about fretting, scuffing and interaction of surfaces and lubricants cannot be done without considering the effects of additives and coatings. Additives are responsible for facilitating interactions of the lubricant and the tribological surfaces and therefore have a widespread application in lubrication industry. On the other hand, surface engineering and coating techniques in particular are considered one of the most common techniques for enhancing the mechanical properties of the surfaces while retaining other desired properties in the bulk material (ref). Therefore, some introductory experiments on this subject will be hereby discussed. Although these experiments are by no means intended to provide a complete account on these matters, they are intended to touch on a fundamental question that were pursued in this study; on the difference between rolling and sliding contacts in interacting with grease. One of the main hypothesis in this work was that while the rolling contact is more affected by the bleed rate, the sliding contact is affected by the surface interactions of the grease, such as is involved with additives and coatings. In previous sections, the first part of this hypothesis was established by showing that the bleed rate of the grease has a major impact on its rolling wear mitigation properties while it did not have a significant effect on its sliding contact performance. Here, the test results will be shown to be in agreement with the second part of the hypothesis; on the dependence of the sliding contact on the surface interactions with grease.

78 3.8.2 The Effect of Boron Additives on Steel-on-Steel Contact

The intensity of the scuffing on HFRR ball and disk in testing 460 LiX grease and the effects of three different boron based additives in different concentrations are shown in

Figure 3.42. The HFRR parameters have been chosen to produce severe conditions with comparatively high Hertzian pressure of 1.8 Gpa and lowest fretting amplitude employed in this research (e=0.4). Not surprisingly, the samples were scuffed without exception and the scuff intensity measured on the ball and disks did not show a significant effect of the additives as can be seen in Figure 3.42. It can be said that the conditions are severe enough in a very low amplitude of sliding that would not allow time for heat to dissipate away from contact, resulting into occurrence of scuffing on all of the experiments.

4500 4000 3500 3000 2500 2000 1500 1000 Scuff Intensity (µm^3) Scuff 500 0 ISO 1% 3% 5% 1% 3% 5% 1% 3% 5% 460 I-5 I-5 I-5 BA BA BA TIPBTIPBTIPB

79 3500

3000

2500

2000

1500

1000

Scuff Intensity (µm^3) Intensity Scuff 500

0 ISO 1% 3% 5% 1% 3% 5% 1% 3% 5% 460 I-5 I-5 I-5 BA BA BA TIPB TIPB TIPB

Figure ‎3.42 Intensity of scuffing on HFRR ball (a) and disk (b) in testing LiC#3 and that mixed with three different

boron based additives in 1, 3 and 5 wt% concentrations.

Despite the severity of the contact, the TIPB shows a slight reduction in the scuff intensity with increasing the concentration to 5%. One reason that the other two types have not shown a significant improvement maybe due to the fact that they were in powder form and much of them had remained undissolved and in powder form, which is an undesirable condition for the grease and is reasonable to assume that it would interfere with the oil flow. Although it could be argued that this could have minor effect on the performance of the grease due to the small dimension of the contact compared to the dimensions of the undissolved additive particles.

80

3.8.3 The Effect of Boron Additives on Steel-on-WC/a-C:H Contact

The results of similar testing conditions for the grease additized with TIPB under a Steel on WC/a-C:H Contact can be seen in Figure 3.43. Here, the effect of additive interactions with the coated surface can be shown. To further verify this result the test has been extended to 120K cycles and the results are shown in Figure 3.44. It can be seen that scuffing reduces in intensity with increasing TIPB content and stops at 5%TIPB. This result is consistent with reports in literature [63,66] that scuffing can be prevented by the use of WC/a-C:H for commercial grease with different additives.

Another observation here is that the softer steel ball shows smaller scuff intensity than the coated surface which is harder. This is important and rather counter intuitive because the damage is expected to be more severe in steel for ordinary circumstances. This implies a different underlying mechanism for the scuffing of coating which may include certain decomposition reactions.

8000 14000

7000 12000 6000 10000 5000 8000 4000 6000 3000 2000 4000

1000 (µ^3) ScuffIntensity 2000 Scuff Intensity (µ^3) ScuffIntensity 0 0 460 + 1% 460 + 3% 460 + 5% 460 + 1% 460 + 3% 460 + 5% TIPB TIPB TIPB TIPB TIPB TIPB

(a) (b)

Figure ‎3.43 Intensity of scuffing on HFRR of 30K cycles with contact composed of steel ball (a) on WC/a-C:H disk (b)

in testing LiC#3 mixed with TIPB in 1, 3 and 5 wt% concentrations.

81

12000 20000

10000 16000 8000 12000 6000 8000 4000

2000 4000 Scuff Intensity (µ^3) ScuffIntensity

0 (µ^3) ScuffIntensity 0 460 + 1% 460 + 3% 460 + 5% 460 + 1% 460 + 3% 460 + 5% TIPB TIPB TIPB TIPB TIPB TIPB

(a) (b)

Figure ‎3.44 Intensity of scuffing on HFRR of 120K cycles with contact composed of steel ball (a) on WC/a-C:H disk

(b) in testing LiC#3 mixed with TIPB in 1, 3 and 5 wt% concentrations.

3.8.4 Boron Additives and DLC Coatings in Rolling Contact

Now, it is intended to see if the addition of TIPB would have a beneficial effect on the rolling contact fretting as well. Due to the limited access to the Modified Fafnir, only the main comparison of the 5 wt% TIPB added 460 grease was made. As can be seen in

Figure 3.45 the addition of TIPB to the grease does not provide any improvement in rolling fretting performance over the baseline grease.

82

Figure ‎3.45 Mass loss of bearing raceways after Fafnir test with baseline LiC#3 grease and that mixed with 5% TIPB.

83

CHAPTER IV

4 MODELING

4.1 Static Bleed Based on Permeability of the Thickener Structure.

4.1.1 Background

As mentioned in section 3.1.1 an unexplained observation was made with the bleed behavior of the calcium sulfonate thickened grease. While the grease is expected to have a gradual increase in bleed rate with increase in its oil content, the calcium sulfonate has an unusual behavior in which the grease maintains zero or little bleed even after a considerable amount of oil is added to it and then suddenly increases with an unusually high steepness. It is hypothesized here that this behavior may be related to the geometry of the thickener particles (see section 2.4).

With this introduction, a model for the oil flow behavior is developed as follows.

4.1.2 Permeability Model

For modeling the bleed behavior based on the permeability of porous structures made of thickeners in general (i.e. not only spherical), as the initial step, an essential assumption needs to be made in an attempt to obtain a more realistic interpretation. We assume that the oil in a certain space surrounding the thickener particles is completely immobilized.

Therefore, in our model, we assume the thickener particles to be made of spheres and the

84 radius of these spheres are assumed to be what we call the “effective radius,” , that are larger than the actual radius of the thickener particles . The spheres of radius will be called “effective spheres” constituting the “effective porous medium”, inside of which (=outside the effective spheres) the oil may be free to have Darcian flow. The ratio of the effective radius to the actual radius will be called the “effective ratio” :

The conceptual reasoning behind this assumption comes from an understanding of the physical structure of actual thickener particles. Firstly, considering the large distance between the thickener particles (as seen in the AFM images of the thickener structures shown in Figure 1.5 and also based on the data obtained in section 3.2) makes us wonder how have the thickener structure been successfully modeled using Darcy’s law where the thickener particles, contributing only to a minor part of the total mass of the grease, act as a porous matrix (as opposed to the oil being the matrix)? Secondly, considering the spherical shape of calcium sulfonate particles and the fact that they are diluted in the oil, it is virtually impossible for an array of separate spheres to form a structure throughout the body of the grease. As explained before, such structure is supposed to exist to provide the consistency of the grease. However, if we take effective thickener structure all of the mentioned problems will be resolved as will be discussed shortly and a more precise description of the grease structure will be achieved.

Now that the dimensions of the particles have been determined, the thickener structure can be assumed to be a three-dimensional array of spheres with a cubic lattice spaced at the distance from one another as depicted in Figure 4.1.

85

Figure ‎4.1 The thickener particle lattice with particles of actual radius r, effective radius r’ and spacing a.

The distance is determined by the amount of oil in the grease. The higher the oil content the more separated the spheres and the more porous the thickener structure.

Therefore, the real porosity of the grease , which is equal to the voidage or the volumetric void fraction is determined by the volume fraction of oil in the grease

related to the parameters of the lattice as:

[ ( ) ]

where is the volume of the lattice and is the volume of the spherical thickeners. The

will be related to the oil mass fraction as:

Where and are respectively, density and mass of oil or grease depending on their subscript. As defined before, the modeled porous structure is based on the effective

86 media. Therefore the effective porosity of the thickener structure based on the effective volume of the spheres is calculated as:

[ ( ) ]

assuming that the effective spheres do not overlap one another. The fraction which is proportional to by , can be considered as a measure of the dilution of the thickener particles. From now on the effective porosity is referred to as porosity or .

Thanks to the concept of the effective radius we can now assume zero permeability vis-à- vis bleed for the model structure, analogous to the first few data points in experimental results of the calcium sulfonate bleed seen in Figure 3.2. In values smaller than √ all of the oil will be contained in the effective medium. In this case, the particles are spaced so close to each other that the effective spheres for each particle overlaps those of all of its 26 neighbors. At √ the surfaces of the effective spheres meet at the center of the unit cell. With further dilution from this point, the effective spheres will lose contact with 8 of its farthest overlapping neighbors, which will result in the formation of isolated pores at the center of the unit cell as shown in Figure 4.2.

Figure ‎4.2 Isolated pores for √

87

Whether the oil is trapped inside the effective medium or completely surrounded by it, it will not be able to leave the structure, and therefore the bleed rate will be equal to zero.

This situation will continue to hold until the spheres are diluted to the point where the surfaces of the effective spheres meet at the center of the faces of the lattice at √

(Figure 4.3). Therefore, if the concentration of the thickener is high enough to have values smaller than √ the bleed of the grease will equal zero.

Figure ‎4.3 Point of transition to a permeable structure at √ .

Dilution beyond √ will separate each effective sphere with 12 more of its overlapping neighbors. This will open passages at the center of each face of the lattice as shown in Figure 4.4. This results into the connection of the pores forming a permeable structure with inter-related pores throughout the grease structure. Such grease is capable of bleeding as the oil outside the effective media is able to have a Darcian flow through channels formed in the structure such as is shown in Figure 4.5.

88

Figure ‎4.4 Formation of channels for √

Figure ‎4.5 The shape and geometry of the pores.

To determine the bleed in the as-described structure, an approach similar to that of

Carman-Kozeny solution [109] for describing laminar flow through a packed bed of particles was taken. For this, we come back to the Darcy’s law explained before. The

Darcy law could be rearranged in terms of the pressure gradient as [39]:

89

Where is superficial fluid velocity through the porous media, is the frictional pressure drop across the length of , is the fluid viscosity, and is the permeability

(same as it was defined before).

To be more descriptive of the permeability of the porous media regarding its geometrical characteristics, we analyze the flow in terms of the fluid flow through a set of tubes. The starting point for such analysis is the Hagen-Poiseuille equation [109] for laminar flow through a tube:

Where is the tube diameter. Considering the pores as tubes with an equivalent diameter

through which the fluid follows the tortuous path of equivalent length with a velocity then the above equation turns into:

The values of , and should be calculated according to the geometry of the porous structure. To calculate the velocity one must consider the actual cross-sectional area of flow (shown in Figure 4.4). If the fraction of the cross-sectional area available for flow is assumed to be equal to porosity, then being the actual velocity of fluid through the pores is related to the superficial fluid velocity by:

where is the apparent cross-sectional area of bleed.

90

It should be noted that the value of is different from the one calculated in equation 10 due to the fact that the effective spheres are still overlapping all 6 of their closest neighbors. Two neighbor spheres spaced at overlap one another in a volume consisted of two spherical caps. The volume of the effective media and pores within the unit cell will therefore be calculated by:

where is the volume of the spherical cap which itself is calculated using the equation:

where and are the diameter and height of the cap respectively and are calculated by:

√

With the calculated values of the effective media, the porosity of such structure can be obtained using:

91

Now, the tortuous path of the fluid inside such a pore ( ) should be calculated. An estimate of can be obtained by visualizing a cross-section of the lattice (Figure 4.6) as:

where is the same value as defined before, contributing to the radius of the overlapping spherical caps. Assuming is proportional to the actual length it can be written as:

Where is a parameter called tortuosity.

Figure ‎4.6 The tortuous path of fluid inside the pores.

92

The tube equivalent diameter can be determined as:

where flow area as noted before and wetted perimeter in which is the particle surface area (effective) per unit volume of the grease that can be calculated as:

Where is the surface area within the pores that can be calculated (analogous to the calculation of ):

where is the area of spheres ( and is the area of the overlapping spherical caps ( )

can now be calculated as:

By substituting the calculated formulas for , and equation 13 becomes:

93

Moreover, therefore the permeability of the porous structure will be calculated by:

Where is the combination of all the constant values in the equation. This description of permeability will be valid for dilutions with . Further dilution beyond

will totally separate the effective spheres and the flow condition starts to change drastically. For the sake of future discussions, this point in the range of dilution will be called “the point of structural opening”. In such condition, the value of the permeability maybe estimated more easily using Carman-Kozeney equations for a packed bed of spheres [110] as:

where is the so called Kozeney constant.

4.2 Dynamic Bleed

4.2.1 Overview

In the previous section, the static bleed, meaning the kind that happens in the standard bleed rate tests, was modeled. However, in the tribological contact the oil release

94 mechanism of the grease could be very different. It was seen in the results section that under sliding contact the protective properties of greases did not show any kind of dependence on or correlation with their measured bleed rate. The oil is responsible for lubricity, and therefore its release from the thickener structure should happen one way or another within the contact to enable lubrication. Therefore, the fact that no dependency on the standard bleed rate is seen in certain conditions is a confirmation that the underlying oil release or “bleed” mechanism in those conditions is substantially different from that in the standard bleed testing. In the model proposed in the last section, a couple of assumptions were made that might help with understanding the dynamic bleed as well.

The first assumption was the existence of the “effective media” that constitutes the effective thickener structure beyond the thickeners themselves. The second assumption which is rather a more realistic description of the thickener structure is that while the effective media surrounding the thickener particles builds up a mechanically stable structure, the thickener particles are separate particles suspended in a matrix of oil and should be treated as such when interacting with external forces. This essential understanding of the grease structure that was gained and confirmed through the modeling of the static bleed can help with the understanding of “dynamic bleed.” With certain modifications and assumptions, the difference in static and dynamic bleed will be discussed first. Furthermore, we will attempt to describe the different features of dynamic bleed in rolling and sliding contact. In the following modeling of the dynamic bleed, there was no particular measured quantity for building correlations as far as the flow of oil is concerned. Therefore, the uses of oil flow equations are avoided, and the models explain the overall qualitative damage results.

95 4.2.2 Interactions of the Thickener Particles with the Advancing Body

While bleed cannot be easily described in any way other than the viscous flow of oil in the effective thickener structure, for the case of dynamic bleed, the effect of the force applied by the moving body (in the contact) on the thickener structure should also be considered. Based on that, the following step is to assess the effect of the mentioned structural change in the oil flow to be able to describe the dynamic bleed. When a body slides or rolls through the thickener structure, it applies a force on the thickener particles to push them away from the contact. Based on the modeling in the previous section, as a thickener particle moves, its surrounding effective medium is also expected to be dragged with it. With that being said, the following effects are expected for dynamic bleed.

The Thickener Particle Transport Effect and Particle Accumulation

The “transport” effect is concerned about the number or fraction of thickener particles that are dragged along the moving contact. In other words, when the contact encounters some thickener particles, what fraction of those particles are affected to the maximum extent of the motion? The first layer of particles that are transported by the moving body would then transport the next layer of thickener particles and so on. This sequence would lead to the accumulation of thickener particles. The transport effect depends mostly on the geometry of the thickeners. For obvious reasons, the higher the aspect ratio of the thickener particle the higher the particle transport and consequently the accumulation effect. For example, for the case of spherical thickener particles, the moving body would mostly slip through the particles and a small number of the particles, if any, would be transported along with the contact. Therefore, the spherical thickener particles with the

96 lowest aspect ratio would experience the minimum transport effect. On the other hand, the rod-like thickener particles with the highest aspect ratio under this study are very likely to be transported with the contact and layers of the particles would accumulate in the way of the moving contact. It should be noted that if the thickener particles are too diluted, the particle transport effects could be similar or close but the accumulation effect would not be the same and may be neglected. This is because if the thickener particles were too separated from each other, the particles would have time to rotate and reposition themselves to avoid transport or accumulation before meeting the next particle.

The Structural Squeeze Effect and Gradual Clogging

The most obvious effect of the force applied by the advancing body onto on a set of transported thickener particles is that it will presumably squeeze the structure by pushing the first layer confronted by the surface towards the rest of the structure. This is depicted in the schematic drawing Figure 4.7 in which the thickener particles are taken to be an ordered lattice of spheres in two dimensions for simplicity, and it is assumed that two spheres (from a single layer) are being transported by the contact. If the body continues to advance further in that direction, the first layer of particles will start to apply force on the next layer and so on. While the effects of the so-called “squeeze flow” of oil through the closing structure will contribute to the total flow rate, the fact that the particles are coming closer together could make the pores within the effective structure narrower. In this case the squeeze effect would lead to the reduction of total oil flow rate or dynamic bleed to an eventual clogging.

97

Figure ‎4.7 Interactions of the advancing body with a layer of thickener particles causing squeeze flow towards the

contact (yellow arrow).

The Thickener Particle Rotation Effect and Immediate Clogging

An essential effect on transported particles that depends highly on their shape is particle rotation. One way to describe this effect is to consider the porous structure of the effective medium surrounding just one thickener particle and calculate the permeability of the resulting structure accordingly. It is assumed for simplicity that the effective media of other randomly oriented particles have created a tube. This structure will be called the equivalent oil flow duct, which is depicted in Figure 4.8. The size of such duct is in that case determined by the thickener to oil volume ratio, having to be equal to the ratio of the volume of the thickener particle to the volume of the tube (equation 3). The rotation of the thickener particle when it meets the advancing body will now be discussed as shown in Figure 4.9. As the advancing surface meets some section of a randomly oriented thickener particle, a force will be applied to that point. While it can also be assumed that

98 the same force in the opposite direction will be applied to a point on the opposite side of the particle being exerted from another particle behind it or simply from the equivalent force of the resistance forces of the liquid. This is shown in the figure as one large block of medium next to the particle exerting the opposite force. The balance on the forces would eventually demand rotation of the thickener particle in a plane that is shared by the vector of the exerted force and the normal of the broadest surface of the particle. If the motion continues, the rotation of the thickener particle also continues until the broadest surface of the particle is perpendicular to the direction of the motion of the advancing body. This is schematically shown in Figure 4.10 for the case of a plate-like thickener particle, which would show the maximum of rotational effects as will be discussed shortly. From that point forward, any further motion would only increase the structural squeeze and thickener accumulation effect.

Figure ‎4.8 Analyzing particle geometry effects by estimating flow surrounding a single effective thickener particle.

99

Figure ‎4.9 Rotation of a single thickener particle (blue) stuck between the advancing contact (gray) and another thickener particle or the remaining thickener structure (green) opposite of the direction of the motion (purple arrow) in

response to the forces (yellow arrows) applied by the advancing body and the remaining thickener structure.

Figure ‎4.10 Total clogging of the equivalent oil flow duct by the rotation of a plate-like effective thickener particle.

After establishing the rotation of thickener particles in the contact, it is time to analyze the effect of the orientation of a single thickener particle on the oil flow in the duct surrounding the effective thickener. It will be shown that the orientation of the thickener particle could have a substantial effect on the geometry and the opening size of the pores around it or the permeability of the equivalent oil flow duct. Let’s consider the area of the

100 cross-section of the effective thickener at the middle of the equivalent oil flow duct (AT) in Figure 4.8. The thickener particle is, at first, positioned in a way that the broadest surface of the particle is parallel to the axis of the duct and the surrounding void is at its maximum value (AC). With the rotation of the particle AT would increase and the area of the void would decrease. Looking at another cross-section of the thickener particle, this time along the axis of the duct as seen in Figure 4.11, and limiting the rotation of the particle in one plane for simplicity, AT can be calculated as a function of the displacement of the moving body (x). Therefore, AT would be equal to tXtR and tXtP for a rod-like and plate-like thickener particle respectively, where tR is the diameter of the rod, tP is the width of the plate and tX is calculated by:

where L is the maximum value of x, which is achieved when the normal of the broadest surface of the thickener particle is parallel to the direction of the motion, and no further rotation would occur. Figure 4.12 shows the cross-sectional area of the equivalent oil flow duct (Ap for plates, Ar for rods and As for spheres) with respect to the orientation of the thickener particle as a function of the displacement of the advancing body for a spherical, rod-shaped and plate-like thickener particle. A spherical particle has zero rotational effect since it does not change under rotation. On the other hand, a platelet, which has a high surface area, would strongly affect the porosity if rotated. It can be shown that even minor rotation of a plate could result in a significant blockage of the tube, which will be referred to as clogging.

101

It was shown that the orientation of a single thickener particle with respect to the position of the tube determines the permeability of the tube, and that is primarily affected by the thickener particle geometry. An important implication of this model is that the higher the area to volume ratio of the thickener particle, the higher is the effect of rotation on the permeability of the resulting structure. Spheres have the minimum possible area to volume ratio and have zero rotational effect. While the plate-like thickener particles have an extremely high sensitivity to rotation and could easily cause clogging.

Figure ‎4.11 Cross-section thickness of a particle (tx) located at the center of the equivalent oil flow duct showing

increase with rotation due to motion of the advancing body.

102

Figure ‎4.12 The change in cross-sectional area of the equivalent oil flow duct with displacement normalized with the

length of each particle.

4.2.3 Dynamic Bleed in Rolling vs. Sliding Contact

The Action of the Advancing Body with Respect to the Direction of Motion

The most significant difference between the body/thickener interactions in rolling and sliding conditions that is possibly responsible for the difference in their bleed mechanism is the fact that the advancing body that causes the displacement of the thickener particles acts in two different directions for the rolling and sliding contact. For sliding contact, the advancing body applies force to the thickener particles in the same direction as the motion of the sliding body. For rolling contact, however, the advancing body applies force on the thickener particles, mostly in the direction perpendicular to the direction of

103 the motion of the spherical roller pushing the particles away from the middle of the contact to the sides (Figure 4.13). This difference would be responsible for other effects seen in results and will be discussed further in the following sections.

The Difference in the Magnitude of Displacement

For the sliding contact, in determining the magnitude for each of the thickener structure squeeze effect and the thickener particle rotation effect, the displacement can be taken as infinity and the mentioned effects to be at their maximum amount. However, as can be seen in Figure 4.14 for the spherical roller the magnitude of the displacement, which is in the direction perpendicular to its motion, is in the same order of magnitude as the dimensions of the particle itself. Considering this, both the squeeze effect and the particle rotation effect would be at low to moderate levels and only to the limited extent of the displacement magnitude. This results in the dynamic bleed under the rolling condition with spherical rollers to be somewhat close to their static bleed depending on the geometry and size of the thickener particles, which is very small. This could explain the reason behind the fact that grease protection performance under rolling contact for most of the greases under study to be related to their standard bleed rate which is basically the static bleed.

104

Figure ‎4.13 The direction of the motion of the thicker particles (purple arrows) and flow of oil (yellow arrows) with

respect to the direction of the motion of the ball (gray arrows) for both rolling (top picture) and sliding (bottom

picture).

105

Figure ‎4.14 The two-dimensional schematic drawing, showing the magnitude displacement of the thickener particle

under rolling contact.

106

CHAPTER V

5 DISCUSSION

5.1 Different Bleed Mechanisms

An important understanding obtained from the proposed models, which will enable us to explain many of the observed phenomena, is the fact that the oil release mechanism or bleed in the static condition (i.e. the bleed rate test) is different from that under rolling and sliding conditions. It was seen in the results that the performance of the greases under sliding fretting showed no dependency on the grease bleed rate while under rolling some correlation was found as long as the thickener type was not changed. As discussed before, while it is evident that fretting under rolling and sliding are two different phenomena, they are often both referred to as “fretting” which is not correct. Here we emphasize that like fretting, the term “bleed” had also been previously loosely defined and used to describe phenomena that could be very different. In this study, at least three different types of bleed mechanisms are discussed, static bleed, dynamic bleed under sliding and dynamic bleed under rolling. This terminology will be used for the discussions on the experimental observations as follows.

5.2 Experimental and Model Fit for Static Bleed

Figure 5.1 shows the permeability values calculated from the experimental results shown before in Figure 3.1, presented along with the model results with a good fit.

107 The permeability values were calculated from the bleed rate results using Darcy’s law as explained in section 1.1.9. By comparing the experimental and model results, the otherwise unexplained bleed behavior of the calcium sulfonate thickened greases can be contributed to the spherical geometry of its particles. The sudden rise in the bleed is explained by the “structural opening” explained in section 4.1.2, which happens at 30% added oil.

0.7

0.6 K (Experiment) K (Model)

0.5

0.4

m^2) Structural Opening μ

K( 0.3

0.2

0.1

0 0 0.2 0.4 0.6 Added Oil Ratio

Figure 5.1 Permeability of calcium sulfonate grease model vs. experimental result.

It should be noted that in Figure 5.1 the mixture with 30 percent added oil, according to the model, falls in the first bleed regime which for the case of CaS-100 has resulted in overall zero percent bleed due to the narrow range for this regime. To widen this range, the CaS-Mix was developed. In these mixtures a calcium sulfonate grease with 75 cSt base oil viscosity was mixed with a 485 cSt oil. This way as the structure is expanding, the viscosity of the oil mix is also increasing resulting in to a wider range for the

108 transitional flow regime where bleed rates are limited. As can be seen in Figure 5.2 the structural opening for this case happens at 50% added-oil which shows a bleed rate higher than zero. The discontinuity in the fitted bleed rate from the model is due to the bleed regime change at 50% added oil.

0.45 0.4 K (Model) 0.35 K (Experiment) 0.3

0.25

m^2) 0.2 μ

K ( K 0.15

0.1

0.05

0 0 0.2 0.4 0.6 0.8 -0.05 Added Oil Ratio

Figure ‎5.2. The model and experimental results for the permeability of the thickener structure vs. the ratio of added oil

to the CaS-Mix grease.

A similar approach for cylindrical or rod-like thickener particles has been taken.

Supposedly, the actual structure of the lithium complex grease thickeners has a random distribution of fibers. Other researchers have used Kozeney-type solutions to model the permeability of a packed bed of randomly distributed cylinders[111] as:

109

Where is the Kozeney Constant. A comparison between the experimental results of the lithium complex grease with this model is shown in Figure 5.3.

0.3

K (Experimental) 0.25 K (Model) 0.2 m^2) μ 0.15 K(

0.1

0.05

0 0 0.2 0.4 0.6 0.8

Added Oil Ratio

Figure 5.3. Permeability of lithium complex (LiX-100) grease model vs. experimental

result.

The reason for the more linear behavior of the lithium complex thickened greases is explained by considering the point of structural opening. To understand the point of structural opening for cylindrical thickener particles, a unit cell is constructed in which the fibers are assumed to be orthogonally arranged for simplicity. An important conclusion from such modeling is that the point of structural opening for cylinders would

be proportional to the length of the fibers . In cylinders, which have very high aspect ratios, the fiber length is appreciably larger than the effective radius

110 (which is proportional to the point of structural opening for the case of spheres).

Therefore the opening of the thickener structure does not happen for lithium complex grease with applied ranges of dilution. That is why the increase in bleed with oil content has been gradual and no sharp increase such as seen for calcium sulfonates has been observed for lithium complex thickened grease.

5.3 Overall Comparison of the Static Bleed of Different Thickener Types

Figure 5.4 shows the same bleed measurements that were shown before in Figure 3.1, this time, drawn against the total oil content of the grease mixture rather than the added oil to the baseline grease. As can be seen, the amount of calcium sulfonate thickener particles used is much larger than that of polyurea and lithium complex that have higher aspect ratios. This means that the range of dilution for lithium complex is even higher than that of calcium sulfonate grease and yet it does not surpass the point of structural opening.

The higher the aspect ratio, the further away is the point of structural opening, and the less thickener would be needed to yield the desired structurally-dependent grease properties.

111 55

Polyurea 45 Calcium Sulfonate

35 Lithium Complex 25

15 Cone Bleed Bleed (%) Cone

5

60 70 80 90 -5 Percent Oil in Grease (%)

Figure 5.4. Bleed vs. grease oil content for different thickener types.

5.4 Rolling vs. Sliding Contact Fretting and their Relationship with Bleed Rate

As was discussed in section 4.2 the remarkably different relationship between rolling and sliding fretting and the static bleed rate of the grease, is caused mainly by the fact that the direction of the forces and motion on thickener particles are different in the two cases, resulting in very different displacement amplitudes. The displacement amplitude is practically infinity for sliding contact, where the particle motion is parallel to the direction of the motion of the contact, resulting in maximum dynamic effects, namely, particle transport and accumulation effects, structural squeezing effects, thickener particle rotational effects and consequently clogging. This, results in the sliding dynamic bleed behavior of the grease to be very different from its static bleed. That is why no relationship is found between the fretting wear performance of the grease under sliding contact and standard bleed rate of the grease as seen in the results sections 3.4, 3.5

112 and 3.7.1. However, for rolling contact, the motion of the particles is to the sides, resulting in low displacement amplitudes. Therefore, the effects specific to the dynamic bleed will be minimal, and the dynamic bleed behavior will be comparable to that of static bleed. This explains the dependence of the rolling fretting performance results seen in sections 3.3.3 and 3.7.2 on the standard bleed rates of the grease samples.

It should be noted, that this dependence only holds as long as there is no change in grease types. In other words, when one is comparing rolling fretting performance of greases with same thickener type, the results show dependence on the bleed rates of the samples, whereas, if the same is compared for different grease types, no correlation could be found between rolling fretting performance results and the grease bleed rate. Another exception for the dependence of rolling bleed on static bleed rate is for the case of polyurea thickened greases that have plate-like thickener particles. As discussed in section 4.2.3 the displacement for rolling contact is in the same order of magnitude as the thickener particle size which is not large, but according to the model as shown in Figure 4.12 even that much of displacement could have a substantial rotational effect for a plate-like thickener, noting that the horizontal axis in this figure is normalized to the size of the particle. The effect of particle geometries will be discussed further in the following section.

5.5 Different Grease Thickeners Under Rolling Contact

Now that that the motion of particles in rolling dynamic bleed has been established, it is time to explain specific behaviors observed for each thickener type. Firstly, the general trends in the rolling fretting results of sections 3.3.3 and 3.7.2 meaning the fact that the

113 rolling fretting performance of the grease improves with the addition of oil accompanied by the increase in static bleed rate could be attributed to the similarity of the oil release mechanisms in the rolling bleed and the static bleed. This begs the question, then why is it that the correlation of rolling fretting damage and the grease bleed rate does not hold when changing from one grease type to another? (See results in section 3.7.2) In other words, why the calcium sulfonate grease with practically zero bleed rate performs so much better than the other two grease types? And why does the polyurea grease perform worse than a lithium complex mix with similar or even lower bleed rate?

These questions can be answered by considering the geometry of the particles and their relationship with the dynamic bleed effects such as thickener rotation and structural squeezing and thickener transport. For example, the thickener rotational and transport effects that are usually the most consequential for the bleed rate is virtually zero for the spherical calcium sulfonate grease thickener particles. This quickly explains the superior performance of this grease, not only under rolling but in all of the tested conditions. On the other hand, for the case of the polyurea, the difference arises from the strong dependence of the thickener rotational effects on displacement. Although the displacement amplitude for rolling contact is assumed to be close to the dimensions of the thickeners, the same displacement is enough to render the bleed condition from normal to complete blockage or clogging. Needless to say that for polyurea grease the same holds for other contact conditions as well.

With this understanding, and going back to the discussion of the strong dependence of the rolling fretting on the static bleed rate, the wear-drop-threshold (5% in the case of lithium

114 complex) is probably the limit in which the rotational and squeeze effects approach zero due to the expansion of the thickener structure arising not from the increase of static bleed rate itself but rather the increase in oil content which also happens to result in the increase in bleed rate to the certain “threshold” amount. In other words, although the relationship between rolling bleed and static bleed is established, the threshold at which zero fretting is achieved does not necessarily have anything to do with the bleed rate and the effect could be the result of dilution which seperates the particles and eliminates most of the dynamic bleed effects.

5.6 Grease Thickeners Under Sliding Contact and their Effect on Scuffing

Most of what is to be said about the explanation of the behavior of different grease thickener types under the sliding condition, namely the exceptional behavior of spherical calcium sulfonate thickener particles and the adverse effects of the plate-like polyurea thickeners have already been discussed sufficiently in the previous sections. However, coming back to the results shown in Figure 3.37, some dependence was seen between the scuffing performance and the bleed rate only for the lithium complex grease, which requires more discussion. Furthermore, the general scuffing performance of the grease samples could also be attributed to the geometry of the thickener particles. Although the intrinsic EP and antiwear properties of the calcium sulfonate as a thickener should not be neglected, the immunity of the grease against starvation and subsequent scuffing may simply be attributed to the spherical geometry of its thickener particles which allows the advancing surface to slide through the thickener particles without causing any thickener transport and accumulation that could result in clogging and starvation. The same would not happen for the lithium complex greases with cylindrical particles which would be

115 dragged with the contacting body resulting into further entanglement which accompanied by the rotational effects results in clogging. For the lithium complex greases, scuffing reduces with increasing oil. This, again, may not be an effect of increasing bleed rate but rather the dilution of the thickener particles and the expansion of their structure that would reduce the thickener transport and clogging effects. The scuffing behavior of the polyurea-thickened greases may again be attributed to the higher surface to volume ratio of the thickener particles. The plate-like polyurea particles have higher thickener rotational effects and therefore are more prone to clogging and blocking the oil flow. In other words, clogging could again be the result of rotation of the platelets by the force applied by the motion of the contacting body. .

116 CHAPTER VI

6 CONCLUSION

The main objective of this study was to understand the bleed or the oil release mechanism of different types of greases with a focus on their effects on fretting. Another goal was to acknowledge and explain the difference between rolling and sliding fretting. To this end, the bleed and consistency of grease samples with lithium complex, calcium sulfonate and polyurea thickeners and varying oil contents to achieve different bleed rates were studied.

This was followed by testing the grease mixtures for their protection properties under rolling contact fretting or false brinelling as well as sliding contact fretting with different loads and displacement amplitude. The most important experimental result, firstly, was that there was no dependence of sliding fretting on the grease bleed rate and a strong dependence for rolling fretting. Secondly, different grease types showed unique bleed behavior and performance under different tribological contacts. Specifically, the calcium sulfonate grease had an exceptionally high performance under all tested tribological conditions whereas the polyurea grease performed poorly in all of those conditions. This was observed despite the fact that the nominal grease bleed rate that is expected to correlate to the protection of the grease was zero for most calcium sulfonate samples and comparatively high for polyurea thickened grease samples. This baffling and counterintuitive result was explained using a model based on the thickener particle.

117 geometry. The model takes a fundamental approach to a rather complex system and provides a new and more precise interpretation of the grease which is capable of explaining the observed phenomena as well as solving a number of previously unapproachable problems about different grease thickener types and accounts for some of their exceptional behaviors under certain tribological conditions that had not been explained before. The most important findings of this work accompanied by the essential understanding of the grease structure in the light of the proposed model based on the micellar thickener structure is explained as follows:

Fretting under rolling and sliding contacts are two different damage mechanisms.

Accordingly, oil release mechanism of the grease under rolling and sliding contacts are essentially different and are both “dynamic bleed” mechanisms different than “static bleed” which occurs during the standard bleed rate experiment. The difference between the model proposed in this study and those that it precedes come from two major assumptions, that considering the close agreement of the model and the experimental results for static bleed, they could be stated, strongly, as an improved description of the grease microstructure and its function. First of these assumptions is that the grease is made of thickener particles suspended in a matrix of oil (and not the other way around) and it should be treated as such to be able to describe its function. Secondly, the thickener particles trap and immobilize its surrounding oil to a certain extent and the resulting structure called the “effective thickener media” could now be taken as a matrix that holds the remaining free-to-flow oil. The static bleed could be precisely calculated as the viscous flow of oil in the effective thickener structure. While the effective thickeners in the static mode, meaning when the grease is sitting still, constitute some kind of a

118 structure that explains the consistency of the grease, such structure is not rigid and the thickener particles are only loosely connected. Therefore the thickener particles are able to move independently from one-another –breaking the loose structure– if a force is applied on them, which explains the rheology of the grease. This also applies to the function of the grease in a tribological contact in which the advancing contact would apply force on the thickener particles and the dynamic bleed of the grease could be simply described as the motion of the effective thickener particles out of the contact, leaving oil behind. Along with the motion of the effective thickener particles in the contact, other dynamic effects such as rotation and accumulation of thickener particles are also incorporated. With this interpretation of the dynamic bleed, the difference in rolling bleed and sliding bleed could be attributed to the way the particles are affected in each scenario. In sliding contact, the motion of thickener particles is along the direction of the motion of the contact resulting in practically unlimited displacement of thickener particles. Whereas, under rolling contact the motion of the thickener particles are perpendicular to the motion of the contact as they are pushed to the sides giving a limited amount of displacement. This limited displacement is often small enough that the dynamic effects in rolling could be neglected and therefore, unlike sliding, the dynamic bleed under rolling is closely related to static bleed. This explains the experimental observation that unlike sliding fretting, rolling fretting showed dependence on the standard bleed rates of the grease samples. With such modeling of the grease bleed, the geometry of the thickener particles are also taken into account and they often play an important role in explaining the observed tribological phenomena. For example, it was shown that the higher the aspect ratio of the thickener particle, the more feasible is the

119 yield of a consistent grease structure. Thickener particles with high aspect ratios are more prone to some of the dynamic bleed effects such as being transported and accumulated in the tribological contact causing eventual clogging. Another dynamic effect, which could play a detrimental role in the performance of the grease is the particle rotational effect and the thickener particles with the highest surface areas such as plates are most prone to that. The DLS/SLS technique was successfully used to measure the thickener particle geometry of the grease, and the lithium complex, calcium sulfonate and polyurea greases were rods, spheres and platelets respectively. The significance of this technique was that it characterized the grease thickeners while still interacting with the oil and to the best of our knowledge this was never done on grease before. The unique static bleed behavior of the calcium sulfonate grease as well as its superior performance can be attributed to its spherical thickener particles, the former being the result of structural opening of the effective thickener spheres and the latter being the result of the ability of the spherical thickeners to avoid being transported by the contact as well as their zero rotational effect, together, effectively eliminating the adverse dynamic bleed effects resulting in no resistance to the separation of the oil from the thickener and its flow into the contact. The poor performance of the polyurea grease was attributed to its plate-like geometry with a high surface area, which makes it extremely sensitive to rotation effects and would therefore easily cause clogging even in low displacement amplitude of the rolling contact fretting. The model proposed in this study is justifiably one of the most important contributions to understanding the grease and its function and could lead to profound and revolutionary advances in the grease industry with the newly

120 found ability to tailor the grease for any particular application by manipulating and tuning the geometry of its micellar thickener structure.

.

121 CHAPTER VII

7 SUMMARY

7.1 Summary of the Most Important Findings of this Study

 Fretting under rolling and sliding contacts are two distinct damage mechanisms

and have dissimilar “dynamic bleed” mechanisms, which are different than “static

bleed” which occurs during the standard bleed rate experiment.

 The grease is made of thickener particles suspended in a matrix of oil (and not the

other way around) and it should be treated as such to be able to describe its

function.

 The thickener particles trap and immobilize its surrounding oil to a certain extent

and the resulting structure called the “effective thickener media” could now be

taken as a matrix that holds the remaining free-to-flow oil.

 The consistency of the grease comes from the loose or semi-stable connection of

the effective thickener particles when the grease is sitting still, and the rheology of

the grease is a result of the independent motion of the effective thickener particles

when the grease is under shear.

 The static bleed could be precisely modeled as the viscous flow of oil in the

effective thickener structure. While the dynamic bleed of the grease could be .

122 simply described as the motion of the effective thickener particles out of the

contact simply described as the motion of the effective thickener particles out of

the contact.

 Along with the motion of the effective thickener particles in the contact, other

dynamic effects such as rotation and accumulation of thickener particles are also

incorporated.

 The difference in the dynamic rolling bleed and sliding bleed could be attributed

to the fact that in the sliding contact, the displacement of thickener particles are

practically unlimited along the motion of the contact, whereas under rolling

contact the motion of the thickener particles are to the sides giving a limited

amount of displacement. This explains the dependency of rolling fretting results

to the standard bleed rate of the grease.

 The geometry of the thickener particles specified in quantities such as the aspect

ratio and the surface area, play an important role in the dynamic bleed model

when the transport, accumulation and the rotation of the thickener particles are

being considered.

 Thickener particles with high aspect ratios are more prone to some of the dynamic

bleed effects such as being transported and accumulated in the tribological contact

that may lead to gradual clogging. Whereas, thickener particles with the highest

surface areas are most prone to thickener particle rotational effects.

 The unique static bleed behavior of the calcium sulfonate grease is the result of

the structural opening of the spherical effective thickener spheres due to their low

aspect ratio.

123  The superior performance of calcium sulfonate grease can be attributed to the

ability of the spherical thickeners to avoid being transported by the contact as well

as their zero rotational effect, effectively eliminating the dynamic bleed effects.

 The poor performance of the polyurea grease was attributed to the high surface

area of its plate-like thickeners, which makes them extremely vunerable to

rotation that would lead to immediate clogging.

 Further experimental results are detailed as follows:

7.1.1 Lubricant Test Results Summary

 Bleed rate did not change linearly with increasing oil content for calcium

sulfonate grease unlike other greases that act linearly.

 Polyurea greases showed the highest bleed rate.

 Calcium sulfonate mixtures have less bleed rates than other greases at up to 60%

added oil.

 While there is a huge difference in the bleed rates of calcium sulfonate and other

greases, all exhibit a linear increase in penetration with increase in oil content,

and polyurea has the sharpest increase.

 The DLS/SLS technique was used to characterize the thickener particles in

micellar level which was later used for the inputs of the developed model.

7.1.2 Wear in Rolling and Sliding Fretting Results Summary

 For both reciprocating sliding and fretting regimes calcium sulfonate showed

superior protection while polyurea showed the least protection.

124  Lithium complex and calcium sulfonate greases produced different wear

characteristics in reciprocating sliding contact while polyurea showed signs of

scuffing.

 Non-uniform wear pattern was seen with calcium sulfonate, composed of unworn

regions and comparatively deep and narrow scars.

 With increasing bleed rate of calcium sulfonate, low wear regions in wear track

increased in size and total wear was reduced.

7.1.3 Scuffing, B-based Additives & WC/a-C:H Coating:

 Reducing the stroke length increased the occurrence of “scuffing” in the contact

during the sliding fretting tests.

 For lithium complex greases scuffing happened in about 50% of the tests for

bleed rates below 12%. Scuffing did not occur at higher bleed rates.

 While all the polyurea grease mixtures and about 80% of lithium complex grease

mixtures experienced scuffing in at least one test, calcium sulfonate mixtures did

not.

 Boron-based additives do not statistically affect scuffing for steel on steel contact.

 Applying WC/a-C:H coating on disk in combination with TIPB is an effective

way for preventing scuffing under reciprocating sliding contact. .

125 CHAPTER VIII

8 FUTURE WORK

The improved understanding of grease structure and its function which was achieved from the model developed in this study paves the way to further investigations into many different aspects of grease properties, functionalities, and applications. The modeling of the grease bleed in this study owed its success to two essential assumptions that reflected a comparatively precise description of the grease structure. Based on the same assumptions, many other properties of grease could also be modeled that may lead to an even better understanding of its structure and function. For example, if the rheology of a grease is modeled within the effective thickener structure, not only can the effect of thickener geometry on rheology be investigated (which is a considerable achievement in itself) with great practical implications, but also, the nature and strength of the connections and interactions of the effective thickener particles with one another as separate entities could be investigated. Another example of future work is the aging and thixotropy of the grease. The modeling of these qualities could lead to an understanding of the kinetics of separation and re-joining of effective thickeners that would be crucial parameters in furthering the understanding of dynamic bleed.

The dynamic bleed itself is another subject that could be studied further with advanced modeling. Other than spherical rolling and sliding conditions that were studied in this research, there are other tribological contact conditions pertaining to entirely different

126 applications that are now ready to be analyzed in the context of the new description of the grease that is obtained in this research. Ultimately, the goal of these models would be to determine a priori the thickener geometry that best suits a given application. With such knowledge, the grease thickener microstructure could be manipulated and tailored for any given application and grease manufacturers could “design” the microstructure to achieve most desired grease properties. Therefore, with the recent breakthrough in understanding grease function and its relationship to thickener geometry, the grease industry could be on the verge of a new revolution.

127 REFERENCES

[1] V.L. Popov, Contact Mechanics and Friction, in: Contact Mech. Frict., 2010: pp.

55–70. doi:10.1007/978-3-642-10803-7.

[2] M. Amiri, M.M. Khonsari, On the thermodynamics of friction and wear-A review,

Entropy. 12 (2010) 1021–1049. doi:10.3390/e12051021.

[3] A. Ramalho, J.C. Miranda, The relationship between wear and dissipated energy

in sliding systems, Wear. 260 (2006) 361–367. doi:10.1016/j.wear.2005.02.121.

[4] T. Cousseau, B. Gra??a, A. Campos, J. Seabra, Friction torque in grease

lubricated thrust ball bearings, Tribol. Int. 44 (2011) 523–531.

doi:10.1016/j.triboint.2010.06.013.

[5] P.M. Lugt, A Review Grease Lubrication in Rolling Bearings, Grease Lubr. Roll.

Bear. (2009) 37–41. doi:10.1002/9781118483961.

[6] Z.R. Zhou, L. Vincent, Lubrication in fretting - A review, Wear. 225–229 (1999)

962–967. doi:10.1016/s0043-1648(99)00038-1.

[7] P.L. Hurricks, The mechanism of fretting - A review, Wear. 15 (1970) 389–409.

doi:10.1016/0043-1648(70)90235-8.

[8] S. Fouvry, P. Kapsa, L. Vincent, Analysis of sliding behaviour for fretting

loadings: determination of transition criteria, Wear. 185 (1995) 35–46.

doi:10.1016/0043-1648(94)06582-9.

[9] S. Fouvry, P. Kapsa, L. Vincent, K. Dang Van, Theoretical analysis of fatigue

cracking under dry friction for fretting loading conditions, Wear. 195 (1996) 21–

34. doi:10.1016/0043-1648(95)06741-8.

[10] S. Fouvry, T. Liskiewicz, P. Kapsa, S. Hannel, E. Sauger, An energy description 128 of wear mechanisms and its applications to oscillating sliding contacts, Wear. 255

(2003) 287–298. doi:10.1016/S0043-1648(03)00117-0.

[11] R.B. Waterhouse, Fretting wear, Wear. 100 (1984) 107–118. doi:10.1016/0043-

1648(84)90008-5.

[12] R.B. Waterhouse, D.E. Taylor, Fretting debris and the delamination theory of

wear, Wear. 29 (1974) 337–344. doi:10.1016/0043-1648(74)90019-2.

[13] Y. Berthier, L. Vincent, M. Godet, Fretting fatigue and fretting wear, Tribol. Int.

22 (1989) 235–242. doi:10.1016/0301-679X(89)90081-9.

[14] M. Godet, The third-body approach: A mechanical view of wear, Wear. 100

(1984) 437–452. doi:10.1016/0043-1648(84)90025-5.

[15] S. SODERBERG, U. BRYGGMAN, T. MCCULLOUGH, Frequency effects in

fretting wear, Wear. 110 (1986) 19–34. doi:10.1016/0043-1648(86)90149-3.

[16] U. Bryggman, S. Soderberg, Contact conditions in fretting, Wear. 110 (1986) 1–

17. doi:10.1016/0043-1648(86)90148-1.

[17] U. Bryggman, S. Soderberg, Contact conditions and surface degradation

mechanisms in low amplitude fretting, Wear. 125 (1988) 39–52.

doi:10.1016/0043-1648(88)90192-5.

[18] D.A. Hills, D. Nowell, J.J. O’Connor, On the mechanics of fretting fatigue, Wear.

125 (1988) 129–146. doi:10.1016/0043-1648(88)90198-6.

[19] O. Vingsbo, J. Schon, Gross slip criteria in fretting, Wear. 162–164 (1993) 347–

356. doi:10.1016/0043-1648(93)90518-Q.

[20] D. Nowell, D.A. Hills, Mechanics of fretting fatigue tests, Int. J. Mech. Sci. 29

(1987) 355–365. doi:10.1016/0020-7403(87)90117-2.

129 [21] O. Vingsbo, S. Soderberg, On fretting maps, Wear. 126 (1988) 131–147.

doi:10.1016/0043-1648(88)90134-2.

[22] O. Vingsbo, M. Odfalk, Ning-E Shen, Fretting maps and fretting behavior of some

F.C.C. metal alloys, Wear. 138 (1990) 153–167. doi:10.1016/0043-

1648(90)90174-9.

[23] L. Vincent, Y. Berthier, M.C. Dubourg, M. Godet, Mechanics and materials in

fretting, Wear. 153 (1992) 135–148. doi:10.1016/0043-1648(92)90266-B.

[24] S. Fouvry, P. Kapsa, L. Vincent, Quantification of fretting damage, Wear. 200

(1996) 186–205. doi:10.1016/S0043-1648(96)07306-1.

[25] M. Shima, H. Suetake, I.R. McColl, R.B. Waterhouse, M. Takeuchi, On the

behaviour of an oil lubricated fretting contact, Wear. 210 (1997) 304–310.

doi:10.1016/S0043-1648(97)00078-1.

[26] T. Kolodziejczyk, S. Fouvry, G.E. Morales-Espejel, A friction energy approach to

quantifying lubrication under fretting sliding, Lubr. Sci. 22 (2010) 53–71.

doi:10.1002/ls.105.

[27] B.D. Leonard, F. Sadeghi, R.D. Evans, G.L. Doll, P.J. Shiller, Fretting of WC/a-

C:H and Cr2N Coatings Under Grease-Lubricated and Unlubricated Conditions,

Tribol. Trans. 53 (2010) 145–153. doi:10.1080/10402000903312323.

[28] Z.R. Zhou, Q.Y. Liu, M.H. Zhu, L. Tanjala, P. Kapsa, L. Vincent, Investigation of

fretting behaviour of several metallic materials under grease lubrication, Tribol.

Int. 33 (2000) 69–74. doi:10.1016/S0301-679X(99)00100-0.

[29] Z.R. Zhou, P. Kapsa, L. Vincent, Grease lubrication in fretting, J. Tribol. 120

(1998) 737–743. doi:10.1115/1.2833773.

130 [30] M. Kalin, J. Vižintin, The tribological performance of DLC coatings under oil-

lubricated fretting conditions, Tribol. Int. 39 (2006) 1060–1067.

doi:10.1016/j.triboint.2006.02.040.

[31] Q.Y. Liu, Z.R. Zhou, Effect of displacement amplitude in oil-lubricated fretting,

Wear. 239 (2000) 237–243. doi:10.1016/S0043-1648(00)00323-9.

[32] Z. a. Wang, Z.R. Zhou, G.X. Chen, An investigation of palliation of fretting wear

in gross slip regime with grease lubrication, Ind. Lubr. Tribol. 63 (2011) 84–89.

doi:10.1108/00368791111112207.

[33] Y. Wang, T. Lei, L. Guo, B. Jiang, Fretting wear behaviour of microarc oxidation

coatings formed on titanium alloy against steel in unlubrication and oil lubrication,

Appl. Surf. Sci. 252 (2006) 8113–8120. doi:10.1016/j.apsusc.2005.10.032.

[34] I.R. McColl, R.B. Waterhouse, S.J. Harris, M. Tsujikawa, Lubricated fretting

wear of a high-strength eutectoid steel rope wire, Wear. 185 (1995) 203–212.

doi:10.1016/0043-1648(95)06616-0.

[35] L. Haviez, S. Fouvry, R. Toscano, G. Yantio, An energy-based approach to

understand the effect of fretting displacement amplitude on grease-lubricated

interface, Wear. 338–339 (2015) 422–429. doi:10.1016/j.wear.2015.07.015.

[36] G. Pagnoux, S. Fouvry, M. Peigney, B. Delattre, G. Mermaz-Rollet, A model for

single asperity perturbation on lubricated sliding contact with DLC-coated solids,

Tribol. Int. 82 (2015) 423–430. doi:10.1016/j.triboint.2014.05.009.

[37] ASTM D5707-11 Standard Test Method for Measuring Friction and Wear

Properties of Lubricating Grease Using a High-Frequency, Linear-Oscillation

(SRV) Test Machine, ASTM Stand. (n.d.).

131 [38] Astm, ASTM D4170: Standard Test Method for Fretting Wear Protection by

Lubricating Greases 1, ASTM Stand. XLVII (2011) 6–11. doi:10.1520/D4170-

10.2.

[39] M.C. Sánchez, J.M. Franco, C. Valencia, C. Gallegos, F. Urquiola, R. Urchegui,

Atomic force microscopy and thermo-rheological characterisation of lubricating

greases, Tribol. Lett. 41 (2011) 463–470. doi:10.1007/s11249-010-9734-x.

[40] L. Salomonsson, G. Stang, B. Zhmud, Oil/Thickener Interactions and Rheology of

Lubricating Greases, Tribol. Trans. 50 (2007) 302–309.

doi:10.1080/10402000701413471.

[41] P. Baart, B. van der Vorst, P.M. Lugt, R. a. J. van Ostayen, Oil-Bleeding Model

for Lubricating Grease Based on Viscous Flow Through a Porous Microstructure,

Tribol. Trans. 53 (2010) 340–348. doi:10.1080/10402000903283326.

[42] F. Cyriac, P.M. Lugt, R. Bosman, C.J. Padberg, C.H. Venner, Effect of Thickener

Particle Geometry and Concentration on the Grease EHL Film Thickness at

Medium Speeds, Tribol. Lett. 61 (2016) 18. doi:10.1007/s11249-015-0633-z.

[43] G. Fish, CALCIUM SULPHONATE GREASES Performance and application

overview, White Pap. Lubrisense. 16 (2014).

[44] ASTM D1742 - 06 Standard Test Method for Oil Separation from Lubricating

Grease During Storage, ASTM Stand. (n.d.).

[45] A. Adhvaryu, C. Sung, S.Z. Erhan, Fatty acids and antioxidant effects on grease

microstructures, Ind. Crops Prod. 21 (2005) 285–291.

doi:10.1016/j.indcrop.2004.03.003.

[46] Y.L. Ishchuk, Lubricating Grease Manufacturing Technology, (n.d.).

132 [47] T.A. Renshaw, Effects of Shear on Lithium Greases and Their Soap Phase, Ind.

Eng. Chem. 47 (1955) 834–838.

[48] G. Palazzo, Wormlike reverse micelles, Soft Matter. 9 (2013) 10668.

doi:10.1039/c3sm52193a.

[49] M.A. Delgado, C. Valencia, M.C. Sánchez, J.M. Franco, C. Gallegos, Influence of

soap concentration and oil viscosity on the rheology and microstructure of

lubricating greases, Ind. Eng. Chem. Res. 45 (2006) 1902–1910.

doi:10.1021/ie050826f.

[50] Using Fatty Acids in Lubricating Greases, Assoc. Am. Soap Glycerine Prod. Fat.

Acid Div. (n.d.).

[51] M. Paszkowski, Some Aspects of Grease Flow in Lubrication Systems and

Friction Nodes, Tribol. Adv. chapter 3 (2013) 78–106. doi:44847.

[52] B.W. Hotten, D.H. Birdsall, Fine structure and rheological properties of lithium

soap-oil dispersions, J. Colloid Sci. 7 (1952) 284–294.

doi:http://dx.doi.org/10.1016/0095-8522(52)90074-3.

[53] M. Paszkowski, S. Olsztynska-Janus, Grease thixotropy: evaluation of grease

microstructure change due to shear and relaxation, Ind. Lubr. Tribol. 66 (2014)

223–237. doi:10.1108/ILT-02-2012-0014.

[54] Grease Static Oil Bleed, (2012) 2012.

[55] MOBIL, Grease Oil Bleed — An Essential Characteristic, (n.d.).

[56] R.M. Hurley, S , Cann, Examination of Grease Structure by SEM and AFM

Techniques, in: NLGI Spokesm., n.d.: pp. 17–26.

[57] P.J. Blau, K.G. Budinski, Development and use of ASTM standards for wear

133 testing, Wear. 225–229 (1999) 1159–1170. doi:10.1016/S0043-1648(99)00045-9.

[58] C.H. Hager, P.J. Shiller, Effect of Grease Composition on Fretting and Wear, in:

STLE Anu. Meet., 2011.

[59] A. Cristina, M. De Farias, A. Augusto, S. Medeiros, Correlation Between Fuel

Lubricity and Vibration Signals Obtained in Ball-disc Analysis Using Fourier

Transform, 18 (2015) 210–219.

[60] M.D. Bryant, Entropy and dissipative processes of friction and wear, FME Trans.

37 (2009) 55–60.

[61] B.R. Gebart, Permeability of Unidirectional Reinforcements for RTM, J. Compos.

Mater. 26 (1992) 1100–1133. doi:10.1177/002199839202600802.

[62] D.-H. Cho, J.-S. Kim, J. Jia, Y.-Z. Lee, Comparative analysis based on adiabatic

shear instability for scuffing failure between unidirectional and reciprocating

sliding motion, Wear. 297 (2013) 774–780.

doi:http://dx.doi.org/10.1016/j.wear.2012.10.014.

[63] O.O. Ajayi, C. Lorenzo-Martin, R.A. Erck, G.R. Fenske, Analytical predictive

modeling of scuffing initiation in metallic materials in sliding contact, Wear. 301

(2013) 57–61. doi:10.1016/j.wear.2012.12.054.

[64] R. Lumbreras, M. Majcher, L. Guessous, G. Barber, J.D. Schall, Q. Zou,

Numerical modeling of the transient temperature rise during ball-on-disk scuffing

tests, (2014).

[65] J. Wang, Modern thermodynamics – New concepts based on the second law of

thermodynamics, Prog. Nat. Sci. 19 (2009) 125–135.

doi:10.1016/j.pnsc.2008.07.002.

134 [66] O.O. Ajayi, C. Lorenzo-Martin, R.A. Erck, G.R. Fenske, Scuffing mechanism of

near-surface material during lubricated severe sliding contact, Wear. 271 (2011)

1750–1753. doi:10.1016/j.wear.2010.12.086.

[67] M. Chandrasekaran, A.W. Batchelor, N.L. Loh, Frictional seizure of aluminium

observed by X-ray imaging, Tribol. Int. 35 (2002) 297–308. doi:10.1016/S0301-

679X(02)00006-3.

[68] J. Sundh, U. Olofsson, K. Sundvall, Seizure and wear rate testing of wheel–rail

contacts under lubricated conditions using pin-on-disc methodology, Wear. 265

(2008) 1425–1430. doi:10.1016/j.wear.2008.03.025.

[69] J.-W. Kim, Y.-Z. Lee, The residual stresses on lubricated sliding surfaces during

break-in and up to scuffing, Wear. 251 (2001) 985–989. doi:10.1016/S0043-

1648(01)00701-3.

[70] S. Kajita, K. Yagi, T. Izumi, J. Koyamachi, M. Tohyama, K. Saito, et al., In Situ

X-Ray Diffraction Study of Phase Transformation of Steel in Scuffing Process,

Tribol. Lett. 57 (2015) 6. doi:10.1007/s11249-014-0443-8.

[71] K. Yagi, S. Kajita, T. Izumi, J. Koyamachi, M. Tohyama, K. Saito, et al.,

Simultaneous Synchrotron X-ray Diffraction, Near-Infrared, and Visible In Situ

Observation of Scuffing Process of Steel in Sliding Contact, Tribol. Lett. 61

(2016) 19. doi:10.1007/s11249-015-0636-9.

[72] M. Chandrasekaran, A.W. Batchelor, N.L. Loh, Direct observation of frictional

seizure of mild steel sliding on aluminum by X-ray imaging Part I Methods, J.

Mater. Sci. 35 (2000) 1589–1596. doi:10.1023/A:1004727124456.

[73] H. Li, K. Yagi, J. Sugimura, S. Kajita, T. Shinyoshi, Role of Wear Particles in

135

Scuffing Initiation, Tribol. Online. 8 (2013) 285–294. doi:10.2474/trol.8.285.

[74] K. Yagi, Y. Ebisu, J. Sugimura, S. Kajita, T. Ohmori, A. Suzuki, In Situ

Observation of Wear Process Before and During Scuffing in Sliding Contact,

Tribol. Lett. 43 (2011) 361–368. doi:10.1007/s11249-011-9817-3.

[75] H. Blok, Recent developments in gear tribology, Arch. Proc. Inst. Mech. Eng.

Conf. Proc. 1964-1970 (Vols 178-184), Var. Titles Label. Vol. A to S. 184 (1969)

21–29. doi:10.1243/PIME_CONF_1969_184_438_02.

[76] W.J.S. Grew, A. Cameron, Thermodynamics of Boundary Lubrication and

Scuffing, Proc. R. Soc. London A Math. Phys. Eng. Sci. 327 (1972).

[77] H.A. Spikes, A. Cameron, Scuffing as a Desorption Process—An Explanation of

the Borsoff Effect, A S L E Trans. 17 (1974) 92–96.

doi:10.1080/05698197408981442.

[78] A. Dyson, The failure of elastohydrodynamic lubrication of circumferentially

ground discs, Arch. Proc. Inst. Mech. Eng. 1847-1982 (Vols 1-196). 190 (1976)

699–711. doi:10.1243/PIME_PROC_1976_190_074_02.

[79] P. Eriksson, V. Wikström, R. Larsson, Grease soap particles passing through an

elastohydrodynamic contact under side slip conditions, Proc. Inst. Mech. Eng. Part

J J. Eng. Tribol. 214 (2000) 317–325. doi:10.1243/1350650001543205.

[80] M.R. Riggs, N.K. Murthy, S.P. Berkebile, Scuffing Resistance and Starved

Lubrication Behavior in Helicopter Gear Contacts: Dependence on Material,

Surface Finish and Novel Lubricants, Tribol. Trans. (2016) 0.

doi:10.1080/10402004.2016.1231358.

[81] G.T.Y. Wan, H. Lankamp, A. de Vries, E. Ioannides, The effect of extreme

136

pressure (EP) lubricants on the life of rolling element bearings, Arch. Proc. Inst.

Mech. Eng. Part J J. Eng. Tribol. 1994-1996 (Vols 208-210). 208 (1994) 247–252.

doi:10.1243/PIME_PROC_1994_208_379_02.

[82] E. Forbes, Antiwear and extreme pressure additives for lubricants, Tribology.

(1970).

[83] D. Godfrey, Chemical Changes in Steel Surfaces During Extreme Pressure

Lubrication, A S L E Trans. 5 (1962) 57–66. doi:10.1080/05698196208972453.

[84] U. Muralidharan, D.C. Witte, Evaluation of Scuffing Resistance of Grease-

Lubricated Tapered Roller Bearings, Tribol. Trans. 38 (1995) 317–322.

doi:10.1080/10402009508983411.

[85] Z. Song, M. Fan, Y. Liang, F. Zhou, W. Liu, Lithium-Based Ionic Liquids: In

Situ-Formed Lubricant Additive Only by Blending, Tribol. Lett. 49 (2013) 127–

133. doi:10.1007/s11249-012-0046-1.

[86] L. Chen, X. Zhang, H. Xu, J. Dong, Tribological Investigation of Two Different

Layered Zirconium Phosphates as Grease Additives Under Reciprocating Sliding

Test, Tribol. Lett. 64 (2016) 1. doi:10.1007/s11249-016-0730-7.

[87] J. Palacios, Films formed by antiwear additives and their incidence in wear and

scuffing, Wear. (1987).

[88] W. Piekoszewski, M. Szczerek, W. Tuszynski, The action of lubricants under

extreme pressure conditions in a modified four-ball tester, Wear. (2001).

[89] W. Piekoszewski, M. Szczerek, W. Tuszynski, A method for testing lubricants

under conditions of scuffing. Part II. The anti-seizure action of lubricating oils,

Tribotest. 9 (2002) 35–48. doi:10.1002/tt.3020090105.

137

[90] N. Dhopatkar, J.H. Park, K. Chari, A. Dhinojwala, Adsorption and Viscoelastic

Analysis of Polyelectrolyte−Surfactant Complexes on Charged Hydrophilic

Surfaces, (n.d.). doi:10.1021/la5043052.

[91] X. Wu, X. Wang, W. Liu, H.K. Trivedi, N.H. Forster, C.S. Saba, et al.,

Tribological properties of naphthyl phenyl diphosphates as antiwear additive in

polyalkylene glycol and polyurea grease for steel/steel contacts at elevated

temperature, RSC Adv. 4 (2014) 6074. doi:10.1039/c3ra46591h.

[92] H. Schultheiss, J.-P. Stemplinger, T. Tobie, K. Stahl, Influences on Failure Modes

and Load-Carrying Capacity of Grease-Lubricated Gears, (n.d.).

[93] H. Schultheiss, T. Tobie, K. Stahl, The Effect of Selected Grease Components on

the Wear Behavior of Grease-Lubricated Gears, J. Tribol. 138 (2015) 11602.

doi:10.1115/1.4031278.

[94] H. Schultheiss, T. Tobie, K. Michaelis, B.-R. Höhn, K. Stahl, The Slow-Speed

Wear Behavior of Case-Carburized Gears Lubricated with NLGI 00 Grease under

Boundary Lubrication Conditions, Tribol. Trans. 57 (2014) 524–532.

doi:10.1080/10402004.2014.883005.

[95] P. Shiller, Measuring the “Worms” in Grease, (2009).

[96] ASTM International, Standard Test Method for Kinematic Viscosity of

Transparent and Opaque Liquids ( and Calculation of Dynamic Viscosity ) 1,

Annu. B. ASTM Stand. (2010) 1–10. doi:10.1520/D0445-11A.In.

[97] ASTM D2265 - 15 Standard Test Method for Dropping Point of Lubricating

Grease Over Wide Temperature Range, ASTM Stand. (n.d.).

[98] ASTM D2266-01(2015) Standard Test Method for Wear Preventive

138

Characteristics of Lubricating Grease (Four-Ball Method), ASTM Stand. (n.d.).

[99] ASTM D2509-14 Standard Test Method for Measurement of Load-Carrying

Capacity of Lubricating Grease (Timken Method), ASTM Stand. (n.d.).

[100] ASTM D2596-15 Standard Test Method for Measurement of Extreme-Pressure

Properties of Lubricating Grease (Four-Ball Method), ASTM Stand. (n.d.).

[101] ASTM D1743 - 05 Standard Test Method for Determining Corrosion Preventive

Properties of Lubricating Greases, ASTM Stand. (n.d.).

[102] ASTM D4049 - 06 Standard Test Method for Determining the Resistance of

Lubricating Grease to Water Spray, ASTM Stand. (n.d.).

[103] Page 1 of 32 results ASTM D1403 - 02 Standard Test Methods for Cone

Penetration of Lubricating Grease Using One-Quarter and One-Half Scale Cone

Equipment, ASTM Stand. (n.d.).

[104] ASTM D217 - 10 Standard Test Methods for Cone Penetration of Lubricating

Grease, ASTM Stand. (n.d.).

[105] S. Kucharski, Z. Mrz, Identification of wear process parameters in reciprocating

ball-on-disc tests, Tribol. Int. 44 (2011) 154–164.

doi:10.1016/j.triboint.2010.10.010.

[106] S. Heredia, S. Fouvry, Introduction of a new sliding regime criterion to quantify

partial, mixed and gross slip fretting regimes: Correlation with wear and cracking

processes, Wear. 269 (2010) 515–524. doi:10.1016/j.wear.2010.05.002.

[107] K.C. Ludema, Friction, wear, lubrica tion, 1996. doi:10.1201/9781439821893.

[108] T.D. Cores, Standard Test Method for Evaluating Lubricity of Diesel Fuels by the

High-Frequency Reciprocating Rig (HFRR), i (2012) 1–6. doi:10.1520/C0496.

139

[109] M. Rhodes, Introduction to Particle Technology – Second Edition, 2nd ed., Wiley

Subscription Services, Inc., A Wiley Company, 2008.

[110] Y. Shi, Y.L. Taek, A.S. Kim, Permeability calculation of sphere-packed porous

media using dissipative particle dynamics, Desalin. Water Treat. 34 (2011) 277–

283. doi:10/5004/dwt.2011.2802.

[111] M. V. Bruschke, S.G. Advani, Flow of generalized Newtonian fluids across a

periodic array of cylinders, J. Rheol. (N. Y. N. Y). 37 (1993) 479.

doi:10.1122/1.550455.

140