Tribological Behaviour of Polymers in Lubricated Contacts

ISSN: 1402-1757 ISBN 978-91-7439-XXX-X Se i listan och fyll i siffror där kryssen är

LICENTIATE T H E SIS

Department of Engineering Sciences and Mathematics, Division of Machine Elements Arash Golchin Tribological Behaviour Contacts of Polymers in Lubricated

ISSN: 1402-1757 Tribological Behaviour of Polymers ISBN 978-91-7439-610-2 (tryckt) ISBN 978-91-7439-611-9 (pdf) in Lubricated Contacts Luleå University of Technology 2013

Arash Golchin

Tribological Behaviour of Polymers in Lubricated Contacts

Arash Golchin

Luleå University of Technology Department of Engineering Sciences and Mathematics, Division of Machine Elements

May 2013

Printed by Universitetstryckeriet, Luleå 2013

ISSN: 1402-1757 ISBN 978-91-7439-610-2 (tryckt) ISBN 978-91-7439-611-9 (pdf) Luleå 2013 www.ltu.se Preface

The work presented in this thesis is based on the research carried out at the Division of Machine Elements at Luleå University of Technology. The experimental work has been mainly carried out at Luleå University of Technology and a part of the work has been accomplished at Ghent University, Belgium.

This work has been partially funded by “StandUp for Energy” which is a collaboration initiative between Uppsala University, Royal Institute of Technology (KTH), Swedish University of Agricultural Sciences (SLU) and Luleå University of Technology (LTU).

Swedish Research School in Tribology is greatly acknowledged for financing the research visit to Ghent University, Belgium.

I would like to express my sincere gratitude to my supervisors Professor Braham Prakash and Professor Sergei Glavatskih for their fruitful discussions, support and encouragements. I would also like to thank my colleagues in the Division of Machine Elements for providing an enjoyable place to work.

Finally, I would like to express my greatest gratitude to my family especially my wife, Nassim, for her immense love and support during my studies and beyond.

iii

Abstract

An important issue in hydropower and other industries is the increasing demand for introduction of environment friendly solutions. Mineral oil based lubricants have long been used in various sliding bearings in hydropower stations. Their use in aqueous environments however raises concerns about the environmental impact if they leak into downstream water. This has necessitated research in replacing mineral oils with more bio-degradable lubricants and the ultimate goal of ‘oil-free’ hydropower machines.

Replacing oil with water however poses many challenges. Due to considerably lower viscosity of water compared to that of turbine oils, water lubricated bearings are likely to operate in boundary/mixed lubrication regime for a relatively longer period. Therefore choice of the materials and their tribological performance are very important for these bearings. Application of compliant polymers in water lubricated bearings introduces many advantages which cannot be achieved with coatings (DLC, etc.) or ceramics. However, most previous tribological studies on polymers have been carried out in dry conditions and few studies in presence of water have been reported. This work is thus aimed at investigating the tribological behaviour of some selected polymer materials in water lubricated conditions. The results of these studies provide an insight into polymers’ tribological performance and associated wear mechanisms in the presence of water.

Polymers also enhance performance of oil lubricated bearings. Application of polymers in oil lubricated bearings provides a smooth transition from oil lubricated Babbitt bearings to water lubricated polymer bearings. Therefore a part of this thesis is also aimed at investigating the tribological characteristics of several polytetrafluoroethylene based materials at the onset of sliding (break-away friction) at different pressures and temperatures. The results of this study show significantly lower breakaway friction of PTFE materials compared to Babbitt at all pressures and temperatures. SEM investigations revealed wear modes of the PTFE materials and the abrasive nature of hard fillers. Bronze-filled, carbon-filled and pure PTFE were found to provide lower and more stable break-away friction and generally superior properties compared to the other materials.

Appended Papers

[A] Golchin, A., Simmons, G.F., Glavatskih, S., Break-away Friction of PTFE Materials in Lubricated Conditions, Tribology International, 2012, Vol. 48, pp.54-62.

[B] Golchin, A., Simmons, G. F., Glavatskih, S., Prakash, B., Tribological Behaviour of Polymeric Materials in Water Lubricated Contacts, Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology, Published online before print March 12, 2013, doi: 10.1177/1350650113476441

[C] Golchin, A., Nguyen, T. D., De Baets, P., Glavatskih S., Prakash, B., Effect of Shaft Roughness and Pressure on Friction of Polymer Bearings in Water, To be communicated.

vii

Table of Contents 1.Introduction ...... 1 1.1. Hydrodynamic Sliding Bearings ...... 3 1.2. Water Lubricated Bearings ...... 5 1.3. Research Gaps ...... 10 1.4. Objectives of Present Work ...... 10 2. Experimental Work ...... 11 2.1. Paper A...... 11 2.2. Paper B ...... 13 2.3. Paper C ...... 15 3. Results ...... 21 3.1. Paper A...... 21 3.2. Paper B ...... 26 3.3. Paper C ...... 31 4. Conclusions ...... 37 5. Future Work ...... 39 References ...... 41

Appended Papers A: Break-away friction of PTFE materials in lubricated conditions ...... 43 A.1. Introduction ...... 45 A.2. Experimental study ...... 46 A.2.1. Materials...... 47 A.2.2. Experimental setup...... 47 A.2.3. Break-away friction...... 48 A.2.4. Materials preparation ...... 49 A.3. Results and discussion ...... 49 A.3.1. Contact pressure effects ...... 50 A.3.2. Temperature effects ...... 52 A.3.3. Pressure and temperature interaction ...... 52 A.3.4. Prolonged stop and restart ...... 52 A.3.5. Counter-surface wear ...... 53 A.3.6. SEM investigation ...... 54

A.4. Conclusions...... 57

B: Tribological Behaviour of Polymeric Materials in Water Lubricated Contacts ...... 61 B.1. Introduction ...... 63 B.2. Experimental work...... 64 B.2.1. Materials ...... 64 B.2.2. Experimental Setup ...... 66 B.2.3. Sample Preparation ...... 66 B.3. Results and Discussion ...... 67 B.3.1. Friction...... 67 B.3.2. Evolution of Friction ...... 68 B.3.3. Water Absorption ...... 70 B.3.4. Wettability ...... 70 B.3.5. Solubility Parameters...... 72 B.3.6. Wear...... 75 B.3.7. Effect on Stainless Steel Counter Surface ...... 80 B.4. Conclusions ...... 81

C: Effect of Shaft Roughness and Pressure on Friction of Polymer Bearings in Water ...... 85 C.1. Introduction...... 87 C.2. Experimental Work ...... 88 C.2.1. Materials ...... 88 C.2.2. Test specimens ...... 89 C.2.3. Experimental setup and procedure ...... 90 C.3. Results ...... 91 C.3.1. Friction evolution during running-in ...... 91 C.3.2. Friction Maps ...... 94 C.4. Discussions ...... 96 C.4.1. Running-in behaviour ...... 96 C.4.2. Influence of shaft roughness ...... 99 C.4.3. Influence of pressure ...... 100 C.5. Conclusions ...... 101

Introduction

One of the most important demands in hydropower and other industries today stems from the increasing awareness of the environment and limited mineral resources. The current push towards introducing more environmental friendly solutions has led to many efforts in utilizing renewable and bio-degradable lubricants in machinery as well as in hydropower plants. The recent efforts in replacing conventional mineral oils with more bio-degradable and renewable lubricants have contributed to reduction of the risks posed to the environment. However the environmental impact of these lubricants cannot be neglected if they leak into downstream water. Figure 1.1 shows pollution of the Luleälv which was caused by turbine oil leakage from Laxede hydropower station in October 2003.

Figure 1.1: Oil pollution of the Luleälv caused by turbine oil leakage from Laxede hydropower station, Oct. 2003 [1]

2 Introduction

These concerns have led to the concept of using water as a lubricant and introduced the concept of “Oil-Free Plant” for hydropower generating companies. This however poses many engineering challenges which require re-thinking of the different aspects of bearings e.g. design, operating conditions and selection of shaft and bearing materials. Due to considerably lower viscosity of water (0.66 cSt at 40 ºC) compared to that of turbine oils (32-68 cSt at 40 ºC), water lubricated bearings are likely to operate in boundary/mixed lubrication regime for relatively longer periods. Therefore choice of the materials and their tribological behaviour are very important for the proper performance and extended lifespan of such bearings operating in boundary/mixed lubrication regime. Application of compliant polymers in water lubricated bearings introduces many advantages which cannot be achieved with conventional metallic bearing materials, coatings or ceramics. While most previous tribological studies on polymers have been carried out in dry conditions, few studies in presence of water have been reported. This work is thus mainly aimed at investigating the tribological behaviour of polymeric materials in water lubricated conditions. The results of these studies provide an insight into polymers’ tribological performance and associated wear mechanisms in the presence of water. Application of polymers as bearing materials can as well improve the performance of the existing oil-lubricated hydrodynamic sliding bearings in hydropower plants. Currently there are strong environmental and political demands towards production of more energy from renewable sources such as wind, wave, solar and tidal energy which are all intermittent in nature. However, introducing more intermittent sources of energy to the grid requires more regulating power. This demand has altered the role of hydropower from base electricity production to more participation in regulation and stabilization of the grid frequency through primary and secondary regulation. Primary regulation in hydropower plants is done by automatic adjustment of guide vane angle and flow of water while secondary regulation is done by start-up and shut-down of the hydropower turbines. The new operating conditions (starts and stops) imposed on hydropower plants have great impact on hydrodynamic sliding bearings which are susceptible to damage from frequent starts and stops. In steady state operation, the shaft and bearing surfaces are separated by a hydro-dynamic fluid film, providing low friction and long service life of the mating components. During the transient

Introduction 3 phases of start-ups and stops however a contact between shaft and bearing surfaces is inevitable resulting in wear. To start up the turbine, considerable torque is required to overcome the breakaway friction in a steel-Babbitt contact. In order to prevent bearing wear and reduce start- up friction, a hydraulic jacking system may be employed. This provides a hydrostatically pressurized oil film between the contacting surfaces to ensure smooth start-up of the turbine and minimize damage of the bearings. Application of polymers as bearing lining material has the potential to reduce the start-up friction and eliminate the need for hydraulic jacking system. Therefore, a part of this work is aimed at investigating the breakaway friction of some commercially available PTFE based materials using synthetic ester turbine oil as lubricant at a wide range of operating pressures and temperatures. The results have been compared to that of Babbitt (white metal) which is conventionally used as bearing lining material in hydrodynamic sliding bearings of hydropower plants.

1.1. Hydrodynamic Sliding Bearings In hydrodynamic sliding bearings, lubricant is dragged into a converging gap by relative motion of the mating surfaces providing a hydrodynamically pressurized lubricant film. At certain operating conditions, when the pressure in the film is large enough to carry the load, separation of the surfaces occurs by a self-acting lubricant film. In hydropower plants, two main hydrodynamic sliding bearing types include guide (journal) bearings (Figure 1.2.a) and thrust bearings (Figure 1.2.b).

Figure 1.2: a) Tilting pad journal bearing [2] b) Tilting pad thrust bearing [2]

4 Introduction

Figure 1.3: Schematic of a Kaplan turbine with a vertical shaft showing journal bearings (#1,2,3) and thrust bearing (#4) [3]

In turbines with a vertical shaft, the thrust bearing supports the weight of the shaft and the thrust force of the flowing water, while the main guide bearings support the shaft to maintain its position in the radial direction. Figure 1.3 shows a schematic of a typical turbine-generator assembly and the position of the bearings. In a lubricated contact at rest or at very low rotational speeds, the load is mainly transferred through the asperity-asperity contacts of the mating surfaces. This lubrication regime is referred to as boundary lubrication which is generally associated with relatively high friction and wear due to severity of contact between the asperities. As the sliding speed is increased, hydrodynamic pressure is built up and the load is partly supported by the fluid film. This lubrication regime is referred to as mixed lubrication which is generally associated with lower friction and wear compared to the boundary lubrication regime. If sliding speed or lubricant viscosity is further increased and/or load is decreased, the interacting surfaces may be fully separated by a hydrodynamic fluid film. This lubrication regime is referred to as full film lubrication which is generally associated with very low friction and theoretically no wear of the mating surfaces.

Introduction 5

Figure 1.4: A typical Stribeck curve showing different lubrication regimes [4]

Figure 1.4 shows the three different lubrication regimes in a typical Stribeck curve. Hydrodynamic sliding bearings are designed to operate in the full film lubrication regime at the nominal speed of the turbine shaft where the bearing surfaces are separated by a thick film of lubricant to ensure low frictional losses and long service life of the bearings. Using a low viscosity lubricant or imposing frequent starts and stops can adversely affect the hydrodynamic pressure build up. This can lead to longer operation in boundary/mixed lubrication regime where tribological behaviour of the mating surfaces plays a significant role in proper performance of the bearings.

1.2. Water Lubricated Bearings Water as lubricant is environmental friendly, non-toxic, readily available and provides higher specific heat capacity compared to typical turbine oils [5]. However a major drawback of using water as a lubricant is its low viscosity. Figure 1.5 shows the kinematic viscosity of water at different temperatures. Despite the low viscosity of water, water lubricated bearings can be found in machines used in pharmaceutical and food industry, rolling mills [6], mining and textile machinery, submerged water and process fluid pumps, etc. Another common application of water lubricated sliding bearings is in marine applications.

6 Introduction

2.00 /s] 2 1.75

mm 1.50 1.25 1.00 0.75 0.50

Kinematic Viscosity [ Viscosity Kinematic 0.25 0.00 0 20 40 60 80 100 Temperature [ºC]

Figure 1.5: Kinematic viscosity of water vs. temperature

Oil lubricated stern tube bearings which support the propeller shaft of sub-marines and ships (Fig. 1.6), are accountable for 10 million liters of oil leakage per year into the marine environment [7]. Therefore water lubricated stern tube bearings are being employed to eliminate the environmental impact of these lubricants on the marine environment.

Figure 1.6: Propeller shaft and stern tube bearing [8]

Conventionally, “Lignum vitae” was used in water lubricated stern tube bearings from 1854 [5]. Lingum vitae is a dense wood which exhibits low friction against steel in water lubricated contacts [9]. Low friction characteristic of lignum vitae is attributed to the lubricating action of wood waxes which leach from the wood during sliding. Figure 1.7 shows pads of a bearing made of lignum vitae.

Introduction 7

Figure 1.7: Bearing pads made of lignum vitae [10]

Rubber bearings gradually became common in pumps, naval and many commercial ships by the 1940s. The need of a better material for water lubricated bearings was realized in 1942 when a number of U.S. ships suffered extensive bearing damage at the battle of Midway. The bearing damage was due to hysteretic softening of natural rubber which was caused by high speed impact forces of bent shafts and damaged propeller blades. The natural rubber bearing material was rapidly replaced by nitrile rubber, a synthetic elastomer that did not exhibit hysteretic softening [11].

Nitrile rubber bearings are still being used in water lubricated contacts, however high friction and wear of the counter-surface are the main drawbacks of these bearings. Current commercial materials for application as bearing lining in water lubricated conditions include compound polymers of mainly nitrile rubber, UHMWPE, PEEK or PTFE with commercial names such as Vesconite ®, Thordon SXL ® and Romor ® to name a few (Fig 1.8).

a) b) c) Figure 1.8: Water lubricated bearings made of commercial polymers a) Romor ® [12] b) Vesconite ® [13] c) Thordon SXL® [14]

8 Introduction

Water lubricated hydrodynamic sliding bearings typically consist of a metallic shell with inner polymer lining with grooves along the length of the bearing as shown in Figure 1.8. These grooves provide a pathway out for the abrasive particles which find their way into the clearance of the bearing and supply water to lubricate and remove heat from the bearing surfaces. However, hydrodynamic pressure build-up is disturbed as a result of these relatively low pressure regions and the maximum operating projected bearing pressure, is reduced to approximately 280 kPa. While almost all commercial solutions for water lubricated bearings consist of polymeric materials, one may think of an alternative material for such application. One of the main requirements of the materials used in water lubricated contacts is corrosion resistance which is easily fulfilled by ceramic materials [15]. Earlier studies show superior tribological performance of self-mated SiC and Si3N4 in water lubricated conditions [16,17,18,19,20]. This is attributed to the ability of these ceramics to undergo tribo-chemical polishing in water according to the following reactions [21, 22]:

SiC + 2H 2O → SiO2 + CH 4 Si N + 6H O → 3SiO + 4NH 3 4 2 2 3

Under favorable conditions, the tribo-chemical wear results in polishing of the surfaces and enhances the hydrodynamic lubrication of the contacting surfaces. At certain operating condition this would allow for fluid film lubrication and separation of the surfaces resulting in very low friction and wear of the materials. However there are many limitations for application of ceramics in large bearings. Silicon containing ceramics are expensive, brittle and hard to machine to close tolerances for industrial components [23]. Therefore manufacturing of precision ceramic bearings in large scales is a challenge in itself. Relatively high friction coefficients at initiation of motion obtained with self-mated SiC (~0.4) and self-mated Si3N4 (~1) [16, 17] requires a considerable amount of torque and power to start the turbine. The vibrations due to stick-slip at start-up and frictional heating can potentially result in thermal or mechanical fracture and failure of the bearing.

Introduction 9

Application of coatings is one of the methods to achieve the desired tribological performance of tribo-pairs. One of the coating materials which exhibits high wear resistance and low friction under both dry and wet conditions is Diamond-Like Carbon (DLC). DLC comprises a family of materials with sp2 and sp3 bonds between carbon atoms. As is well known, carbon-carbon inter- atomic bonds can be of two types: the near-planar trigonal or sp2 form found in graphite, or the tetragonal sp3 variety that occurs in diamond. DLC is intermediate in that it contains both types of bonding and clearly it is harder and more brittle if the sp3:sp2 ratio is high. Properties of DLC can be far more readily tailored than those of diamond due to its microstructure which allows incorporation of other elements such as nitrogen, silicon, sulfur, tungsten, silver, etc.. However high internal compressive stress of DLC coatings, usually several GPa, inhibits good adhesion of the coatings to the substrate [24]. This limits deposition thickness of DLC coatings to around 1 µm [25]. Another limitation for application of DLC in large bearings is the possible thermal deflection of bearing pads. This imposes additional interfacial stresses between DLC and the substrate and further weakens coating adhesion to the substrate. The rather small deposition thickness of DLC and its susceptibility to substrate deformation make it unsuitable for application in large bearings. Application of compliant polymers as a bearing lining material can provide many advantages in performance of these bearings which cannot be achieved with ceramics or coatings. Polymer linings can follow the substrate deformations caused by mechanical or thermal loading of the bearing and reduce the risk for delamination of the bearing lining from the substrate. Compliant liners can undergo elastic and/or plastic deformation in case an abrasive particle finds its way into the contact region of the bearing. These deformations can provide decreased contact pressure between the abrasive particles and the bearing surfaces and reduce the extent of the damage caused by the particle to the shaft. Another important characteristic of some polymers is their so called “self-lubricating” property which allows bearings to tolerate short periods of operation in marginally or starved lubrication conditions. Polymers exhibit many favorable properties for application as bearing lining materials; however some of the drawbacks include rather high wear rates, low thermal conductivity, viscoelastic deformation and water absorption which should be considered during material selection and design of bearings.

10 Introduction

1.3. Research Gaps

Tribological behaviour of polymers has been extensively studied in dry sliding contacts during the last decades. However, despite the significant potential of application of polymers in lubricated conditions relatively limited numbers of studies have been undertaken in this field and these are mainly directed towards biomedical applications. Tribological behaviour of polymers in lubricated conditions may significantly differ from dry contacts due to the effects arising from presence of lubricant. Lubricants may interact with the contacting surfaces in many different ways and alter the friction and wear behaviour of polymers. Adsorption and Absorption of lubricant by polymers, plasticization, thermal effects, tribo-corrosion of counter-surfaces and interference with build-up of transfer films are some parameters which may result in unpredictable tribological behaviour of a polymer in lubricated conditions compared to dry contact. Further research is required to bridge the knowledge gaps in tribology of polymers in dry and lubricated conditions. The potential of application of polymers in lubricated conditions to address some engineering issues necessitates further research to provide an in-depth understanding of the mechanisms involved in tribology of polymers in lubricated contacts.

1.4. Objectives of Present Work As mentioned earlier, application of polymers as bearing lining materials can potentially address the issues arising from frequently imposed transient operating conditions in hydroelectric power plants. Furthermore, application of polymers can provide the possibility of using water as a lubricant for bearing applications. The present work is thus aimed at investigating the following: - The potential of application of polymer composites to improve the tribological performance of bearings during the transient phases of start-up and stop. - The tribological behaviour of polymers in water to provide further understanding of the friction and wear mechanisms involved in a polymer/metal sliding contact in the presence of water. - The tribological response of polymers to variation in counter surface characteristics and operating conditions in water lubricated contacts.

Experimental Work

The experimental work carried out throughout these studies involved analysis of the tribological behaviour of polymeric materials in lubricated sliding contacts. However, the experiments were designed and tailored individually to fulfil the specific objectives of each study. Therefore, various test configurations, operating conditions and setups were used for these studies; the details of which are briefly described in this section.

Paper A) In this study, the breakaway friction characteristics of some commercially available PTFE based composite materials were studied sliding against steel plates under lubricated conditions. The tests were conducted over a wide range of contact pressures and temperatures to investigate the materials’ tribological behaviour with variations in pressure and temperature. Further tests were conducted to determine how breakaway friction was affected by extended periods of loading without operation in the presence of lubricant.

Materials Tests were performed using four commercially available PTFE-based composites together with pure PTFE and Babbitt material. Details of the PTFE-based composites are given in Table 2.1. The materials were chosen due to their availability and to develop upon earlier study of PTFE- based composites by McCarthy and Glavatskih [26]. The counter-surface was low carbon steel plate, ground and polished to a surface roughness of Ra=0.4 µm with roughness orientation parallel to the sliding direction. This is a typical surface roughness and orientation for counter surfaces in hydrodynamic sliding bearings. Lubricant used in all tests was commercially available synthetic ester based turbine oil. This lubricant was chosen as it is representative of the lubricants being used in new and renovated hydrodynamic turbine bearings.

12 Experimental Work

Table 2.1: Characteristics of PTFE-based composites

Compression modulus Specific Materials [GPa] density Virgin PTFE 0.46 2.16 PTFE+40% bronze 0.99 3.96 PTFE+25% carbon 0.85 2.15 PTFE+25% black Glass 0.66 2.19

PTFE+20% glass fibre+5% MoS2 0.78 2.28

Experimental setup The experiments were carried out using a TE77-Cameron-Plint tribo-meter with a reciprocating block on plate test configuration. A full description of the experimental arrangement is detailed in Table 2.2 and a diagram showing the test configuration is provided in Figure 2.1.

Figure 2.1: Diagram of block on plate test arrangement

Table 2.2: Experimental conditions for short term tests with pressure and temperature variations

Load 80-320N Contact pressure 1-8 MPa Oil bath temperature 25, 45, 65, 85 °C

Stroke length 5 mm Stroke duration 5 s Test duration 3 hr Total sliding distance 10.8 m Steel surface roughness (Ra) 0.4 µm

Testing for the dependence of break-away friction on contact pressure was conducted at 25 °C and tests for the dependence of break-away friction on temperature were conducted at 2MPa

Experimental Work 13 contact pressure. Additional tests were conducted at the maximum temperature and load conditions, 85 °C and 8MPa to determine whether, or not, temperature effects and contact pressure effects acted independently of each other. Tests were conducted three times each to ensure repeatability and minimize uncertainty. Further tests were conducted to determine how break-away friction was affected by extended periods of loading without operation in the presence of lubricant. For these tests, specimens were run-in to a steady state condition (10 min). They were then shifted to the start/stop point at the end of the stroke and left stationary in that position under load for 72h. The test was then restarted and the first several start–stop cycles were recorded to determine the breakaway friction for each cycle. These tests have been carried out with 2 MPa of loading at room temperature. The effects of the materials on topography of the steel counter surfaces were analyzed using an optical surface profilo-meter (WYKO NT1100). Roughness profiles were measured at the same points in the mid region of the wear track both before and after testing to determine the degree of polishing or roughening caused by the test. Further investigations were carried out utilizing a scanning electron microscope (SEM) to reveal the wear patterns and wear mechanisms involved; confirming the experimental findings regarding the suitability of the materials for the application.

Paper B) In this study, the tribological behaviour of several unfilled polymer materials sliding against 316L stainless steel in water lubricated contacts was studied using a uni-directional pin-on-disc tribo-meter. This configuration was chosen to avoid a converging gap at the contact surfaces in order to minimize the influence of hydrodynamic effects on materials’ tribological behaviour; a challenge faced in a block-on-ring configuration.

Materials In this work, eleven polymeric materials were studied namely ultrahigh molecular weight polyethylene (UHMWPE), polyoxymethylene (POM), polyethylene terephthalate (PET), polyamide 6 (PA 6), polyamide 66 (PA 66), polytetrafluoroethylene (PTFE), polypropylene (PP), polyvinylidene fluoride (PVDF), polyether ether ketone (PEEK), polycarbonate (PC) and

14 Experimental Work

polymethyl methacrylate (PMMA) along with lignum vitae which is a dense wood traditionally used as bearing material in water lubricated sliding contacts. The characteristics of the polymers are shown in Table 2.3. The selection of polymers was based on their availability, commercial application, and earlier studies in water lubricated contacts [27, 28].

Table 2.3: Characteristics of the polymer materials

Elastic Tensile Flexural Yield Elongation Melting Density Hardness Tg Material Modulus Strength Strength Strength at Break point [g/cm3] [MPa] [ºC] [GPa] [MPa] [MPa] [MPa] [%] [ºC] UHMWPE 0.93 0.8 23 40 23 >450 65 135 -160 [29] PET 1.38 3.4 90 * 90 50-80 170 225 76 [30] PP 0.91 1.3-2 33 * 30-32 700 60 175 -17 [30] POM 1.42 3 70 100 70 70-75 140 175 -75 [30] PTFE 2.15 0.4 25-36 18-20 30 400 30 300-310 127 [30] PEEK 1.30 * 70 * 100 50 M99 334 145 [30] PMMA 1.15 1.9 * * 46 5 * 230-260 107 PVDF 1.76 0.8 40 65 55-60 25/500 110R 170 -35 [30] PA6 1.15 1.5 50 40 50 200 70 220 -8 [31] PC 1.15 2.3 65 90 60 >80 100 230 150 [30] PA66 1.19 1.7 70 42 70 150 100 255 -6 [31] *Not available

The counter-surface in all experiments was AISI 316L, also known as marine grade stainless steel. The discs were polished to surface roughness of Ra=0.2±0.02 µm with a circular lay to simulate the lay orientation in relation to sliding direction in practical applications.

Experimental Setup The experiments were carried out using a polymer pin on stainless steel disc configuration. A schematic diagram of the test configuration is shown in Figure 2.2.

Load Water 316L Disc Polymer Pin

Figure 2.2: Schematic diagram of test configuration

Experimental Work 15

Table 2.4: Experimental Conditions Load 62.8 N Initial Contact Pressure 5 MPa Temperature R.T. (21-23 ºC) Sliding Speed 0.13 m/s Test Duration 20 h* Total Sliding Distance 9360 m* Steel Surface Roughness Ra 0.2 µm Lubricant Distilled Water *Except for polypropylene which was tested for 18 hours with total sliding distance of 8424 meters.

Tests were performed at room temperature (21-23 ºC), under an applied load of 62.8 N producing an initial apparent contact pressure of 5 MPa. This contact pressure was chosen to accelerate testing of the materials and to imitate the maximum apparent contact pressure at the loaded region of bearing. The experiments were done at constant sliding speed (0.13 m/s) considering the lowest practical rotational speed of the tribo-meter. These conditions were chosen to increase the interaction of the friction surfaces during sliding to allow characterization of the materials in the boundary/mixed lubrication regime. A full description of experimental conditions is detailed in Table 2.4. The wettability of the polymers was examined by contact angle measurements, using a 4 µL drop of distilled water deposited for 1 second on the polymer surface at room temperature (23 °C). Further investigations were carried out on worn surfaces and wear debris utilizing scanning electron microscope (SEM) to reveal the wear mechanisms involved in a polymer/metal contact.

Paper C) In this study the influence of shaft roughness and contact pressure on frictional behaviour of polymer bearings was studied.

Materials Tribological studies were carried out using four different unfilled thermoplastic polymers namely polytetrafluoroethylene (PTFE), polyether ether ketone (PEEK), polyethylene terephthalate (PET) and ultra-high molecular weight polyethylene (UHMWPE). The characteristics of the materials are listed in Table 2.5.

16 Experimental Work

Table 2.5: Characteristics of the polymer materials obtained from the materials supplier

Elastic Tensile Yield Elongation Density Hardness Melting Material Modulus Strength Strength at Break [g/cm3] [MPa] point [ºC] [GPa] [MPa] [MPa] [%]

UHMWPE 0.93 0.8 23 23 >450 65 135 PET 1.38 3.4 90 90 50-80 170 225

PTFE 2.15 0.4 25-36 30 400 30 300-310

PEEK 1.30 3.6 70 100 50 M99 334

Table 2.6: Limiting chemical composition of Inconel 625 Element Ni Cr Fe Mo Nb & Ta C Mn Si P S Al Ti Co Percentage > 58 20-23 < 5 8-10 3.15-4.15 < 0.1 < 0.5 < 0.5 < 0.015 < 0.015 < 0.4 < 0.4 < 1

The selection of the polymers was based on the results obtained from paper B. The shafts were made from Inconel 625 and its chemical composition is given in Table 2.6.

Test specimens The test shafts (Inconel 625) were machined from rods and polished to the desired surface roughness using various grades of SiC sand papers (#800P-#4000P). Surface roughness measurements were carried out using a stylus profilo-meter. The measurements were performed along the shaft periphery’s four quadrants in the axial direction. Table 2.7 shows the shaft roughness characteristics, averaged from four measurements. Plain polymer bearings were machined from polymer rods and press-fitted into stainless steel bushings (Fig. 2.3). The bearing linings were prepared with different wall thicknesses for each material, considering the large differences in elastic moduli of the polymers (0.4 - 3.6 GPa). This was done to achieve similar lining deformations and contact pressures in the loaded region of the bearings.

Figure 2.3: Machined polymer bearings

Experimental Work 17

Table 2.7: Shaft roughness characteristics, averaged from four measurements at the quadrantal angles Shaft Ra Rt Rz Rq Rpk Rvk Denotation [µm] [µm] [µm] [µm] [µm] [µm] 0.02R 0.02 0.30 0.19 0.03 0.02 0.05 0.1R 0.11 0.71 0.63 0.14 0.02 0.32 0.2R 0.22 1.17 1.01 0.25 0.03 0.69 0.4R 0.42 1.83 1.63 0.48 0.02 1.47

Table 2.8: Bearing dimensions, deformations and contact pressures at load of 141 N corresponding to bearing pressure of 0.3 MPa Specific Semi- Max Mean Wall Inner Radial Elastic Bearing Length Load bearing contact contact contact thickness diameter clearance Indentation lining [mm] [N] pressure width pressure pressure [mm] [mm] [µm] [µm] [MPa] [mm] [MPa] [MPa] PTFE 2 30.1 15 50 141 0.3 5.4 6.45 1.08 0.73 UHMWPE 3 30.1 15 50 141 0.3 4.5 6 1.18 0.79 PEEK 8 30.1 15 50 141 0.3 3.2 5.1 1.39 0.93 PET 8 30.1 15 50 141 0.3 3.2 5.1 1.36 0.91

Table 2.8 shows the polymer lining dimensions, deformations and contact pressures for different materials at specific bearing pressure of 0.3 MPa.

Experimental setup and procedure The experiments were carried out using a small-scale journal bearing test rig. The detailed description of the test rig can be found in [32]. Distilled water was supplied to the clearance of the bearing at atmospheric pressure and room temperature (~23 ºC) through a hole located opposite to the contact region of the bearing. The load was applied to the bearing housing using a lever arm loaded with mass. The friction force was continuously measured by a strain gauge force transducer coupled to the bearing housing. All the shafts and polymer bearings were ultrasonically cleaned with ethanol and dried in air prior to testing. Each bearing was run-in at sliding speed of 0.06 m/s (40 rpm) for four hours. Dynamic friction curves were obtained with variation in speed for each material and shaft roughness combinations at different time intervals during running-in.

18 Experimental Work

Table 2.9: Experimental conditions for running-in tests Radial Load 141 N Specific Bearing Pressure 0.3 MPa Rotational Speed 40 - 690 rpm Sliding Speed 0.06-1.08 m/s Time Intervals 10, 55, 115, 175, 235 min Test Duration 4 h Bearing Materials PTFE, PEEK, PET, UHMWPE Shaft Ra Roughness 0.02, 0.1, 0.2, 0.4 µm Lubricant Distilled Water Lubricant Temperature R.T. (23 ºC)

This allowed study of the influence of shaft roughness on the evolution of friction during the running-in period as the surfaces experienced physical/chemical changes. The experimental conditions for obtaining dynamic friction curves during running-in are shown in Table 2.9. Following the running-in of the bearings, breakaway and dynamic friction maps were obtained for each material, shaft roughness and pressure combination. This was achieved through intermittent motion of the shaft (start-stop cycles) at the lowest practical rotational speed of the tribo-meter. Figure 2.4 gives an example of the data stream from a typical intermittent test showing a number of start/stop cycles. Breakaway friction was determined from the maximum static friction occurring at the initiation of relative motion during each start/stop cycle. An average of one hundred peaks over the length of each test was used to determine the breakaway friction of the materials. A full description of the test conditions for the intermittent tests is provided in Table 2.10.

50 Breakaway 40 Friction

Friction Force [N] Dynamic 10 Friction

0 0 500 1000 1500 2000 2500 Data points One start/stop cycle Figure 2.4: A typical friction data stream obtained with start/stop cycles

Experimental Work 19

Table 2.10: Experimental conditions for intermittent tests for obtaining friction maps Load 141, 224, 336, 449, 562, 679 N Specific Bearing Pressure 0.3, 0.5, 0.75, 1, 1.25, 1.5 MPa Rotational Speed 40 - 690 (rpm) Sliding Speed 0.06-1.08 m/s Bearing Materials PTFE, PEEK, PET, UHMWPE Shaft Ra Roughness 0.02, 0.1, 0.2, 0.4 µm No. of Cycles 100 Cycle Duration 12 Sec Test Duration 20 min Lubricant Distilled Water Lubricant Temperature R.T. (~23 ºC)

Table 2.11: Experimental conditions to prepare the Inconel plates for contact angle measurements Load 20 N Temperature R.T. (21-23°C) Stroke length 5 mm Frequency 7 Hz Test Duration 4 hours Total Sliding Distance 1008 m lnconel Ra Roughness 0.03 µm Lubricant Distilled Water

At least three repetitions were carried out for each test combination throughout this study and the results were averaged from these test runs. Screening tests were carried out to investigate the possible changes in wettability of Inconel counter-faces when sliding against the polymeric materials. Due to the practical limitations for contact angle measurements on curved shaft surfaces, these measurements were carried out on flat Inconel plates which were pre-rubbed against the polymers in a simplified test configuration. Cylindrical polymer pins were rubbed against Inconel plates using a reciprocation Cameron-Plint tribo-meter. Prior to rubbing, the metal plates were polished to surface roughness (Ra) of 0.03 µm in order to minimize the effect of surface roughness on contact angle measurements [33]. The experimental conditions for preparation of Inconel plates for contact angle measurements are summarized in Table 2.11.

20 Experimental Work

Contact angle measurements were carried out on polymer and Inconel surfaces using distilled water. Sessile drops of 4µL were used and contact angles were recorded after one second of drop deposition on the surfaces. At least six measurements were performed for each sample and the results were averaged from these measurements.

Results

This section contains the main results obtained during the course of these studies. Detailed results and discussions can be found in the appended papers.

Breakaway friction of PTFE materials in lubricated conditions (Paper A) In this work, the breakaway friction characteristics of some commercially available PTFE based composite materials sliding against steel plates were studied under lubricated conditions. The tests were conducted over a wide range of contact pressures and temperatures to investigate the materials’ tribological behaviour with variations in pressure and temperature. The experimental setup was found to provide highly reproducible results, allowing for clear comparisons to be made between the performances of the various materials. The results of breakaway friction obtained with variations in contact pressure and oil temperature are shown in Figures (3.1-3.4). These results show that a considerably lower breakaway friction can be obtained with PTFE based materials in comparison to Babbitt; and this behaviour is consistent with variations in pressure and temperature.

0.25

0.20 Babbitt Black Glass +PTFE

0.15 MoS2/Glass fiber +PTFE

0.10 Bronze +PTFE Carbon +PTFE

Friction Coefficient Friction 0.05 Pure PTFE

0.00 1 2 3 4 5 6 7 8 Contact Pressure (MPa) Figure 3.1: Breakaway friction vs. contact pressure at 25 °C

22 Results

0.15

Babbitt 0.10 Black Glass +PTFE MoS2/Glass fiber +PTFE Bronze +PTFE 0.05 Carbon +PTFE Pure PTFE

Variation in Friction Coefficient Friction in Variation 0.00 1 2 3 4 5 6 7 8 Contact Pressure (MPa)

Figure 3.2: Variation in breakaway friction vs. contact pressure at 25 °C (µmax-µmin during the course of the tests)

0.30 Babbitt

Black Glass +PTFE 0.20 MoS2/Glass fiber +PTFE

Bronze +PTFE 0.10 Carbon +PTFE Friction Coefficient Friction Pure PTFE 0.00 25 45 65 85 Temperature (C)

Figure 3.3: Breakaway friction vs. oil bath temperature at 2MPa

0.15

Babbitt

0.10 Black Glass +PTFE

MoS2/Glass fiber +PTFE

Bronze +PTFE 0.05 Carbon +PTFE

Pure PTFE Variation in Friction Coefficient Friction in Variation 0.00 25 45 65 85 Temperature (C)

Figure 3.4: Variation in breakaway friction vs. oil bath temperature at 2MPa (µmax-µmin during the course of the tests)

Results 23

Pure PTFE consistently provided the lowest and most stable breakaway friction among the materials in all test conditions with carbon filled and bronze filled PTFE also providing low and stable breakaway friction. Glass filled PTFE composites did not perform well as the highest level of friction with largest variation was obtained with these materials in comparison to the rest of the PTFE based composites. Testing of the effects of an extended stop under load, such as in the practical case of a machine stop, was conducted with Babbitt, carbon filled, bronze filled, and pure PTFE. The poor performance of black glass and fiberglass filled PTFEs in the earlier testing disqualified these types of fillers from further testing so they were not included in the time consuming extended stop tests. Results of this series of testing are displayed in Figure 3.5 showing that Babbitt produced a much larger break-away friction after a prolonged stop than the PTFE based materials. The breakaway friction observed in the case of Babbitt was also much larger in proportion to the steady state friction measured prior to the prolonged stop. It is felt that the large increase in break-away friction for Babbitt was caused by the Babbitt material conforming to the steel counter-surface and ‘squeezing-out’ the lubricant from the contact. In the case of PTFE materials however reduction of lubricant in the contact is believed to allow PTFE to operate in the ranges of low friction that is known for PTFE based composites in sliding contacts. The effect of the materials on topography of steel counter surfaces was analyzed using an optical surface profilo-meter.

0.70 Babbitt 0.60 Bronze +PTFE 0.50 Carbon +PTFE 0.40 Pure PTFE 0.30 0.20 Friction Coefficient Friction 0.10 0.00 0 1 2 3 4 5 Cycle Figure 3.5: Breakaway friction before and after 72h pause under loading at 2MPa and 25 °C

24 Results

100 Babbitt 0 Black Glass +PTFE -100 MoS2/Glass fiber +PTFE Bronze +PTFE -200 Carbon +PTFE Pure PTFE Roughness change (nm) change Roughness -300

-400 1 MPa Tests 6 MPa Tests Figure 3.6: Change in roughness of steel counter surface

Figure 3.6 shows the results of roughness change in the middle of the wear track after the tests finding that only the glass filled PTFE based materials had a significant effect on counter surface at high loads. Although not statistically significant, similar trends could also be observed for glass filled composites at low loads. Investigations utilizing a scanning electron microscope revealed a number of tribological characteristics of the materials under sliding conditions and further confirmed the experimental findings regarding the suitability of the materials for the application.

SEM images of glass filled PTFE composites (black glass filled- and glass fiber/MoS2 filled- PTFE) revealed the abrasive nature of glass fillers clarifying the reason for the change in counter surface roughness when sliding against these materials. Figure 3.7.a shows the worn surface of black glass filled PTFE featuring sharp black glass fibers protruding from the polymer surface with collections of iron particles piled up in front of the fibers in the direction of sliding. In this figure, the darkest areas are PTFE matrix, the gray areas are fibers and the white areas were found to be rich in iron. Given that the PTFE material did not initially contain any iron, it can only be assumed that the iron detected was the result of wearing of the counter surface. In this case, it appeared that the PTFE did very little to reduce the friction in the contact as large portions of the load seemed to have ridden on the harder glass fibers.

Similar effects were also observed with the fiberglass and MoS2 filled PTFE. As in the case of black glass filled PTFE, the sharp glass fibers tended to abrade and polish the counter surface material. The worn surface of fiberglass and MoS2 filled PTFE however showed large areas of mixed MoS2 and iron wear debris as shown in Figure 3.7.b.

Results 25

a) b)

Figure 3.7: SEM images of worn a) black glass filled PTFE b) MoS2 and fiber glass filled PTFE

Although MoS2 appeared to have smoothed out the PTFE surface, it did not seem to have contributed to reduction of friction as the friction levels were generally the same or greater for

MoS2 and fiberglass filled PTFE than for black glass filled PTFE.

26 Results

Tribological behaviour of polymeric materials in water lubricated conditions (Paper B) This study is aimed at investigating the tribological behaviour of several unfilled thermoplastic polymers in water lubricated contacts to provide further understanding of friction and wear mechanisms involved at polymer/metal interface in the presence of water. The experiments were carried out using several unfilled thermoplastic polymers sliding against 316L stainless steel in water. The results were compared to that of lignum vitae; a dense wood conventionally used for application in water lubricated sliding contacts. The results of tribological tests allowed for clear comparison of the materials’ tribological performance. Results for the average friction coefficient over the entire test period of 20 hours are shown in Figure 3.8 and clearly indicate that PTFE consistently provided the lowest level of friction with UHMWPE also providing low friction.

PET, PVDF and PEEK provided friction coefficients of 0.16-0.21, statistically higher than those of PTFE and UHMWPE, however lower than that obtained with lignum vitae.

POM and PP exhibited similar friction to lignum vitae while the remaining materials, namely PC, PMMA, PA 6 and PA 66 provided higher friction compared to lignum vitae with PA6 and PA66 exhibiting the highest level of friction coefficient of all tested materials.

Analysis of the variation in friction coefficient over the course of the test demonstrated the stability or instability of the material’s friction performance during the tests.

0.45 PTFE 0.40 UHMWPE PET 0.35 PVDF 0.30 PEEK 0.25 POM 0.20 LV 0.15 PP PC Friction Coefficient 0.10 PMMA 0.05 PA 66 0.00 PA 6

Figure 3.8: Average friction coefficient over the test period of 20 hours.

Results 27

0.10 POM 0.09 PTFE PET 0.08 LV 0.07 PEEK 0.06 PVDF 0.05 PC 0.04 PA 66 0.03 PA 6 0.02 PMMA

Variation in FrictionVariation Coefficient in 0.01 UHMWPE 0.00 PP

Figure 3.9: Variation in friction coefficient calculated as the standard deviation of friction curve over the test period of 20 hours.

Figure 3.9 shows the standard deviation of the friction coefficient obtained with all materials during the entire test period of 20 hours.

In the case of POM, PTFE, PET, lignum vitae, PEEK and PVDF relatively small variation in friction during the test time was observed. However, it was a different case for the remainder of the materials which showed higher variation in friction coefficient along with large scatter.

Further investigations of wettability and solubility of polymers in water revealed the correlations of these characteristics to polymers’ frictional behaviour in presence of water.

In lubricated conditions, solubility of polymers can potentially provide information about the extent of the interaction of the polymer chains at the interface with lubricant. The latter can significantly influence the tribological behaviour of polymers in water as penetration of water molecules at the surface alters polymers’ affinity to water and influences the wettability characteristics and possibly surface free energy of the polymers.

Solubility of a polymer in a medium is characterized by its relative energy difference (RED) in regards to that medium. Figures 3.10 and 3.11 show the average friction coefficient of the materials versus water contact angle and relative energy difference (RED) of the polymers in regard to water respectively.

28 Results

0.45 PTFE 0.40 UHMWPE 0.35 PET 0.30 PVDF PEEK 0.25 POM 0.20 PP 0.15 PC

Friction Coefficient Friction 0.10 PMMA 0.05 PA 66 0.00 PA 6 50 60 70 80 90 100 110 120 Contact Angle [Degree]

Figure 3.10: Average friction coefficient over the test period of 20 hours vs. water contact angle of polymers

0.45 PTFE 0.40 PET

0.35 PVDF

0.30 POM 0.25 PP 0.20 PC 0.15

Friction Coefficient PMMA 0.10 PA 66 0.05 0.00 PA 6 3 5 7 9 11 Relative Energy Difference

Figure 3.11: Average friction coefficient over the test period of 20 hours vs. relative energy difference (RED) of polymers

In general, a trend of decreasing friction coefficient can be observed with increasing hydrophobicity and relative energy difference of polymers in regards to water. A careful look at Figures 3.10 and 3.11 shows that among the polymers with similar RED values (e.g.PA66 and POM, or PA6 and PC) those having a higher contact angle (POM and PC) exhibit lower friction coefficient and among the polymers with similar contact angles (e.g. POM and PVDF) the one having a higher RED value (PVDF) shows a lower friction coefficient. Considering the latter and the general trends observed in friction of polymers with respect to their water contact angle and

Results 29

8 UHMWPE 7 PET PP 6 POM 5 PTFE 4 PEEK 3 PVDF PA6 2 PC 1 Water Absorption[Weight %] PA66 0 PMMA 0 50 100 Time [Days]

Figure 3.12: Water absorption of polymers versus exposure time

relative energy difference it is inferred that the frictional behaviour of the polymers is inversely related to their water contact angle and relative energy difference (RED).

Due to practical significance of swelling of bearing materials, assessment of water absorption of polymers was carried out by long-term water absorption tests. Figure 3.12 shows the weight percentage water intake of polymers for an exposure period of 4 months. The results show the considerably larger and continuously progressing water absorption of polyamides in comparison to the rest of the materials. This is generally attributed to the presence of the amide groups in the molecular chain of polyamides [34, 35]; favouring water absorption by forming hydrogen bonds with water molecules [34].

Wear of the materials was calculated from weight loss measurement of the samples after each test. The measurements were corrected for the amount of water absorbed by samples during the test time. The results of the wear rates obtained for all tested materials are shown in Figure 3.13. UHMWPE consistently provided the lowest wear rate of all materials with PET and POM also providing low wear rates. The wear resistance of PVDF, PA 6 and PA 66 was similar to that of lignum vitae while PTFE, PC, PEEK, PMMA and PP provided higher wear rates with PP showing the highest wear rate of all tested materials in water lubricated contacts.

Further investigations utilizing scanning electron microscopy revealed the wear mechanisms of the materials in water.

30 Results

1.E-03 UHMWPE PET

/Nm] POM 3 LV mm 1.E-04 PVDF PA 6 PA 66 1.E-05 PTFE PC

Specific Wear Rate [ SpecificWear PEEK PMMA 1.E-06 PP

Figure 3.13: Specific wear rate of the materials. Error bars represent one standard deviation of the three repeated tests for each material.

Figure 3.14: SEM image of a) worn POM surface b) polypropylene wear debris c) PA6 wear debris d) worn stainless steel surface after sliding against polypropylene

Results 31

Analysis of worn surfaces and wear debris of PMMA, PEEK, PC, PTFE, PA 66, PVDF, PET and UHMWPE showed that abrasive wear is the dominant wear mechanism of these materials in water. Examination of worn POM surfaces showed mild abrasive wear with some regions undergoing plastic deformation and micro-cracking (Fig. 3.14.a).

The worn surface of PA6 and PP were however rather smooth and the wear debris featured a rolled morphology (Fig. 3.14.b and 3.14.c) suggesting that adhesive wear is the dominating wear mechanism of these materials when sliding against stainless steel in water. Some irregularly shaped wear debris could also be noticed in case of polypropylene which is believed to be caused by the abrading action of the stainless steel counter surface undergoing tribo-corrosion when sliding against polypropylene (Fig. 3.14.d). The highest wear rate of polypropylene among the materials was attributed to the combined action of adhesive wear and corrosion-induced abrasive wear.

Effect of Shaft Roughness and Pressure on Friction of Polymer Bearings in Water (Paper C)

In this study the frictional behaviour of some commercially available unfilled polymers namely PEEK, PTFE, PET and UHMWPE was investigated against Inconel shafts in water lubricated sliding contacts. Experiments were carried out to investigate the influence of shaft surface roughness on running-in behaviour and steady state friction of polymers in a journal bearing test configuration.

Furthermore, the influence of shaft surface roughness and pressure on frictional response of polymers was studied through intermittent tests by obtaining dynamic and breakaway friction maps for different materials, shaft roughness and pressure combinations. Figure 3.15 shows the dynamic friction behaviour of the materials at different time stages of running-in when sliding against shafts of Ra=0.02 µm and Ra=0.4 µm. In general, all the materials showed a trend of decreasing friction coefficient during the running- in period and this trend was consistent with shafts of different surface roughness values.

32 Results

0.25 0.25 PEEK PEEK Ra=0.02 µm Ra= 0.4 µm 0.20 0.20

0.15 10 min 0.15 10 min µ 55 min µ 55 min 0.10 115 min 0.10 115 min 175 min 175 min 0.05 235 min 0.05 235 min

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s] 0.30 0.30 PET PET 0.25 Ra=0.02 µm 0.25 Ra=0.4 µm

10 min 0.20 10 min 0.20 55 min 55 min 0.15 0.15 115 min µ 115 min µ 175 min 0.10 175 min 0.10 235 min 235 min 0.05 0.05

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s] 0.16 0.16 PTFE PTFE 0.14 Ra=0.02 µm 0.14 Ra=0.4 µm 0.12 0.12 0.10 10 min 0.10 10 min 55 min 55 min 0.08 0.08 µ 115 min µ 115 min 0.06 175 min 0.06 175 min 0.04 235 min 0.04 235 min 0.02 0.02 0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s] 0.10 0.10 UHMWPE UHMWPE Ra=0.02 µm Ra=0.4 µm 0.08 0.08

0.06 10 min 0.06 10 min µ 55 min µ 55 min 0.04 115 min 0.04 115 min 175 min 175 min 0.02 235 min 0.02 235 min

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

Figure 3.15: Dynamic friction behaviour of tested materials at different time stages of the running-in process (specific bearing pressure= 0.3 MPa, test duration=4 h, shaft Ra roughness=0.02 and 0.4 µm)

Results 33

5 Unworn

Worn by shaft Ra=0.02 µm 4 Worn by shaft Ra=0.1 µm

3 Worn by shaft Ra=0.2 µm

Worn by shaft Ra=0.4 µm 2 Ra Ra [µm] Roughness 1

0 PTFE UHMWPE PEEK PET

Figure 3.16: Average Ra roughness of bearing surfaces sliding against shafts with different roughness

This was mainly attributed to the reduced roughness of polymer bearings during running-in as all of the polymer materials except UHMWPE showed a considerable and significant reduction in surface roughness regardless of the shafts used. Figure 3.16 shows the average Ra roughness of the polymer bearings before and after sliding against shafts of different surface roughness values. The frictional behaviour of the polymer bearings was significantly influenced by shaft surface roughness during the running-in period. PEEK, PTFE and PET exhibited a lower initial friction followed by 50-75% reduction in friction during running-in when using the smoothest shaft. This is while only a marginal reduction in friction during running-in (approx. 10%) could be obtained with the roughest shaft. After the running-in period, frictional behaviour of the materials obtained with smooth and rough shafts was significantly different. Figure 3.17 shows the average dynamic friction curves obtained for each material after 4h of running-in with shafts of different surface roughness. Further investigations were carried out to examine possible changes in wettability of Inconel counter-faces when sliding against the polymeric materials as it is considered to play an important role in frictional behaviour of tribo-pairs. Contact angle measurements were carried out on polymer and Inconel surfaces using distilled water. Sessile drops of 4µL were used and contact angles were recorded after one second of drop deposition on the surfaces. At least six measurements were performed for each sample and the results were averaged from these measurements. Figure 3.18 shows the results of water contact angle measurements for unworn and worn Inconel plates and polymer materials.

34 Results

0.20 0.30 PEEK PET 0.16 0.25 0.20 0.12 0.02 µm 0.02 µm µ 0.1 µm 0.15 0.1 µm µ 0.08 0.2 µm 0.2 µm 0.10 0.4 µm 0.4 µm 0.04 0.05

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s] 0.16 0.10 PTFE UHMWPE 0.14 0.08 0.12

0.10 0.06 0.02 µm 0.08 µ µ 0.02 µm 0.1 µm 0.06 0.1 µm 0.04 0.2 µm 0.04 0.2 µm 0.4 µm 0.4 µm 0.02 0.02 0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

Figure 3.17: Dynamic friction behavior of materials after 4h of running-in (specific bearing pressure=0.3 MPa)

The results show a significant reduction in hydrophobicity of Inconel surfaces when sliding against the polymers in water. The considerably lower water contact angle of worn Inconel surfaces from those of unworn Inconel or polymers suggests formation of a reaction layer on Inconel surfaces rather than deposition of a physical polymer transfer layer.

100 95 90 85 Unworn Inconel 80 Unworn Polymer

Degrees 75 Inconel worn by polymer 70 65 60 PEEK PET PTFE UHMWPE

Figure 3.18: Water contact angles of unworn Inconel, unworn polymer and worn Inconel surfaces on the wear track

Results 35

Figure 3.19: Average dynamic and breakaway friction of polymer bearings versus specific bearing pressure and shaft surface roughness. The results are averaged from at least three test runs (300 start/stop cycles) for each combination. Error bars show one standard deviation at the measured points. The dynamic friction map of PET is not shown due to the large scatter in results.

36 Results

To investigate the influence of bearing pressure, shaft surface roughness and their interaction on frictional behaviour of polymers in water, breakaway and dynamic friction maps were obtained for all materials after the four hour running in period. Figure 3.19 shows the results of average breakaway and dynamic friction obtained with variations in specific bearing pressure and shaft surface roughness for each material. In general, a trend of decreasing friction could be observed with increasing contact pressure for all materials. This is while the materials’ frictional responses to variations in shaft surface roughness varied from one material to another. A trend of increasing friction was observed with increasing shaft roughness for PTFE whereas an opposite trend was obtained with PEEK at breakaway. No significant dependence of friction on shaft roughness was found with UHMWPE and PET. The various trends observed in friction of the polymers with variations in shaft roughness may be due to the contribution of adhesive and deformation components of friction for each material considering the vast differences in elastic modulus and surface free energy of the polymers studied. These results indicate that although using a shaft of reduced surface roughness can decrease the dynamic friction of polymer bearings; it may adversely affect the materials’ frictional response at breakaway and increase the torque required for machine start-up; a critical issue in some applications such as pumped storage hydropower plants.

Conclusions

In this work, extensive tribological studies on several polymeric materials having potential for application as bearing lining material for hydropower applications under lubricated conditions have been carried out. The main conclusions of this work are as follows:

- The breakaway friction can be significantly reduced using PTFE based materials in comparison with Babbitt; this behaviour being consistent with variations in contact pressure and oil temperature. - The glass filled polymeric materials, in view of the abrasive nature of glass fillers, are not considered suitable for application as bearing lining material. - The increase in breakaway friction due to prolonged machine stop can be significantly reduced by using PTFE based materials. This can eliminate the need for utilizing hydraulic jacking system for turbine startup after prolonged machine stop. - Wettability and solubility of polymers appear to correlate with their frictional behaviour in presence of water as an increased contact angle and relative energy difference with regard to water result in lower friction. - In view of low water absorption, comparable friction to that of PTFE, low wear rate and marginal effect on counter surface, UHMWPE exhibits promising characteristics for further development as bearing lining material for water lubricated applications. - Dynamic friction of polymers can be significantly reduced using a shaft of reduced surface roughness. However, depending on the bearing material, this can adversely affect the material’s frictional response at breakaway.

Future Work

Further experimental work should be carried out to investigate the possibility of enhancement of tribological behaviour of UHMWPE by incorporating nano-reinforcements. While assessment of the materials’ tribological response is generally carried out using non-aged samples, in real applications, the materials mechanical characteristics and tribological behaviour may vary during the lifespan of a bearing due to polymer/water interactions and the effects arising from oxidation of polymers. Therefore the influence of hygrothermal aging on oxidation behaviour and tribological response of polymers in water lubricated sliding contacts should be studied and the results should be compared to those of commercially available polymers used for water lubricated sliding bearings. As mentioned earlier, although a reduced counter surface roughness can favorably decrease the dynamic friction of a polymer; it may adversely affect the materials’ frictional response at breakaway and result in increased breakaway torque during machine start-up; a critical issue in some applications such as pumped storage hydropower plants. Due considerably lower deformation component of friction of polymer when sliding against smooth shafts, the choice of the counter surface material can significantly influence the frictional behaviour of polymers through its key role in determining adhesion forces between the interacting bodies. Further studies should therefore be carried out to investigate the influence of the counter surface material and its significance on frictional behaviour of polymers in the presence of water. Most studies on friction and wear of polymers are carried out in contaminant-free environments. However, in real application abrasive particles may find their way into the contact region of the bearings and alter friction and wear mechanisms of polymers as bearing lining material. Further studies are therefore also required to investigate the influence of abrasive particles on three-body abrasive wear of polymers in water lubricated conditions.

References [1] Åstrand, S., Miljöeffekter av turbinoljeläckage från vattenkraftverk till älvar, Master thesis, 2008, http://www.w-program.nu/filer/exjobb/Stina_%C3%85strand.pdf [2] Tilting pad bearings, http:// www.marunda.com.sg [3] Glavatskih S.B. and Paramonov G.A., 20 years of experience with a PTFE-faced tilting pad bearing operating at 11 MPa thrust load, HYDRO, 2007. [4] Cha M., Nonlinear Isoviscous Behaviour of Compliant Journal Bearings, Licentiate thesis, KTH Royal Institute of Technology, 2012, http://kth.diva-portal.org [5] Smith, W., Material Selection Criteria for Water Lubrication, Wear, Vol. 25, Issue 2, Aug. 1973, pp. 139-153. [6] Lancaster, J. K., Lubrication of Carbon Fibre-Reinforced Polymers part I—Water and Aqueous Solutions, Wear, Vol. 20, Issue 3, July 1972, pp. 315-333. [7] Higgenbottom, Adrian, Coastguard Non-Polluting Sterntube Sealing System, RINA International Conference for the Design and Operation of Container Ships, 23-24 April 2003, London, UK , pp. 53-60. [8] Stern tube bearing, http://www.seaquipment.com [9] McLaren K. G., Tabor D., The Frictional Properties of Lignum Vitae, British Journal of Applied Physics, Mar. 1961, Vol. 12, No 3, pp. 118-120. [10] Lignum-vitae pads, http:// www.lignum-vitae.com [11] Orndorff, R., New UHMWPE/Rubber Bearing Alloy, Journal of Tribology, Vol. 122, Issue 1, Jan. 2000, pp. 367-373. [12] Romor bearing, http:// www.naval-technology.com [13] Vesconite bearing, http:// www.vesconite.com [14] Thordon SXL, http:// www.thordonbearings.com [15] Andresson P., Lintula P., Load-carrying capability of water-lubricated ceramic journal bearings, Tribology International, Oct 1994, Vol. 27, No. 5, pp. 315-321. [16] Chen M., Kato K., Adachi K., Friction and Wear of Self-Mated SiC and Si3N4 Sliding in Water, Wear, Oct 2001, Vol. 250, Issues 1-12, pp. 246-255. [17] Chen M., Kato K., Adachi K., The-comparisons-of-sliding-speed-and-normal-load-effect- on-friction-coefficients-of-self-mated-Si3N4-and-SiC-under-water-lubrication, Tribology International, Mar. 2002, Vol. 35, No. 3, pp. 129-135. [18] Andersson P., Nikkila A., Lintula P., Wear-characteristics-of-water-lubricated-SiC-journal- bearings-in-intermittent-motion, Wear, Vol. 179, No. 1-2, Dec. 1994, pp. 57-62. [19] Wang X., Kato K., Adachi K., Running-in Effect on the Load-Carrying Capacity of a Water-Lubricated SiC Thrust Bearing, Journal of Engineering Tribology, 2005, Vol. 219, Part J, pp. 117-124. [20] Li J. F. et al., Tribological Properties of Silicon Carbide under Water-Lubricated Sliding, Wear, July 1998, Vol. 218, No. 2, pp. 167-171.

[21] Chen M., Kato K., Adachi K., The Difference in Running-in Period and Friction Coefficient between Self-mated Si3N4 and SiC under Water Lubrication, Tribology Letters, 2001, Vol. 11, No. 1, pp. 23-28. [22] Tomizava H., Fischer T., Friction and Wear of Silicon Nitride and Silicon Carbide in Water: Hydrodynamic Lubrication at Low Sliding Speed obtained by Tribochemical Wear, ASLE Transactions, 1987, Vol. 30, No. 1, pp. 41-46.

[23] Yamamoto K., Sato T., Takeda M., Structural analysis of (Cr1−xSix)N coatings and tribological property in water environment, Surface and Coatings Technology, Apr. 2005, Vol. 193, No. 1-3, pp. 167-172. [24] Neerinck, D. et. al., Diamond-Like Nano-Composite Coatings for Low-Wear and Low- Friction Applications in Humid Environments, Thin Solid Films, Vol. 317, Issues 1-2, Apr. 1998, pp. 402-404. [25] Nakahigashi, T. et al., Properties of Flexible DLC Film Deposited by Amplitude-Modulated RF P-CVD, Tribology International, Vol. 37, Issues 11-12, Nov.-Dec. 2004, pp. 907-912. [26] McCarthy DMC, Glavatskih SB. Assessment of polymer composites for hydrodynamic journal-bearing applications. Lubrication Science, 2009, Vol. 21, Issue 8, pp. 331–41. [27] Clarke CG and Allen C., The water lubricated sliding wear behaviour of polymeric materials against steel, Tribol. Int., 1991, Vol. 24, pp. 109-118. [28] Mens JWM., Friction and wear behaviour of 18 polymers in contact with steel in environments of air and water, Wear, 1991, Vol. 149, pp. 255-268. [29] Kurtz SM., UHMWPE Biomaterials Handbook - Ultra-High Molecular Weight Polyethylene in Total Joint Replacement and Medical Devices, 2nd ed. Elsevier, 2009. [30] Subramanian MN., Basics of troubleshooting in plastics processing - An introductory practical guide, Wiley – Scrivener, 2011. [31] Ehrenstein GW., Polymeric materials: structure, properties, applications. Hanser Verlag, 2001. [32] Peels J.A., Meesters C.J.M., Ontwerp en constructieve aspecten van een 2kN- glijlageropstelling, De constructeur, 1996, No. 3, pp. 40-43. [33] Busscher H. J., Van Pelt A.W. J., De Boer P., De Jong H. P., Arends J., The Effect of Surface Roughenning of Polymers on Measured Contact Angles of Liquids, Colloinds and Surfaces, 1984, Vol. 9, pp. 319-331. [34] Rajesh JJ, Bijwe J, Venkataraman B, et al., Effect of water absorption on erosive wear behaviour of polyamides, Journal of Material Science, 2002, Vol. 37, pp. 5107-5113. [35] Chen Z, Li T, Yang Y, et al., Mechanical and tribological properties of PA/PPS blends, Wear 2004, Vol. 257, pp. 696-707.

Break-away Friction of PTFE Materials in Lubricated Conditions

A. Golchin, G. F. Simmons, S. Glavatskih

Tribology International, Vol. 48, April 2012, pp. 54–62

Break-away friction of PTFE materials in lubricated conditions

Arash Golchina, Gregory F. Simmonsa, Sergei B. Glavatskiha,b

a Division of Machine Elements, Department of Applied Physics and Mechanical Engineering, Luleå University of Technology, Luleå 97187, Sweden b Department of Mechanical Construction and Production, Ghent University, Sint-Pietersnieuwstraat 41, 9000 Gent, Belgium

Abstract This study investigates the tribological characteristics at initiation of sliding (break-away friction) of several polytetrafluoroethylene based materials. Four PTFE composites, pure PTFE, and white metal were tested in a reciprocating tribo-meter with the block on plate configuration against a steel counter-surface. Apparent contact pressure and oil temperature were varied from 1 to 8 MPa and 25 to 85 °C respectively. SEM investigations revealed wear patterns of the PTFE materials and the abrasive nature of hard fillers. Bronze filled, carbon filled and pure PTFE were found to provide lower break-away friction and less variation over the course of testing and generally superior properties.

Keywords: Friction; PTFE; Babbitt; Break-away

A.1. Introduction bearing pads and lift the machine prior to start-up. Sliding bearings used in most large power Since its invention in the 1930s, countless generation machines are designed to studies have found PTFE to provide low operate in the fully hydrodynamic regime friction in dry sliding [1]. However, PTFE with a thick oil film separating the sliding is also associated with some of the highest and stationary surfaces. In the full film wear rates among crystalline polymers in regime, these bearings provide low friction dry contacts [2]. This deficiency has led to and extremely long service life. White the use of fillers to improve the mechanical metal (Babbitt) materials have traditionally and wear properties of the PTFE matrix provided acceptable performance. and has been widely documented by However, changes in electricity markets Bahadur and Tabor [3], Briscoe et al. [4] and the introduction of variable power and Xue et al. [5]. A review of work sources have resulted in more frequent related to polymers, including PTFE, with starts and stops of power generating nano-particle fillers is provided by machines. Because the Babbitt material Friedrich et al. [6]. Gong et al. [7] found currently in use is not optimum for these that the choice of the counter-surface conditions due to its potential for being material's chemical properties had no effect damaged by seizure at start-up, hydraulic on the wear rate of PTFE. Presumably, jacking systems are often used to flood the wear of the PTFE is primarily caused by

45 the roughness of the counter-surface as journal bearing applications [17] showed mentioned by Akagaki et al. [8]. A that PTFE-based composites performed comparison of tribo-testing techniques for better than Babbitt in lubricated conditions. polymer materials was accomplished by Pure PTFE has long been used as a bearing Samyn et al. [9] finding that the selection surface material in large hydrodynamic of test arrangement significantly affected bearings [18]. the friction and wear rates in dry Investigating start-up in sliding bearings, it conditions. was found that journal bearings rapidly Little work has been accomplished in become hydrodynamic following an initial regards to PTFE's characteristics in moment of break-away friction [19]. While lubricated conditions. Zhang et al. [10], wear of the bearing was observed after [11] and [12] found that the friction and these tests, it was deemed that the bearing wear of PTFE were greatly reduced by continued to function acceptably. On the adding lubrication. This result can other hand industrial experience of bearing probably be explained by the high sliding failures due to wiping or failure of the speeds used during the tests, 1.5–4 m/s, bearing's Babbitt surface has led to the which could have allowed for widespread use of hydrostatic jacking hydrodynamic lubrication. A recent study systems during machine start-up. on wear rates in lubricated conditions [13] Reduction of the break-away friction by of PTFE composites found a marked changing to low friction and gently decrease in wear rate of PTFE materials wearing bearing materials has potential to filled with MoS2 and fiberglass, black simplify machine start-up while also glass, bronze, and carbon. However, controlling the risk of bearing failure. Akagaki et al. [8] and Zhang et al. [10] This work seeks to address the issue of found that PTFE-based composites exhibit break-away friction in hydrodynamic different tribological behavior compared to bearings through focused testing of bearing pure PTFE material with varying pressures materials in a simplified configuration. and temperatures in lubricated conditions. This configuration is designed to isolate Journal bearing configurations made of a the transition from static to dynamic variety of polymers were tested in dry friction at machine start-up by eliminating conditions by Ünlü et al. [14]. These tests hydrodynamic effects which become more showed that PTFE provided the lowest dominant as speed increases. coefficient of friction and some of the highest wear rates of the materials tested. Further journal bearing tests with both A.2. Experimental study carbon composite and bronze composite PTFE in dry conditions conducted by The break-away friction characteristics of Tevrüz [15] and [16] showed that in both some commercially available PTFE-based cases the friction coefficient decreased composite materials sliding against steel with increasing contact pressure while the plates under lubricated conditions are friction coefficient increased with studied using a reciprocating block on increasing temperature. plate configuration. These tests are Work accomplished to investigate PTFE- conducted over a wide range of contact based composites for hydrodynamic pressures and temperatures. Sliding speeds

46 in all tests are kept extremely slow to used in new and renovated hydrodynamic eliminate the possibility of hydrodynamic bearings in large machines. effects. The results are compared with those of pure PTFE and Babbitt material. Table 2: Test lubricant characteristics Density at 40 °C 0.92 kg/l A.2.1. Materials Density at 100 °C 0.86 kg/l Kinematic viscosity at 40 °C 31 mm2/s Testing was performed using four Kinematic viscosity at 100 °C 6.2 mm2/s commercially available PTFE-based composites together with pure PTFE and Babbitt material. Details of the PTFE- A.2.2. Experimental setup based composites are given in Table 1. Compression modulus was determined on- Experiments were carried out using a site for comparison of the materials to each reciprocating tribo-meter under lubricated other. The materials have been chosen due conditions in the block on plate to their availability and to develop upon configuration. This configuration was earlier study of PTFE-based composites by chosen to eliminate a converging gap McCarthy and Glavatskih [13]. Babbitt which could lead to hydrodynamic specimens were removed from a journal lubrication and to ensure that contact bearing sleeve and consist of a steel pressure remained constant with varying backing with a minimum of 0.5 mm of wear rates of the materials. Providing a aluminium–tin (AlSn40Cu1) coating on constant contact is a common challenge in the bearing surface. block on ring studies of soft materials against hard materials. The conditions are Table 1: PTFE composite material characteristics nearly identical to those in hydrodynamic Compression Specific Filler material modulus thrust bearings at start-up, while because of density (GPa) the small scale the conditions are similar to 40% Bronze 0.99 3.96 those in large hydrodynamic journal 25% Carbon 0.85 2.15 bearings at start-up. The steel counter- 25% Black glass 0.66 2.19 surfaces were fixed in a temperature 20% Glass fiber and 5% MoS 0.78 2.28 2 controlled oil bath while the block Pure virgin PTFE 0.46 2.16 specimens were fastened to the reciprocating arm. Reciprocating motion The counter-surface in all testing was low was provided by an eccentric transmission carbon steel plate ground and polished to a driven by an electric motor which was surface roughness of 0.4 µm Ra with geared down through a gearbox. Loading structure parallel to the sliding direction was accomplished via a spring connected which is a typical surface roughness and to the reciprocating arm and was varied to direction for counter-surfaces in provide specified contact pressures. hydrodynamic bearings. Lubricant used in Friction force measurements were taken all testing was commercially available continuously using a piezoelectric load synthetic ester based turbine oil as detailed cell. A full description of the experimental in Table 2. This lubricant was chosen as it arrangement is detailed in Table 3 and a is representative of the lubricants being diagram showing the test configuration is

47 provided in Fig. 1. Tests were conducted end of the stroke and left stationary in that three times each to ensure repeatability and position under load for 72 h. The test was minimize uncertainty. then re-started and the first several start– Testing for the dependence of break-away stop cycles were recorded to determine the friction on contact pressure was conducted break-away friction for each cycle. at 25 °C and tests for the dependence of break-away friction on temperature were A.2.3. Break-away friction conducted at 2 MPa contact pressure. Additional tests were conducted at the Break-away friction is determined from the maximum temperature and load conditions, friction load data stream provided by the 85 °C and 8 MPa, to determine whether, or test rig. The value of break-away friction is not, temperature effects and contact taken from the maximum value of each pressure effects acted independently of stroke occurring when the reciprocating each other. test specimen goes through the stop and Further tests were conducted to determine start at the end of each stroke. Fig. 2 gives how break-away friction was affected by an example of the raw data stream from a extended periods of loading without typical test of carbon filled PTFE showing operation in the presence of lubricant. For a number of starts and stops with steady these tests, specimens were run-in to a break-away friction. An average of these steady state condition (10 min). They were peak values over the length of the test is then shifted to the start–stop point at the used to determine the average break-away friction for the materials. The lack of symmetry over each stroke of the friction Table 3: Experimental conditions for short term curve results from the flexing of the tests with pressure and temperature variation materials and lubricant presence in the Load 80–320 N contact. During the stop, the surfaces settle Contact pressure 1–8 MPa against each other as lubricant is squeezed Oil bath temperature 25, 45, 65, 85 °C out of the contact, leading to the dramatic Stroke length 5 mm break-away peak. However, because Stroke duration 5 s lubricant is fed into the contact all the way Test duration 3 h until the stop, there is no friction peak at Total sliding distance 10.8 m the end of the stroke. This effect is Steel surface roughness Ra 0.4 µm increased by the flexing that occurs in materials with lower elastic modulus, such as PTFE. Babbitt displayed much more symmetry over the course of each stroke with a lower peak at the end of the stroke. The slight difference in the shape of the positive and negative portions of the curve is the result of the geometry of the test rig arrangement. The load, which is provided

by a spring and beam, varies slightly over Figure 1: Diagram of block on plate test one test cycle by approximately 1–2% of arrangement the total load value.

0.10 0.08 0.06 0.04 0.02 0.00 -0.02 -0.04 Friction Coefficient Friction -0.06 -0.08 -0.10 0 20000 40000 60000 Data Points

Figure 2: Friction curve over approximately 30 s from testing of carbon filled PTFE at 2 MPa and 25 °C

Running-in effects are included in the while for the rest of the loading conditions average friction results and are taken into the contact surface dimensions were 5 mm account in terms of the variation in the parallel to and 8 mm perpendicular to the break-away friction through the duration of sliding direction. This made it possible to the test. This method was used because achieve apparent contact pressures of 1–8 some of the materials did not provide a MPa. The block specimens were cleaned in steady state friction after running-in. an ultrasonic bath using industrial Further supporting this decision is the petroleum (heptane) followed by ethyl application of this test program to alcohol and dried in air prior to each test. hydrodynamic sliding bearings which only experience break-away friction at start-up of the machine. With the exception of A.3. Results and discussion pumped storage machines, these bearings typically experience less than one start per The experimental setup was found to day over the course of their several year produce highly reproducible results, life-cycle. Thus any variation of break- allowing for clear comparisons to be made away friction, including changes during between the performance of the various running-in, can have significant materials. In general, each of the test consequences for machine operation. materials followed one of three different trends. The first of these trends was that of A.2.4. Material preparation Babbitt which provided the highest levels of break-away friction. This high friction The polymer blocks were prepared in two phenomenon reflects typical industrial different dimensions to achieve desired applications of Babbitt bearings which contact pressures with the practical load generally limit mean contact pressures to range of the test rig which is in the range 2.5 MPa without the use of hydrostatic of 80–400 N. The dimensions of the lifting systems at startup. Another trend contact surface of block specimens for 1 was displayed by the materials with glass MPa tests were 5 mm parallel to and 16 and fiberglass fillers. These materials mm perpendicular to the sliding direction, tended to wear the counter-surfaces,

0.25

0.20 Babbitt Black Glass +PTFE

0.15 MoS2/Glass fiber +PTFE

0.10 Bronze +PTFE Carbon +PTFE

Friction Coefficient Friction 0.05 Pure PTFE

0.00 1 2 3 4 5 6 7 8 Contact Pressure (MPa)

Figure 3: Break-away friction vs. contact pressure at 25 °C. Error bars represent one standard deviation of three repeated tests for each point

0.15

Babbitt 0.10 Black Glass +PTFE MoS2/Glass fiber +PTFE Bronze +PTFE 0.05 Carbon +PTFE Pure PTFE

Variation in Friction Coefficient Friction in Variation 0.00 1 2 3 4 5 6 7 8

Contact Pressure (MPa)

Figure 4: Variation in break-away friction vs. contact pressure at 25 °C.

leading to decreased roughness and higher showed a slightly decreasing trend in levels of friction, but not as high as friction coefficient with increasing contact Babbitt. The final response that was pressure. This is consistent with empirical observed was that of the bronze and carbon formulas proposed by Zhang et al.'s [10] filled PTFEs, as well as pure PTFE, which findings in regards to PTFE composites in did not appear to wear the counter-surface lubricated conditions. The exception to this and provided low friction with little was fiberglass and MoS2 material which variation due to temperature or loading. did not show a consistent trend with varied loading. A.3.1. Contact pressure effects Results for the average break-away friction over the entire test period of 3 h are shown Results of break-away friction in relation in Fig. 3 and clearly indicate that pure to variation in contact pressure generally PTFE consistently provided the lowest

0.30 Babbitt Black Glass +PTFE 0.20 MoS2/Glass fiber +PTFE Bronze +PTFE 0.10 Carbon +PTFE Friction Coefficient Friction Pure PTFE 0.00 25 45 65 85 Temperature (C)

Figure 5: Break-away friction vs. oil bath temperature at 2 MPa. Error bars represent one standard deviation of three repeated tests for each point

0.15

Babbitt

0.10 Black Glass +PTFE

MoS2/Glass fiber +PTFE

Bronze +PTFE 0.05 Carbon +PTFE

Pure PTFE Variation in Friction Coefficient Friction in Variation 0.00 25 45 65 85 Temperature (C)

Figure 6: Variation in break-away friction vs. oil bath temperature at 2 MPa.

levels of friction with bronze filled PTFE Analysis of the variation in break-away also providing low friction. Carbon filled friction over the course of the test cycle PTFE was found to provide lower friction provides insight into a material's behavior at higher loads than at lower loads while in an actual application. In the case of the black glass filled PTFE and fiberglass materials tested in this study, a stark filled PTFE provided highest break-away contrast could be seen in the response of friction of the PTFE composites. All PTFE the materials to variations in pressure. Fig. based materials provided significantly 4 shows the difference between the lower friction than Babbitt and these minimum and maximum break-away friction levels were generally 25–50% friction observed for materials during the those of Babbitt. course of each test. This demonstrates the

51 stability or instability of the materials' contact with the counter-surface. The other break-away friction when subjected to PTFE composites appeared to have very increased loading. In the case of carbon little change in their average break-away filled and pure PTFE, relatively little friction values with increased temperature. difference was observed between the When variation in friction was analyzed minimum and maximum break-away with increasing temperature, significant friction. It was quite a different case for the trends were observed, Fig. 6. The remainder of the materials which had large fiberglass filled PTFE had inconsistent levels of variation in break-away friction variation in friction coefficient while with changes in loading and no consistent bronze filled PTFE demonstrated the trends in the variation observed during the opposite trend with a decrease in variation tests. in break-away friction at increasing temperatures coupled with a decrease in A.3.2. Temperature effects the average friction. Black glass filled PTFE had slightly decreasing variation in Results of break-away friction in relation friction coefficient with increasing to changes in temperature in the oil bath temperature. Variation in break-away are displayed in Fig. 5. It was found that friction for carbon filled and pure PTFE at pure PTFE and bronze filled PTFE increased temperature was marginal as produced the lowest friction as temperature both materials provided very stable break- increased with carbon filled PTFE also away friction levels through the entirety of providing low friction. As in testing with the test range. pressure variation, fiberglass filled PTFE yielded higher coefficients of friction than A.3.3. Pressure and temperature the other PTFE composites while Babbitt interaction provided much higher break-away friction Testing for the interaction between than all of the other materials. Black glass temperature and pressure effects was filled PTFE approached carbon filled conducted for all materials except black PTFE at higher temperatures but had glass and fiberglass filled PTFEs at 85 °C equivalent friction to fiberglass filled and 8 MPa loading. Analysis of variance of PTFE at lower temperatures. The these results confirmed the influences of fiberglass filled PTFE demonstrated an temperature and pressure on the break- increasing trend in break-away friction away friction for all materials. However, with increasing temperature which was no interaction was found between the most likely a result of the decreasing effects of changes in temperature and lubricant viscosity which could have changes in pressure meaning that changes allowed for the hard fibers to come more in temperature can be considered to act into contact with the counter-surface. A independently of changes in contact similar effect is believed to be the cause of pressure and vice versa. the increase in friction observed for Babbitt with increased temperature. It is A.3.4. Prolonged stop and restart believed that the lower viscosity lubricant more readily squeezed from the contact, Testing of the effects of an extended stop allowing the Babbitt to have greater area in under load, such as in the practical case of

0.70 Babbitt 0.60 Bronze +PTFE 0.50 Carbon +PTFE 0.40 Pure PTFE 0.30 0.20 Friction Coefficient Friction 0.10 0.00 0 1 2 3 4 5 Cycle

Figure 7: Break-away friction before and after 72 h pause of testing under loading at 2 MPa and 25 °C.

a machine stop, was conducted with conforming to the steel counter-surface and Babbitt, carbon filled, bronze filled, and ‘squeezing-out’ the lubricant from the pure PTFE. The poor performance of black contact. In the case of PTFE materials, glass and fiberglass filled PTFEs in the reduction of lubricant in the contact is earlier testing disqualified these types of believed to allow PTFE to operate in the fillers from further testing so they were not ranges of low friction that it is known for included in the time consuming extended in dry contacts while the lubricant that was stop tests. Results of this series of testing not ‘squeezed-out’ of the contact could are displayed in Fig. 7. The value at the ‘0’ have been sealed under pressure within the cycle is the average break-away friction contact similar to the mechanism proposed from the 1 min prior to the stop while ‘1’, by Persson et al. [20] for the case of ‘2’, etc. are the first, second and vehicle tires on asphalt. subsequent cycles after the extended stop. One of the most notable features of these A.3.5. Counter-surface wear results is that the maximum friction returns to values measured prior to the prolonged The effects the materials had on the steel stop after just one stroke, regardless of the counter-surface were analyzed using an material. However, in regards to the optical profilometer (WYKO NT1100). materials, it was observed that Babbitt Roughness average (Ra) was measured at produced a much greater break-away the same points in the center of the wear friction after a prolonged stop than the track both before and after testing to PTFE based materials. The break-away determine the degree of polishing or friction observed in the case of Babbitt was roughening caused by the test. These also much greater in proportion to the results are displayed in Fig. 8 showing that steady state friction measured prior to the only the glass filled PTFE based materials prolonged stop. It is felt that the large had a significant effect on the counter- increase in break-away friction for Babbitt surface at high load. Error bars in the was caused by the Babbitt material figure represent the accuracy of the

100 Babbitt 0 Black Glass +PTFE -100 MoS2/Glass fiber +PTFE Bronze +PTFE -200 Carbon +PTFE Pure PTFE Roughness change (nm) change Roughness -300

-400 1 MPa Tests 6 MPa Tests

Figure 8: Change in roughness of steel counter-surface

measurement method and error associated A.3.6. SEM investigation with comparison of samples before and after testing. Given the significant surface Further investigations utilizing a scanning polishing provided by these materials at electron microscope revealed a number of the 6 MPa load level it is felt that the trend unique tribological characteristics of the observed at 1 MPa loading also represents materials under break-away sliding. polishing by the glass-based materials even Overall however, SEM examination further though the difference is very near to the confirmed the experimental findings uncertainty of the measurement method. regarding suitability of materials for the Observations by Iwai et al. [17] under application. The sliding direction in all constant sliding conditions with similar SEM images is approximately in the fiberglass filled material also found vertical direction. wearing of the counter-surface with the exception that counter-surface roughness increased. The difference in these observations from the current study can be explained by the testing configurations. The reciprocating action of the current tests likely trapped wear particles in the contact region, breaking them into a media which polished both surfaces instead of removing them from the contact as in the block on ring configuration. This wearing effect made the black glass Figure 9: SEM image of Babbitt tested at 6 MPa and fiberglass filled composites unsuitable and 25 °C for the final application in hydrodynamic bearings and so they were not included in A.3.6.1. Babbitt the, time consuming, long-term break- SEM evaluation of the worn Babbitt away testing. surface revealed a significant wear scar

54 with grooves cut into the material by the steel counter-surface. A representative image of the worn Babbitt surface is displayed in Fig. 9. Counter-surface material was not detected on the worn surface signifying that wearing occurred on the Babbitt material only. Similar wear scars were seen on the Babbitt specimens for both high and low loads regardless of temperature.

A.3.6.2. Black glass filled PTFE Figure 10: SEM image of black glass filled PTFE Investigation of the black glass filled PTFE tested at 8 MPa and 25 °C revealed the reason for the change in surface roughness observed on the counter- surfaces. The fibers of the black glass had in many locations become sharpened cutting surfaces with collections of iron particles piled up in the direction of sliding. This effect is clearly demonstrated in Fig. 10 in which the darkest areas are PTFE, the gray areas are fibers and the white areas were found to be rich in iron. Given that the PTFE material did not initially contain iron, it can only be assumed that the iron presence was the result of wearing on the counter-surface. In Figure 11: SEM image of MoS2 and fiberglass this case, it appears that the PTFE did very filled PTFE tested at 8 MPa and 25 °C little to reduce the friction in the contact as large portions of the load appear to have lubricated conditions as the friction levels ridden on the harder glass fibers. were generally the same or greater for MoS2 and fiberglass filled PTFE than black glass filled PTFE. A.3.6.3. Fiberglass and MoS2 filled PTFE

Analysis of the fiberglass and MoS2 filled PTFE revealed large areas of blended A.3.6.4. Bronze filled PTFE The wearing characteristics of bronze filled MoS2 and iron as well as sharp fibers. As in the case of black glass, these sharp PTFE were very consistent between high fibers tended to collect the other softer and low pressures and high and low materials in the direction of sliding and temperatures. In all cases, the bronze produced significant polishing on the particles had scratches in parallel with the counter-surface. This effect is shown in sliding motion. However, depending on the load, the bronze particles and the PTFE Fig. 11. While the MoS2 appeared to have smoothed out the PTFE surface, it does not surrounding them appeared to have worn at appear to have reduced friction in different rates.

thought that the bronze particles are more wear resistant than the PTFE material and so they collect at the surface and are pushed into the bulk material as it is steadily worn away. At some point equilibrium may be reached when the bronze and PTFE wear at equal rates. This equilibrium appears to be pressure dependent as it was observed to occur at 2 MPa and not at 8 MPa.

Figure 12: SEM image of bronze filled PTFE A.3.6.5. Carbon filled PTFE tested at 2 MPa and 25 °C Carbon filled PTFE had consistent performance at low and high pressures and low and high temperatures.

Figure 13: SEM image of bronze filled PTFE tested at 8 MPa and 25 °C Figure 14: SEM image of carbon filled PTFE tested at 2 MPa and 85 °C

This difference is highlighted by comparison of Figs. 12 and 13 showing bronze filled material tested at 2 and 8 MPa respectively. At 2 MPa, it appears that the surfaces of the worn bronze particles (white regions) and the bulk material (gray) are approximately level with smooth transitions between filler and bulk material. This is not the case at 8 MPa where gaps (black areas) can be distinguished around the edges of the filler particles. Additionally, the bronze particles occupy a much larger area of the contact Figure 15: SEM image of carbon filled PTFE region than at lower pressure loading. It is tested at 8 MPa and 25 °C

In all cases, the carbon particles wore A.4. Conclusions evenly with the PTFE bulk material as shown in Fig. 14 in which the darker Break-away friction was investigated in patches are carbon and the lighter gray reciprocating, sliding, block-on-plate areas are PTFE. The primary difference contact in the presence of lubricant for a between heavier and lighter loads was number of PTFE based materials as well as observed in the smoothness of the worn Babbitt material. The counter-surface was surface with higher loads yielding a mild steel. Sliding speed was kept as low markedly smoother surface than lighter as possible and mean contact pressure was loads as shown in Fig. 15. This smoothing varied from 1 to 8 MPa while oil bath at higher loads could have caused the trend temperature was varied from 25 to 85 °C. of reducing friction with increasing load A summary of the major findings is as that was observed with variations in follows: pressure. • Highest break-away friction levels were A.3.6.6. Pure PTFE provided by Babbitt and these were Pure PTFE showed very similar wear considerably higher than those of the characteristics to carbon filled PTFE with PTFE-based materials. Pure PTFE an even smoothing of the surface in provided the lowest friction levels while general without the complex characteristics carbon and bronze filled PTFEs also seen for the composite materials. As could produced low break-away friction. be expected, counter-surface particles were • Increased loading generally resulted in a not observed as in the case of the glass slight decrease of break-away friction, filled materials. except in the case of fiberglass filled PTFE The smooth surface seen for pure PTFE which had no consistent trend. could have helped to yield the very low • Pure PTFE and carbon filled PTFE friction coefficient observed for all provided very low variation in friction pressures with PTFE. Fig. 16 shows the through changes in both temperature and smoothing of the PTFE surface at high pressure. load and temperature. • The break-away friction coefficient with increasing temperature was found to increase for Babbitt and fiberglass filled PTFE materials while it decreased slightly in the cases of black glass filled and bronze filled PTFE. Break-away friction was not affected by temperature in the cases of carbon filled and pure PTFE. • Contact pressure and temperature were found to influence the break-away friction independently of each other for bronze filled, carbon filled, and pure PTFE as well

as Babbitt, i.e. there was no significant Figure 16: SEM image of pure PTFE tested at interaction between temperature and 8 MPa and 85 °C pressure effects.

• Significant polishing of the steel counter- Science and Technology, 65 (15–16) surface was observed in the cases of both (2005), pp. 2329–2343 black glass filled and fiberglass and MoS2 [7] D.-L. Gong, B. Zhang, Q.-J. Xue, H.-L. filled PTFE materials at all pressures. This Wang, Effect of tribochemical reaction of was clarified by the presence of iron on the polytetrafluoroethylene transferred film PTFE surfaces. Measurable polishing was with substrates on its wear behaviour, not observed on the counter-surfaces for Wear, 137 (2) (1990), pp. 267–273 the other materials. [8] Akagaki T, Kawabata M, Kato K. • Break-away friction for Babbitt increased Lubricated wear of ptfe-based composite at greatly following an extended stop before high sliding velocity. In: Proceedings of returning to the pre-stop value. Bronze the international tribology conference. filled, carbon filled, and pure PTFE Nagasaki, Japan; 2000. p. 917–22. showed a much smaller increase in break- [9] P. Samyn, P. De Baets, R. Keresztes, L. away friction following an extended stop Zsidai, G. Kalacska, E. Kislinder et al., before returning to pre-stop levels. Influence of cylinder-on-plate or block-on- ring sliding configurations on friction and wear of pure and filled engineering References polymers, TriboTest, 13 (2) (2007), pp. 83–100 [1] S.K. Biswas, K. Vijayan, Friction and [10] Z.-Z. Zhang, Q.-J. Xue, W.-M. Liu, wear of PTFE—a review, Wear, 158 (1–2) W.-C. Shen, Tribological properties of (1992), pp. 193–211 metal–plastic multilayer composites under [2] K. Friedrich, R. Reinicke, Z. Zhang, oil lubricated conditions, Wear, 210 (1–2) Wear of polymer composites, Proceedings (1997), pp. 195–203 of the Institution of Mechanical Engineers, [11] Z.-Z. Zhang, Q.-J. Xue, W.-M. Liu, Part J: Journal of Engineering Tribology, W.-C. Shen, Friction and wear behaviors 216 (6) (2002), pp. 415–426 of several polymers under oil-lubricated [3] S. Bahadur, D. Tabor, The wear of conditions, Journal of Applied Polymer filled polytetrafluorethylene, Wear, 98 Science, 68 (1998), pp. 2175–2182 (1984), pp. 1–13 [12] Z.-Z. Zhang, W.-C. Shen, W.-M. Liu, [4] B.J. Briscoe, M.D. Steward, A.J. Q.-J. Xue, T.-S. Li, Tribological properties Groszek, The effect of carbon aspect ratio of polytetrafluoroethylene-based on the friction and wear of ptfe, Wear, 42 composite in different lubricant media, (1) (1977), pp. 99–107 Wear, 196 (1–2) (1996), pp. 164–170 [5] Q.-J. Xue, Z.-Z. Zhang, W.-M. Liu, [13] D.M.C. McCarthy, S.B. Glavatskih, W.-C. Shen, Friction and wear Assessment of polymer composites for characteristics of fiber- and whisker- hydrodynamic journal-bearing reinforced ptfe composites under oil applications, Lubrication Science, 21 (8) lubricated conditions, Journal of Applied (2009), pp. 331–341 Polymer Science, 69 (1998), pp. 1393– [14] B.S. Unlu, E. Atik, S. Koksal, 1402 Tribological properties of polymer-based [6] K. Friedrich, Z. Zhang, A.K. Schlarb, journal bearings, Materials & Design, 30 Effects of various fillers on the sliding (7) (2009), pp. 2618–2622 wear of polymer composites, Composites

[15] T. Tevruz, Tribological behaviours of carbon filled polytetrafluoroethylene (ptfe) dry journal bearings, Wear, 221 (1) (1998), pp. 61–68 [16] T. Tevruz, Tribological behaviours of bronze-filled polytetrafluoroethylene dry journal bearings, Wear, 230 (1) (1999), pp. 61–69 [17] Y. Iwai, T. Honda, T. Onizuka, Y. Taniguchi, M. Kawabata, Wear properties of a glass fiber filled ptfe composite under oil lubrication, Japanese Journal of Tribology, 45 (5) (2000), pp. 389–401 [18] Glavatskih SB, Wasilczuk M, Fillon M. Performance peculiarities of ptfe faced tilting pad thrust bearings, HRW. [19] Bouyer J, Fillon M, Valle V. Stick- slip phenomenon induced by friction in a plain journal bearing during start-up. In: World tribology congress 2009. Kyoto, Japan; 2009. P. 101. [20] B.N.J. Persson et al., Rubber friction on wet and dry road surfaces: the sealing effect Physical Review B, 71 (2005), p. 035428

Tribological Behaviour of Polymeric Materials in Water Lubricated Contacts

A. Golchin, G. F. Simmons, S. Glavatskih, B. Prakash

Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology, Published online before print March 12, 2013, doi: 10.1177/1350650113476441

Tribological Behaviour of Polymeric Materials in Water Lubricated Contacts

A. Golchin1*, G. F. Simmons1, S. Glavatskih2,3, B. Prakash1 1. Division of Machine Elements, Department of Engineering Sciences and Mathematics, Luleå University of Technology, 97187 Luleå, Sweden 2. Machine Design, KTH Royal Institute of Technology,10044 Stockholm, Sweden 3. Department of Mechanical Construction and Production, Ghent University, B-9052 Zwijnaarde, Belgium

Abstract The present study aims at investigating the tribological behaviour of several unfilled polymer materials sliding against 316L stainless steel in distilled water. The tests were carried out in a uni-directional pin-on-disc configuration with an initial apparent contact pressure of 5 MPa at room temperature. The worn surfaces were examined using scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) techniques and the wear mechanisms were discussed. These studies indicate the occurrence of tribo-corrosion of stainless steel during sliding against polypropylene. It is inferred that the frictional behaviour of the polymers is determined by both their wettability and solubility in water where generally an increased hydrophobicity and relative energy difference (RED) in regard to water results in lower friction. The results from friction and wear tests show overall superior tribological performance of UHMWPE compared to the other polymeric materials in water.

Keywords: Water lubrication, Polymer-steel contacts, wettability. *Corresponding author: Arash Golchin ([email protected]).

B.1. INTRODUCTION lubricant in hydro-turbines. However, due to Mineral oils have long been used in much lower viscosity of water (0.66 cSt @ hydrodynamic sliding bearings in 40C) compared to typical turbine oils (32-68 hydropower stations. Their use as lubricant cSt @ 40C), longer operation in in aqueous environments however raises boundary/mixed lubrication regime is concerns about the environmental impact expected for these bearings. Therefore, these lubricants may have if they leak into choice of the materials and their tribological downstream water. Using water in performance are very important since these hydrodynamic sliding bearings is bearings may operate a considerable portion environmentally friendly and eliminates the of their life-time in boundary/mixed concerns associated with the use of oil as a lubrication regimes. The practical choice of

63 materials is however not only determined by dominant wear mechanisms involved in the mechanical and tribological properties, presence of water. but also by the price, simplicity of production, processing [1] and the practical B.2.1. Materials limitations in the real application. Thermoplastics exhibit superior properties Application of compliant polymers in water such as enhanced toughness and prolonged lubricated bearings has many advantages shelf-life [4]. Furthermore, their intrinsic which cannot be readily achieved with hard recyclability is a strong driving force for coatings such as DLC or ceramics. While their application in environmentally friendly most studies on the tribological behavior of lubricated bearings. polymers and composites are carried out in In this study, eleven polymeric materials dry conditions, there only exist a few studies were tested namely ultrahigh molecular under water lubricated condition. weight polyethylene (UHMWPE), Furthermore, there are rare comparative polyoxymethylene (POM), polyethylene studies of tribological behavior of polymer terephthalate (PET), polyamide 6 (PA 6), materials in aqueous environment; and in polyamide 66 (PA 66), some cases with contradictory results [2, 3]. polytetrafluoroethylene (PTFE), This study is therefore aimed at polypropylene (PP), polyvinylidene fluoride investigation of tribological behavior of (PVDF), polyether ether ketone (PEEK), several unfilled polymer materials in water polycarbonate (PC) and polymethyl to characterize their performance with methacrylate (PMMA) along with lignum regard to bearing application and to identify vitae which has been traditionally used as the wear mechanisms involved at the bearing material in water lubricated polymer/metal interface in presence of contacts. water. The characteristics of the polymers are shown in Table 1. The selection of polymers B.2. EXPERIMENTAL WORK was based on their availability, commercial The tribological behaviour of several application, and earlier studies in water unfilled polymer materials sliding against lubricated contacts [2, 3]. 316L stainless steel in water is studied using UHMWPE is an engineering thermoplastic a uni-directional pin-on-disc tribo-meter. with several beneficial properties e.g., high This configuration was chosen to avoid a toughness, excellent wear resistance and low converging gap in order to minimize the friction against metallic counter-faces in hydrodynamic pressure build-up at the aqueous environment. contact surfaces, enabling characterization It has been used as bearing material in total of friction and wear of materials in hip joint replacement for decades because of boundary/mixed lubrication regimes, a its excellent wear resistance and bio- challenge in the block-on-ring configuration. compatibility [5]. Studies were performed using several POM is a semi-crystalline polymer [6] with unfilled polymers to investigate the high crystallinity (70 to 100%) and low tribological performance and to study the water absorption [7]. The attraction of POM

Table 1: Characteristics of the polymer materials

Elastic Tensile Flexural Yield Density Elongation Hardness Toughness Melting Material Modulus Strength Strength Strength Tg [ºC] [g/cm3] at Break [%] [MPa] [kJ/m2] point [ºC] [GPa] [MPa] [MPa] [MPa]

UHMWPE 0.93 0.8 23 40 23 >450 65 * 135 -160 [15] PET 1.38 3.4 90 * 90 50-80 170 3.5 225 76 [16] PP 0.91 1.3-2 33 * 30-32 700 60 * 175 -17 [16] POM 1.42 3 70 100 70 70-75 140 * 175 -75 [16] PTFE 2.15 0.4 25-36 18-20 30 400 30 * 300-310 127 [16] PEEK 1.30 * 70 * 100 50 M99 * 334 145 [16] PMMA 1.15 1.9 * * 46 5 * * 230-260 107 PVDF 1.76 0.8 40 65 55-60 25/500 110R 300 170 -35 [16] PA6 1.15 1.5 50 40 50 200 70 >7 220 -8 [17] PC 1.15 2.3 65 90 60 >80 100 * 230 150 [16] PA66 1.19 1.7 70 42 70 150 100 >5 255 -6 [17] * Not Available is that its creep resistance is ten times that of [11], PTFE is considered as potential UHMWPE. However, the wear rate of POM bearing material for application in polymer- is higher than that of UHMWPE in water metal systems operating in aqueous [8]. environment. PET is an engineering thermoplastic which Polypropylene exhibits a combination of is mostly used as commodity thermoplastic characteristics such as insensitivity to humid due to its low cost. It has a wide range of environments, easy processing, corrosion application from packing and beverage resistance and low cost [12] which makes it containers, to production of synthetic fibers an interesting potential material for this and machine parts in mechanical industry. study. Its low cost, availability and wear resistance PC and PMMA are not usually used for in aqueous environment [2] makes PET a tribological applications; however these potential material for application in water materials were included in this study due to lubricated bearings. their ready availability. Polyamides (PA6 and PA 66) possess high PVDF is a semi-crystalline polymer which tensile and impact strength and exhibit possesses many remarkable properties, such abrasion resistance. Therefore, polyamides as good thermal stability under operating are used for many engineering parts and processing application temperatures, undergoing friction and wear such as excellent resistance to chemicals, hydrolytic bearings and gears [9]. These properties are stability in combination with low creep and attributed to presence of hydrogen bonds in high mechanical strength [13, 14]. polyamide molecular chains [10]. PEEK is a semi-crystalline thermoplastic Due to its low friction, good resistance to polymer with outstanding mechanical chemical attack and low water absorption properties such as high toughness, strength,

Table 2: Experimental Conditions at room temperature (21-23 ºC), under an applied load of 62.8 N producing an initial Load 62.8 N Initial Contact Pressure 5 MPa apparent contact pressure of 5 MPa. This Temperature R.T. (21-23 ºC) contact pressure was chosen to accelerate Sliding Speed 0.13 m/s testing of the materials and to imitate the Test Duration 20 hours* maximum apparent contact pressure at the Total Sliding Distance 9360 m* loaded region of bearing. Steel Surface Roughness Ra 0.2 µm Lubricant Distilled Water The experiments were carried out at constant sliding speed (0.13 m/s) considering the *Except for polypropylene which was tested for 18 hours with total sliding distance of 8424 meters. lowest practical rotational speed of the tribo- meter. The surface roughness of stainless steel discs were Ra=0.2 µm. These conditions were chosen to increase the Load Water 316L Disc Polymer Pin interaction of the friction surfaces during sliding. This would allow for characterization of the materials in boundary/mixed lubrication regimes. The test duration was 20 hours for all materials except for polypropylene which was ended a) b) after 18 hours of sliding due to excessive wear of the polymer. Three replicates were carried out for each material and results Figure 1: Schematic diagram of a) Pin b) Test configuration were averaged from the three test runs. A full description of experimental conditions is detailed in Table 2. and low creep rate. PEEK is widely used to make bearings and gears [18]. The stainless steel disc was immersed in The counter-surface in all experiments was distilled water and fixed in a rotating water AISI 316L, also known as marine grade bath while the polymer pin was attached to a stainless steel. stationary pin holder. The load was applied to the pin using dead weights and the friction force was continuously measured by B.2.2. Experimental Setup a strain gauge force transducer. A schematic The experiments were carried out using a diagram of the test configuration is shown in polymer pin on steel disc configuration. This Figure 1.b. configuration was chosen to avoid any possible edge effect and macro-ploughing of B.2.3. Sample Preparation the pin into the disc which could result in a Cylindrical pins were machined to the form contact configuration not representative of shown in Figure 1.a. The machining marks the real application. Tests were performed were removed from the contact surface of

0.45 PTFE 0.40 UHMWPE 0.35 PET P VDF 0.30 PEEK 0.25 POM 0.20 LV PP 0.15

Friction Coefficientt Friction PC 0.10 PMMA 0.05 PA 66 0.00 PA 6

Figure 2: Average friction coefficient over the test period of 20 hours. Error bars represent one standard deviation of the three repeated tests for each material.

the polymers prior to testing using 800 grit Results for the average friction coefficient SiC abrasive disc in a pin on disc over the entire test period of 20 hours are configuration to decrease the variability in shown in Figure 2 and clearly indicate that the results. The polymer pins were cleaned PTFE consistently provided the lowest level with ethanol in an ultrasonic bath prior to of friction with UHMWPE also providing the experiments to remove any possible low friction. contamination or loose particles attached to The mechanisms of low friction of the contact surfaces. The discs were laser- UHMWPE and PTFE are not yet well cut from 1 mm thick AISI 316L plates and understood. However some studies attribute were polished to surface roughness the low friction of UHMWPE to Ra=0.2±0.02 µm using 400 grit SiC paper. plasticization of polymer in water [19]. The discs were polished to a circular lay, to Wang et al [20] showed that the hardness of simulate the lay orientation in relation to PTFE and UHMWPE do not change after sliding direction in practical applications. immersion in water for 24 hours. Therefore, Cleaning was carried out using ethanol in an it can be concluded that tribological ultrasonic bath followed by air drying. behaviour of PTFE and UHMWPE under lubrication of pure water may not be dominated by plasticization. B.3. RESULTS AND DISCUSSION PET, PVDF and PEEK provided friction B.3.1. Friction coefficients of 0.16-0.21, statistically higher The experimental setup was found to than those of PTFE and UHMWPE, provide highly reproducible results, however lower than that obtained with allowing for clear comparison between the lignum vitae. POM and PP provided similar performances of the various materials. friction to lignum vitae while the remaining

0.10 POM 0.09 PTFE 0.08 PET LV 0.07 PEEK 0.06 P VDF 0.05 PC 0.04 PA 66 0.03 PA 6 0.02 PMMA

Variation in Friction Coefficientt in Friction Variation 0.01 UHMWPE 0.00 PP

Figure 3: Variation in friction coefficient calculated as the standard deviation of friction curve over the test period of 20 hours. Error bars represent one standard deviation of mean values over three repeated tests for each material.

materials, namely PC, PMMA, PA 6 and PA B.3.2. Evolution of Friction 66 provided higher friction than lignum Figures 4, 5, and 6, show the evolution of vitae with PA6 and PA66 exhibiting the friction coefficient during the entire tests. In highest level of friction coefficient of all general, each of the tested materials tested materials. followed one of four different trends. Analysis of the variation in friction The first of these trends was that of PET, coefficient over the course of the test POM, PEEK, PTFE, PC and lignum vitae provides insight into the material’s which provided a relatively stable friction performance in the actual application. Figure coefficient during the test time following a 3 shows the standard deviation of the short running-in period (Fig. 4). The second friction coefficient observed for the trend was that of PMMA which showed materials during the entire test period of 20 very large fluctuations in friction coefficient hours. This demonstrates the stability or during the entire test; however no consistent instability of the material’s friction behavior was observed with this material performance as the steel and polymer (Fig. 5). surfaces undergo chemical and/or Another trend was displayed by PVDF, PA topographical changes during the course of 6, PA 66, and UHMWPE. These materials the tests. exhibited a steady increase in friction In the case of POM, PTFE, PET, lignum coefficient during the test following a short vitae, PEEK and PVDF relatively little running-in period. The final trend was variation in friction during the test time was obtained with PP which exhibited a highly observed. time dependant behaviour during the test It was a different case for the remainder of time (Fig. 6). Formation of a continuous the materials which showed higher variation transfer film on stainless steel counter-face in friction coefficient along with large was inhibited for all tested materials; scatter. however Scanning Electron Microscopy

0.40 influence hydrodynamic pressure build-up PET 0.30 and contribute to the increase of friction; POM 0.20 PEEK adhesive force between these micro-patches PTFE of transfer film and polymer pin is believed 0.10 PC

Friction Coefficient LV to be the dominant parameter leading to the 0.00 0 1500 3000 4500 6000 7500 9000 trend of increasing friction for PVDF, PA 6, Sliding Distance [m] PA 66 and PMMA. The same trend was also observed with Figure 4: Evolution of friction coefficient for PET, POM, PEEK, PTFE, PC and lignum vitae UHMWPE which exhibited an increase in friction coefficient of about two fold during the entire test. Although an adherent transfer film was not 0.40 formed with UHMWPE, cleaning the PVDF 0.30 PA 6 UHMWPE wear track on disc by a paper 0.20 PA 66 cloth at the end of the tests reduced the UHMWPE 0.10

Friction Friction Coefficient PMMA friction coefficient back to the values 0.00 0 1500 3000 4500 6000 7500 9000 obtained at the beginning of the tests. This Sliding Distance [m] simple experiment confirmed the role of the transfer film on increasing friction of Figure 5: Evolution of friction coefficient for PVDF, UHMWPE. PA 6, PA 66, UHMWPE and PMMA The time dependant trend in friction of PP when sliding against stainless steel can be explained by the interaction of wear and 0.40 corrosion of stainless steel counter-face. 0.30 Adhesive wear is the dominant wear 0.20 PP mechanism of PP when sliding against 0.10 stainless steel. Therefore, the larger the area Friction Coefficient 0.00 of contact, the larger the force required to 0 1500 3000 4500 6000 7500 Sliding Distance [m] shear the interface between polymer pin and disc. Polypropylene exhibits a trend of Figure 6: Evolution of friction coefficient for PP increasing friction coefficient at the initial stages of sliding. This is attributed to the (SEM) of worn stainless steel counter-faces progressive wear of the pin resulting in revealed signs of micro-patches of polymer increased contact area between pin and disc transfer when sliding against PA6, PA 66, until the friction reaches its maximum level PEEK, PET, PMMA, POM, PVDF, and when the truncated conical tip of the pin is UHMWPE. No such micro-transfer could be totally removed. However, after some hours observed on counter-faces sliding against of sliding, pitting corrosion of the counter- PTFE, PC and PP. Although presence of face (Fig. 7) results in decreased contact such discrete transfer film may detrimentally area between pin and disc and consequently

PA 66 and PA 6 exhibiting the largest water absorption respectively. The equilibrium state of water absorption is reached after eight weeks with all the polymer materials except for polyamides (PA 6 and PA 66) which continue to absorb water beyond this period. The high water absorption of polyamides is generally attributed to the presence of amide groups in the molecular chain of the polymers [22,

Figure 7: SEM image of worn stainless steel surface 23], favouring the water absorption by after sliding against polypropylene forming hydrogen bonds with water molecules owing to their high polarity [22]. The interaction of water molecules with decreased friction coefficient until it reaches polyamide chains are further discussed in its minimum level at the termination of the [24] suggesting two possible interactions i.e. test. hydrogen bonding and self-association.

B.3.3. Water Absorption The water absorption characteristics of the B.3.4. Wettability materials are very important to consider in The wettability of the polymers was the design of water lubricated bearings. examined by contact angle measurements of Swelling of bearing materials results in 4 µL drop of distilled water deposited for decreased clearance between shaft and one second on polymer surface at room bearing surfaces and depending on the temperature (23 ºC). extent of the water up-take, a bearing could potentially act as a break, resulting in increased frictional heating and wear. Table 3: Average surface roughness of samples used Although assessment of water absorption of for contact angle measurements polymers should be based on long term Material* Ra [µm] PA66 0.23 equilibrium, most specification values are POM 0.38 only based on a few hours of exposure [21]. PVDF 0.20 Figure 8 shows the weight percentage water PA6 0.30 absorption of the polymers for an exposure PET 0.20 period of four months. Lignum vitae was not PEEK 0.49 included in this section of the study because PTFE 0.20 it is generally stored in water to maintain its PP 0.29 stability. The results show that most of the UHMWPE 1.22 polymers exhibited water absorption level of less than 0.2%, while POM, PMMA, PET, *Surface roughness of PC and PMMA could not be measured using optical profile-meter due to their PA 66, and PA6 showed larger values with transparency.

8 UHMWPE 7 PET PP 6 POM 5 PTFE 4 PEEK 3 PVDF PA6 2 PC 1 Water Absorption[Weight %] PA66 0 PMMA 0 50 100 Time [Days]

Figure 8: Water absorption of polymers versus exposure time

The samples for contact angle measurements regarding the surface wettability of the were cut from polymer rods using a materials. precision microtome. The surface roughness The latter was attributed to the enhanced of the samples has been measured using an formation of a water layer at the contact optical profilometer and the average results surface and the hydrostatic lift due to of five measurements are shown in Table 3. presence of water at the pin/disc interface. The hydrostatic lift due to the capillary Table 4 shows the average results of ten forces is in fact more pronounced when both contact angle measurements for each contacting surface are hydrophobic and this material. Figures 9 and 10 show the seems to have a reasonable correlation with measured water contact angle against the the results obtained in the current study. specific wear rate and average friction coefficient of polymers respectively. No correlation could be found between Table 4: Average water contact angle of polymers wettability and wear behavior of the Confidence Material Contact Angle [°] polymers. Although not statistically Interval % significant, in general, a trend of decreasing PA66 58 10 friction coefficient can be observed with POM 70 13 increase in contact angle of the polymers. PVDF 70 7 PA6 77 8 Borutto et al. [25] discussed the influence of PMMA 81 9 surface wettability on friction and wear of PET 84 7 different combination of materials and PC 90 11 PEEK 93 7 suggested that a combination of hydrophilic PTFE 100 7 pin and hydrophobic disc provides the PP 106 6 lowest friction and wear of the materials UHMWPE 118 14

1.E-03 UHMWPE PET POM /N.m] 3 PVDF

mm 1.E-04 PA 6 PA 66 PTFE 1.E-05 PC PEEK

Specific Wear Rate [ PMMA 1.E-06 PP 50 70 90 110 Contact Angle [Degree]

Figure 9: Specific wear rate vs. water contact angle of polymers

0.45 PTFE 0.40 UHMWPE 0.35 PET 0.30 PVDF PEEK 0.25 POM 0.20 PP 0.15 PC

Friction Coefficient Friction 0.10 PMMA 0.05 PA 66 0.00 PA 6 50 60 70 80 90 100 110 120

Contact Angle [Degree]

Figure 10: Average friction coefficient over the test period of 20 hours vs. water contact angle of polymer

Although capillary repulsion may contribute proposed by Hansen [26]. In 1967 he to lower friction at higher contact angles; the introduced Hansen solubility parameters general trend observed cannot be solely (HSP) consisting of several individual parts explained by this effect and a further based on the total energy of vaporization of investigation is required to reveal the a liquid. Materials which have similar mechanisms involved. solubility parameters have high affinity for each other and the extent of their similarity B.3.5. Solubility Parameters influences the extent of the interaction A widely used approach to determine the between the materials, the principle often solubility of polymers in solvents is quoted as “like seeks like”.

Table 5: Hansen solubility parameters and corresponding R , R and RED of polymers in regard to water

Materials* [ ] [ ] [ a ] 0 [ ] [ ] RED 1 1 1 1 1 �2 �2 �2 �2 �2 PVDF 훅퐝 푀푃푎17 훅퐩 12.1푀푃푎 훅퐡 10.2푀푃푎 푹ퟎ 4.1푀푃푎 푹풂 32.5푀푃푎 7.9 PTFE 16.2 1.8 3.4 3.9 41.4 10.6 PA6 17 3.4 10.6 5.1 34.2 6.7 PA 66 16 11 24 3 19.0 6.3 PMMA 18 10.5 5.1 9.5 37.9 4.0 PET 18.2 6.4 6.6 5 37.4 7.5 POM 17.2 9 9.8 5.3 33.4 6.3 PP 18 0 1 6 44.6 7.4 PC 18.1 5.9 6.9 5.5 37.2 6.8

* No information could be found for UHMWPE and PEEK.

The investigation of the Hansen solubility subscripts 1 and 2 represent polymers and parameters of the polymers and their relative water respectively. energy difference (RED) in comparison with RED values less than one indicate good water can provide information about the solubility whereas higher RED values extent of the interaction of the polymer indicate poor solubility of polymers in the chains at the interface with water and solvent. therefore can be of high importance for Figures 11 and 12 show the specific wear studying the tribological behavior of rate and the average friction coefficient of polymers in water lubricated contacts. The the materials versus relative energy Hansen solubility parameters [27] and difference (RED) of the polymers in regard corresponding and RED values of the to water. polymers are provided in Table 5. No clear correlation can be found between 푅푎 The distance between Hansen solubility the specific wear rate of the materials and parameters in Hansen space ( ) and the RED; however a trend of decreasing friction relative energy difference (RED) of the can be observed with an increase in RED. 푅푎 polymers in regard to water is obtained from A careful look at Figures 10 and 12 shows the following formulae that among the polymers with similar RED values (e.g.PA66 and POM, or PA6 and PC) ( ) = ( ) + + ( ) the ones having a higher contact angle ퟐ ퟐ ퟐ ퟐ 풂 풅ퟐ 풅ퟏ 풑ퟐ 풑ퟏ 풉ퟐ 풉ퟏ (POM and PC) exhibit lower friction 푹 = ퟒ 휹 − 휹 �휹 − 휹 � 휹 − 휹 coefficient and among the polymers with 풂 ퟎ 푹푬푫 푹 ⁄푹 similar contact angles (e.g. POM and PVDF) Where is the disperse part, is the polar the one having a higher RED value (PVDF) part and is the hydrogen part of HSP. The shows a lower friction coefficient. 훿푑 훿푝 훿ℎ

1.E-03 PET

POM /N.m]

3 PVDF 1.E-04 mm PA 6

PA 66

PTFE 1.E-05 PC

PMMA

Specific Wear Rate [ Rate Wear Specific 1.E-06 PP 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0

Relative Energy Difference

Figure 11: Specific wear rate vs. relative energy difference (RED) of polymers

0.45 PTFE 0.40 PET

0.35 PVDF

0.30 POM 0.25 PP 0.20 PC 0.15

Friction Coefficient PMMA 0.10 PA 66 0.05 0.00 PA 6 3 5 7 9 11 Relative Energy Difference

Figure 12: Average friction coefficient over the test period of 20 hours vs. relative energy difference (RED) of polymers

Considering the latter and the general trends water at its surface and thus influences the observed in friction of polymers in water wettability characteristics and possibly with respect to their water contact angle and surface free energy of the polymers. This relative energy difference it is inferred that has been experimentally confirmed in this the frictional behavior of the polymers is study with water contact angle measurement inversely related to their water contact angle of PA6 and PA66 which showed 54 and 11 and relative energy difference (RED). degrees reduction in contact angle Penetration of water molecules at the surface respectively after immersion in water for of the polymers alters polymers affinity to more than eight months.

1.E-03 UHMWPE PET

/Nm] POM 3 LV mm 1.E-04 PVDF PA 6 PA 66 1.E-05 PTFE PC

Specific Wear Rate [ SpecificWear PEEK PMMA 1.E-06 PP

Figure 13: Specific wear rate of the materials. Error bars represent one standard deviation of the three repeated tests for each material.

Solubility of polymer at the interface with polymer pin in water for 20 hours and water is determined by its relative energy measuring the water absorption during this difference (RED) in regard to water. period. Knowing the density of the materials Generally, the solubility of polymers with the wear volume was calculated by the higher RED is lower and consequently their following formula: water contact angle does not decrease significantly due to exposure to water. This = may be the reason that polymers with higher ∆푚 ∆푉 RED values generally show lower friction 휌 Where is the wear volume, is the coefficient in water lubricated sliding wear mass and is density of the material. contacts. However, the mechanism as to ∆푉 ∆푚 The results of the wear rates obtained for all why an increased water contact angle results 휌 tested materials are shown in Figure 13. in lower friction coefficient in water is not UHMWPE provided the lowest wear rate yet clearly understood and requires further with PET and POM also providing low wear investigations. rates. The wear resistance of PVDF, PA 6

and PA 66 was similar to that of lignum

vitae while PTFE, PC, PEEK, PMMA and B.3.6. Wear PP provided higher wear rates with PP Wear of the materials was calculated from showing the highest wear rate of all tested weight loss measurement of the samples materials in water lubricated contacts. after each test. The measurements were Further investigations utilizing scanning corrected for the amount of water absorbed electron microscopy revealed the wear by samples during the test time. This was mechanisms of the materials in water carried out by immersing a separate unused lubricated sliding contacts. Overall however,

SEM and EDS examination of worn stainless steel surfaces further determined the suitability of materials for the real application. The sliding direction in SEM images is indicated by an arrow.

Polypropylene Scanning electron microscopy of worn PP pins showed even smoothing of the surface Figure 15: SEM image of worn PMMA surface with few scratches on the polymer pin surface. Figure 14 shows the SEM image of PP wear Polymethyl methacrylate debris, which generally has a rolled Analysis of worn PMMA surface revealed morphology with high aspect ratio abrasion marks along the sliding direction suggesting that adhesive wear is the with some spots undergoing spalling (Fig. dominating wear mechanism of PP when 15) which may have been caused by sliding against stainless steel. However successive loading of brittle PMMA ridges some irregular shaped wear debris was also when sliding against stainless steel noticed which was caused by the change in asperities. surface characteristics of stainless steel due to corrosion when sliding against polypropylene. The combined action of Polyether ether ketone and Polycarbonate adhesive wear and corrosion induced PEEK and PC showed very similar wear abrasive wear led to the high wear rate of characteristics; both undergoing polypropylene. smoothening of the polymer surface with shallow furrowed features along the sliding

Figure 14: SEM image of PP wear debris Figure 16: SEM image of PEEK wear debris

Figure 17: SEM image of PC wear debris Figure 19: SEM image of PTFE wear debris

direction. The wear debris tended to Formation of a PTFE transfer film in dry agglomerate and form large and brittle sliding condition partially protects the bulk particles. polymer from the abrasive action of hard Figure 16 and 17 show the morphology and metal asperities. coalescence of the wear debris of PEEK and However such transfer film is not formed in PC respectively. water lubricated contacts leading to severe abrasive wear of the polymer. As a result the Polytetrafluoroethylene worn surface of PTFE is characterised by With low water absorption, low friction, and scratches along the sliding direction (Fig. excellent chemical stability in various 18). corrosive environments, PTFE is considered The wear debris is in form of millimetre- as potential bearing material for application sized rolled sheets (Fig. 19) generated by in water. However high wear rate is the most ploughing and abrasive action of hard prominent deficiency of PTFE for stainless steel asperities when sliding against application in lubricated contacts. PTFE.

Figure 18: SEM image of worn PTFE surface Figure 20: SEM image of PET wear debris

Polyethylene terephthalate hydrogen bonds between polymer chains Many studies attribute the higher wear rate and form hydrogen bonds with amide groups of polymers in the aqueous environment to while diffusing into the polymer surface. the inhibition of formation of a transfer film This weakens the intermolecular forces and on the metal counter-face; however, the deteriorates the polymer’s mechanical effect of plasticization of polymers caused properties [22, 28]. The reduction of by absorption of lubricants is neglected in attractive forces between polymer chains these studies. allow for easy material removal and high Yamada and Tanaka [6] observed formation wear rate in water lubricated sliding of a PET transfer film on stainless steel contacts. counter-faces in both dry and water The high wear rates of polyamides are lubricated contacts which is consistent with attributed to the plasticization caused by the SEM and EDS examination of the worn water absorption. Although the worn stainless steel surfaces in the present study. surfaces of PA 6 and PA 66 are smooth and They noticed that the wear rates measured in have similar appearances; the wear debris of water lubricated conditions was much higher these polymers has different morphologies than that in dry sliding contacts and (Fig. 21, 22) implying that the wear suggested that the higher wear rate of PET mechanisms of the two polyamides are with water lubrication is due to surface different. plasticization occurring as a result of The wear mechanism of PA 66 is dominated permeation of water molecules into the by abrasive wear, while adhesive wear is the amorphous phase of PET. Plasticization of dominant wear mechanism of PA 6. the polymer surface can soften the polymer surface and make the polymer more Polyvinylidene fluoride (PVDF) susceptible to abrasion. However the latter is Figure 23 shows the worn surface of PVDF not confirmed yet and requires further which is characterized with deep grooves investigations. resulting from severe abrasion of polymer In the current study, the worn surface of the PET pin was smooth with fine scratches along the sliding direction and the wear debris was in form of long and thin chips as shown in Figure 20. This implies that abrasive wear is the dominant wear mechanism of PET in water lubricated sliding contact against stainless steel.

Polyamides (PA 66 and PA 6) The mechanism of interaction of polyamides with water molecules were described in [24] suggesting that water molecules loosen the Figure 21: SEM image of PA 66 wear debris

Figure 22: SEM image of PA 6 wear debris Figure 24: SEM image of PVDF wear debris surface by hard asperities of the counter However in uni-directional testing of surface. The wear debris is in form of very unfilled UHMWPE, no such flake-like thin sheets as shown in Figure 24. This is morphology was observed. thought to be caused by micro-ploughing of Worn polymer surface was relatively hard steel asperities into the polymer and smooth with shallow furrows along the abrasive wear of PVDF as a result of sliding direction indicating that micro- shearing of the interface. Such sheet-like abrasion was the dominant wear mechanism morphology of PVDF wear debris was of UHMWPE. reported earlier in dry sliding contacts when sliding against steel counter-faces [14]. Polyoxymethylene The examination of worn polymer pins Ultra High Molecular Weight Polyethylene shows that the plateau of the polymer is The wear debris of unfilled UHMWPE in relatively smooth with fine abrasion marks reciprocating sliding motion is generally parallel to the sliding direction. The basic reported to have a flake-like morphology failure mechanisms of polyoxymethylene in [29]. dry sliding contacts were described by

Figure 23: SEM image of worn PVDF surface Figure 25: SEM image of worn POM surface

Clerico [30] and were attributed to sub- study when sliding against the remaining surface deformation with crack initiation materials. and propagation. It is very well known that the corrosion This is consistent with the observation of resistance of stainless steel is derived from polymer surface in this study. The worn the formation of a few nano-meter thick

POM surface shows abrasion marks parallel passive film of chromium oxide Cr2O3 on to the sliding direction with some spots the surface of stainless steel [31]. undergoing plastic deformation and micro- Growth of an oxide film provides protection cracking (Fig. 25) suggesting that the wear against corrosion while wear break-up and mechanism of POM is dominated by mild- remove the passive film. Therefore, passive abrasion, plastic deformation and micro- film properties (stability and adherence) and cracking. re-passivation kinetics play important roles in determining the interactions between corrosion and wear [32]. B.3.7. Effect on Stainless Steel Counter Occurrence of corrosion of stainless steel when sliding against polypropylene is Surface primarily attributed to the wear of the The effects of polymers on stainless steel protective oxide layer along the wear track. counter-surfaces were analyzed using Once the passive film is removed due to Scanning Electron Microscopy (SEM) and sliding action of the counter-body, the re- Energy Dispersive Spectroscopy (EDS) passivation kinetics determines the corrosion techniques. behaviour of the nascent stainless steel Formation of a continuous transfer film on during sliding. stainless steel counter-faces was inhibited One of the key operating parameters for all tested materials; however Scanning influencing the re-passivation kinetics of Electron Microscopy (SEM) of worn stainless steel is contact frequency. stainless steel counter-faces revealed signs The influence of contact frequency on tribo- of micro-patches of polymer transfer when corrosion behaviour of stainless steel in sliding against PA6, PA 66, PEEK, PET, corrosive environments has been widely PMMA, POM, PVDF, and UHMWPE. No studied earlier [31, 32, 33], finding that such micro-transfer could be observed on increased contact frequency resulted in counter-faces sliding against PTFE, PC and decreased open-circuit potential and PP. increased anodic current. This was attributed SEM and EDS analysis of counter surfaces to the fact that at higher contact frequencies show occurrence of corrosion of stainless more active material is exposed to the steel when sliding against polypropylene as electrolyte due to the shorter time in shown in Figure 7. Neither measurable between two successive contacts along the change in surface roughness, nor corrosion wear track [32]. If the time between two of the stainless steel counter surface was successive contacts is shorter than the time detected for the conditions of the present needed for re-passivation, the nascent

80 stainless steel is continuously exposed to the increased contact angle and relative energy electrolyte and therefore becomes de- difference in regard to water results in passivated. The latter is believed to be the decreased friction. secondary cause for tribo-corrosion of • Tribo-corrosion of stainless steel occurred stainless steel counter surfaces. when sliding against polypropylene. No However, further investigations of the tribo- such corrosion was observed when sliding chemical behaviour of stainless steel against against remaining materials. polymer materials in aqueous environments are required to reveal the mechanisms • Micro-patches of polymer transfer film involved in relation to tribo-corrosion. were observed on worn stainless steel counter-faces for all polymer materials except for PTFE, PC, and PP. However no B.4. CONCLUSIONS correlation was observed between formation of such transfer films and wear of polymers. The tribological behaviour of several unfilled polymer materials as well as lignum • The wear mechanism of PMMA, PEEK, vitae was investigated in uni-directional PC, PTFE, PA 66, PVDF, PET, and sliding pin-on-disc contact in the presence of UHMWPE is dominated by abrasive wear distilled water. The tribological behaviour, while adhesive wear is the main wear wear mechanisms, water absorption, mechanism of PA 6 and PP. POM undergoes wettability, solubility of the polymers and mild abrasive wear, plastic deformation and their influence on frictional behaviour of the micro-cracking in water lubricated contacts. polymers were discussed. A summary of the • With low water absorption, comparable major findings is as follows: friction to that of PTFE, lowest wear rate and marginal effect on stainless steel • PTFE consistently provided the lowest counter-face, UHMWPE exhibits promising friction coefficient followed by UHMWPE. characteristics for further development in • UHMWPE exhibited the lowest wear rate regard to application in water lubricated with PET and POM also showing low wear bearings. rates. • Relatively small variation in friction Funding and Acknowledgement coefficient was observed with POM, PTFE, The research presented in this paper has PET, lignum vitae, PEEK and PVDF. been funded by “StandUp for Energy”. • POM, PMMA, PET, PA 66, and PA 6 StandUp for Energy is a collaboration absorbed significant amount of water (0.7- initiative between Uppsala University, The 7.7 wt %) compared to the remaining Royal Institute of Technology (KTH), The materials (<0.2 wt %). Swedish University of Agricultural Sciences (SLU) and Luleå University of Technology • The frictional behaviour of the polymers is (LTU). It arose as a result of the influenced by both their wettability and Government’s commitment to high quality solubility in water where generally an

81 research in areas of strategic importance to 11. Chen B., Wang J. and Yan F., society and the business sector. Microstructure of PTFE-based polymer blends and their tribological behaviors under REFERENCES aqueous environment, Tribol. Lett., 2012, Vol. 45, pp. 387-395. 1. Ginzburg BM et al., Polymeric materials for 12. Chow WS et al., Effect of maleic anhydride- water-lubricated plain bearings, Russ. J. grafted ethylene–propylene rubber on the Appl. Chem., 2006, Vol. 79, pp. 695-706. mechanical, rheological and morphological 2. Clarke CG and Allen C., The water properties of organoclay reinforced lubricated sliding wear behaviour of polyamide 6/polypropylene nanocomposites, polymeric materials against steel, Tribol. Eur. Polym. J., 2005, Vol. 41, pp. 687-696. Int., 1991, Vol. 24, pp. 109-118. 13. Wang HG et al., Friction and wear behavior 3. Mens JWM., Friction and wear behaviour of of polyamide 66/poly (vinylidene fluoride) 18 polymers in contact with steel in blends, J. Macromol. Sci. Part B Phys., environments of air and water, Wear, 1991, 2008, Vol. 47, pp. 701-711. Vol. 149, pp. 255-268. 14. Peng Q. et al., The preparation of 4. Li J. and Cai C., Friction and wear PVDF/clay nanocomposites and the properties of carbon fiber reinforced investigation of their tribological properties, polypropylene composites, Adv. Mat. Res., Wear, 2009, Vol. 266, pp. 713-720. 2011, Vol. 284-286, 2380-2383. 15. Kurtz SM., UHMWPE Biomaterials 5. Sujeet K Sinha, Brian J Briscoe, Polymer Handbook - Ultra-High Molecular Weight Tribology, London: Imperial College Press, Polyethylene in Total Joint Replacement and 2009, p. 196. Medical Devices. 2nd ed. Elsevier, 2009. 6. Yamada Y. and Tanaka K., Effect of the 16. Subramanian MN., Basics of degree of crystallinity on the friction and troubleshooting in plastics processing - An wear of poly (ethylene terephthalate) under introductory practical guide. Wiley – water lubrication, Wear, 1986, Vol. 111, pp. Scrivener 2011. 63-72. 17. Ehrenstein GW., Polymeric materials: 7. Ginzburg BM et al., Tribological behaviour structure, properties, applications. Hanser of polyoxymethylene in water-lubricated Verlag 2001. friction against steel, J. Frict. Wear, 2011, 18. Chen B., Wang L. and Yan F., Friction and Vol. 32, pp. 246-250. wear behaviors of several polymers sliding 8. Shen C. and Dumbleton JH., The friction against GCr15 and 316 steel under the and wear behavior of polyoxymethylene in lubrication of sea water, Tribol. Lett., 2011, connection with joint replacement, Wear, Vol. 42, pp. 17-25. 1976, Vol. 38, pp. 291-303. 19. Xiong D. and Ge S., Friction and wear 9. Yang Z. et al., Effects of polyamide 6 on the properties of UHMWPE/Al2O3 ceramic crystallization and melting behavior of β- under different lubricating conditions, Wear, nucleated polypropylene, Eur. Polym. J., 2001, Vol. 250, pp. 242-245. 2008, Vol. 44, pp. 2754-3763. 20. Wang JZ, Yan FY and Xue QJ, Tribological 10. De Baets P. et al., The friction and wear of behaviors of some polymeric materials in different polymers under high-load sea water, Chin. Sci. Bull., 2010, Vol. 54, conditions, J. Synth. Lubr., 2002, Vol. 19, pp. 4541-4548. pp. 109-118.

21. Smith WV., Material selection criteria for 32. Berradja A. et al., Effect of sliding wear on water lubrication, Wear, 1973, Vol. 25, pp. tribocorrosion behaviour of stainless steels 139-153. in a ringer's solution, Wear, 2006, Vol. 261, 22. Rajesh JJ et al., Effect of water absorption pp. 987-993. on erosive wear behaviour of polyamides, J. 33. Henry P., Takadoum J. and Berçot P., Mater. Sci., 2002, Vol. 37, pp. 5107-5113. Depassivation of some metals by sliding 23. Chen Z. et al., Mechanical and tribological friction, Corros. Sci. 2011, Vol. 53, pp. 320- properties of PA/PPS blends, Wear, 2004, 328. Vol. 257, pp. 696-707. 24. Song J. and Ehrenstein GW., The influence of water absorption on the properties of polyamide, Kunstst. Ger. Plast., 1990, Vol. 80, pp. 722-726. 25. Borutto A., Crivellone G., Marani F., Influence of surface wettability on friction and wear tests, Wear, 1998, Vol. 222, pp. 57-65. 26. Hansen C. M., The three-dimensional solubility parameter - key to paint component affinities: solvents, plasticizers, polymers, and resins. II. Dyes, emulsifiers, mutual solubility and compatibility, and pigments. III. Independent calculation of the parameter components, J. Paint Tec., 1967, Vol. 39, Issue 511, pp. 505-510. 27. Hansen CM., Hansen solubility parameters: a user's handbook. Boca Raton, Fla.: CRC Press; 2000. 28. Srinath G. and Gnanamoorthy R., Sliding wear performance of polyamide 6–clay nanocomposites in water, Compos. Sci. Technol., 2007, Vol. 67, pp. 399-405. 29. Oonishi H., Ishimaru H and Kato A., Effect of cross-linkage by gamma radiation in heavy doses to low wear polyethylene in total hip prostheses, J. Mater. Sci. - Mater. Med., 1996, Vol. 7, pp. 753-763. 30. Clerico M., Tribological behaviour of polyacetals, Wear, 1980, Vol. 64, pp. 259- 272. 31. Sun Y. and Rana V., Tribocorrosion behaviour of AISI 304 stainless steel in 0.5 M NaCl solution, Mater. Chem. Phys., 2011, Vol. 129, pp. 138-147.

Effect of Shaft Roughness and Pressure on Friction of Polymer Bearings in Water

A. Golchin, T. D. Nguyen, P. De Baets, B. Prakash, S.B. Glavatskih

To be communicated

Effect of Shaft Roughness and Pressure on Friction of Polymer Bearings in Water

a* c c a b,c Arash Golchin , Tan Dat Nguyen , Patrick De Baets , Braham Prakash and Sergei Glavatskih

*Corresponding Author: [email protected]

a. Division of Machine Elements, Department of Engineering Sciences and Mathematics, Luleå University of Technology, 97187 Luleå, Sweden b. Division of Machine Design, School of Industrial Engineering and Management, Royal Institute of Technology, Stockholm, Sweden c. Department of Mechanical Construction and Production, Ghent University, Gent, Belgium

Abstract In this study, the frictional behavior of selected commercially available unfilled polymers namely PEEK, PTFE, PET and UHMWPE against an Inconel shaft was investigated using a journal bearing test configuration in water lubricated sliding contact. Dynamic friction curves were obtained for various shaft roughness values and polymer combinations. The results showed a significant influence of shaft surface roughness on running-in and steady state friction in water lubricated conditions. Contact angle measurements revealed a significant increase in wettability of Inconel counter faces. This suggests formation of a reaction layer on worn Inconel surfaces when sliding against the polymers. The influences of counter surface roughness and load on frictional response of polymers were studied through intermittent tests by obtaining dynamic and breakaway friction maps for different polymer materials, shaft roughness values and pressure combinations. In general, a trend of decreasing friction was obtained with increasing contact pressure; however the materials’ frictional responses to variations in counter surface roughness were different. These results indicate that although a reduced counter surface roughness may be beneficial for dynamic friction of polymers in all lubrication regimes, it can adversely affect the materials’ breakaway friction response.

Keywords: Water Lubrication, Breakaway, Friction, Polymers.

C.1. Introduction environmental impacts these lubricants may have if they leak into downstream water. One of the most important demands in These concerns have led to efforts to use hydropower industry nowadays stems from water as a lubricant and the emergence of the increasing emphasis on preserving the the concept of an “Oil-Free Plant” for environment. Application of mineral or hydropower generating companies. This synthetic oil as a lubricant in aqueous however poses many challenges which environments raises concerns about the require consideration of the different aspects

87 of bearing design, operating conditions and similar behavior with some thermoplastic selection of shaft and bearing materials. Due polymers under low loads [4]. In another to considerably lower viscosity of water study Quaglini et al. showed that an (0.66 cSt at 40 ºC) compared to that of increased counter surface roughness turbine oils (32-68 cSt at 40 ºC), water influenced the material’s frictional behavior lubricated hydrodynamic sliding bearings depending on the polymers’ elastic modulus operate at lower specific bearing pressures [3] such that an increased counter surface compared to their oil lubricated roughness led to increased friction for soft counterparts. To achieve similar load polymers and decreased friction for carrying capacity, water lubricated bearings polymers with high elastic modulus. Some are generally manufactured with larger other studies suggest the existence of an length to diameter ratios (L/D) which optimal surface roughness for minimal exacerbates the issues associated with friction of different polymers [4, 5]. shaft/bearing misalignment. The influence of contact pressure on It is well known that the combined surface tribological behavior of polymers has been roughness of shaft and bearing surfaces play extensively studied earlier in order to an important role in determining the evaluate the limiting PV values for various lubrication regime of sliding bearings. This polymeric materials in dry sliding role becomes even more crucial in water conditions. However, in lubricated contacts lubricated bearings due to the much larger frictional response of polymers differs from roughness to film thickness ratios in that in dry contacts due to the effects arising comparison to the oil lubricated bearings. from the presence of lubricant. Despite the Earlier simulations show the beneficial practical significance of the latter, very little influence of reduced combined surface work has been accomplished in regards to roughness on hydrodynamic lubrication of the effect of counter surface roughness or journal bearings [1]. Using a smoother shaft contact pressure on friction of polymers in could potentially allow for applying higher lubricated conditions [6, 7]. This study is loads and increased load bearing capacity therefore aimed to study the influence of while maintaining fluid film lubrication counter surface roughness and operating under certain operating condition. However, conditions on frictional behavior of selected this not only affects the bearing performance polymers in a water lubricated journal in the hydrodynamic lubrication regime, but bearing configuration. it can also influence the tribological performance of the system during start-up and stop when a mechanical contact between C.2. Experimental Work shaft and bearing surfaces is inevitable. In dry sliding contacts, Mofidi and Prakash C.2.1. Materials [2] showed that a reduced friction can be Tribological studies were carried out using obtained with elastomers in sliding against a four different unfilled thermoplastic rougher counter surface. Zsidai et al. found polymers namely polytetrafluoroethylene

(PTFE), polyether ether ketone (PEEK), C.2.2. Test specimens polyethylene terephthalate (PET) and ultra- The test shafts were machined from rods and high molecular weight polyethylene polished to the desired surface roughness (UHMWPE). The characteristics of the using various grades of SiC sand paper materials are listed in Table 1. (#800P-#4000P). The selection of the polymers was based on Surface roughness measurements were the results obtained from an earlier study on carried out using a stylus profilo-meter. The tribological behavior of polymers in water measurements were performed along the lubricated contacts [8]. shaft periphery’s four quadrants in the axial The shafts were made from Inconel 625 with direction. Table 3 shows the shaft roughness chemical composition given in Table 2. characteristics averaged from four Inconel exhibits enhanced tribo-corrosion measurements. Plain polymer bearings were resistance in aqueous environments. It is a machined from polymer rods and press- preferred material for hydraulic piston rods fitted into stainless steel bushings (Fig. 1). in offshore applications exhibiting high The bearing linings were prepared with performance with no failures even after different wall thicknesses for each material, twenty years of operation in marine considering the large differences in elastic environment [9]. It is also a preferred moduli of the polymers (0.4 - 3.6 GPa). This material over stainless steel for application was done to achieve similar lining as the shaft sleeve in water lubricated stern deformations and contact pressures in the tube bearings. loaded region of the bearings.

Table 1: Characteristics of the polymer materials obtained from the materials supplier

Density Elastic Tensile Yield Elongation Hardness Melting Material [g/cm3] Modulus Strength Strength at Break [MPa] point [ºC] [GPa] [MPa] [MPa] [%] UHMWPE 0.93 0.8 23 23 >450 65 135 PET 1.38 3.4 90 90 50-80 170 225 PTFE 2.15 0.4 25-36 30 400 30 300-310 PEEK 1.30 3.6 70 100 50 M99 334

Table 2: Chemical composition of Inconel 625

Element Ni Cr Fe Mo Nb & Ta C Mn Si P S Al Ti Co Content % > 58 20-23 < 5 8-10 3.15-4.15 < 0.1 < 0.5 < 0.5 < 0.015 < 0.015 < 0.4 < 0.4 < 1

Table 4 shows the polymer lining This allowed studying the influence of shaft dimensions, deformations and contact roughness on evolution of friction during the pressures for different materials at specific running-in period as the surfaces bearing pressure of 0.3 MPa. experienced physical/chemical changes. The experimental conditions for obtaining dynamic friction curves during running-in C.2.3. Experimental setup and procedure are shown in Table 5. The experiments were carried out using a small-scale journal bearing test rig. The detailed description of the test rig can be found in [10]. Distilled water was supplied to the clearance of the bearing at atmospheric pressure and room temperature (~23 ºC) through a hole placed opposite to the contact region of the bearing. The load was applied to the bearing housing using a lever arm loaded with dead weights. The friction force was continuously measured by a strain gauge force transducer coupled to the bearing housing. All the shafts and polymer bearings were Figure 1: Machined polymer bearings ultrasonically cleaned with ethanol and dried in air prior to tests. Each bearing was run-in at sliding speed of Table 3: Shaft roughness characteristics, averaged 0.06 m/s (40 rpm) for four hours. Dynamic from four measurements at the quadrantal angles friction curves were obtained with variation Shaft Ra Rt Rz Rq Rpk Rvk Denotation [µm] [µm] [µm] [µm] [µm] [µm] in speed for each material and shaft 0.02R 0.02 0.30 0.19 0.03 0.02 0.05 roughness combinations at different time 0.1R 0.11 0.71 0.63 0.14 0.02 0.32 intervals during running-in. 0.2R 0.22 1.17 1.01 0.25 0.03 0.69 0.4R 0.42 1.83 1.63 0.48 0.02 1.47

Table 4: Bearing dimensions, deformations and contact pressures at load of 141 N corresponding to bearing pressure of 0.3 MPa

Specific Semi- Max Mean Wall Inner Radial Elastic Bearing Length Load bearing contact contact contact thickness diameter clearance Indentation lining [mm] [N] pressure width pressure pressure [mm] [mm] [µm] [µm] [MPa] [mm] [MPa] [MPa] PTFE 2 30.1 15 50 141 0.3 5.4 6.45 1.08 0.73 UHMWPE 3 30.1 15 50 141 0.3 4.5 6 1.18 0.79 PEEK 8 30.1 15 50 141 0.3 3.2 5.1 1.39 0.93 PET 8 30.1 15 50 141 0.3 3.2 5.1 1.36 0.91

50 Breakaway 40 Friction

Friction Force [N] Dynamic 10 Friction

0 0 500 1000 1500 2000 2500 Data points One start/stop cycle

Figure 2: A typical friction data stream obtained with start/stop cycles

Table 5: Experimental conditions for running-in tests obtained for each material, shaft roughness Radial Load 141 N and pressure combination. This was Specific Bearing 0.3 MPa achieved through intermittent motion of the Rotational Speed 40 - 690 rpm shaft (start-stop cycles) at the lowest Sliding Speed 0.06-1.08 m/s practical rotational speed of the tribo-meter. Time Intervals 10, 55, 115, 175, 235 min Figure 2 gives an example of the data stream Bearing Materials PTFE, PEEK, PET, from a typical intermittent test showing a Shaft Ra Roughness 0.02, 0.1, 0.2, 0.4 µm number of start/stop cycles. Breakaway Lubricant Distilled Water Lubricant R.T. (23 ºC) friction was determined from the maximum static friction occurring at the initiation of relative motion during each start/stop cycle. Table 6: Experimental conditions for intermittent An average of one hundred peaks over the tests for obtaining friction maps length of each test was used to determine the Load 141, 224, 336, 449, 562, breakaway friction of the materials. A full Specific Bearing 0.3, 0.5, 0.75, 1, 1.25, 1.5 description of the test conditions for the Rotational Speed 40 - 690 (rpm) intermittent tests is provided in Table 6. Sliding Speed 0.06-1.08 m/s At least three repetitions were carried out for Bearing Materials PTFE, PEEK, PET, each test combination throughout this study Shaft Ra Roughness 0.02, 0.1, 0.2, 0.4 µm and the results were averaged from these test No. of Cycles 100 runs. Cycle Duration 12 Sec Test Duration 20 min Lubricant Distilled Water C.3. Results Lubricant R.T. (~23 ºC) C.3.1. Friction evolution during running-in Figure 3 shows the dynamic friction Following the running-in of the bearings, behavior of the materials versus sliding breakaway and dynamic friction maps were speed at different stages of running-in using

0.25 0.25 PEEK PEEK Ra=0.02 µm Ra= 0.4 µm 0.20 0.20

0.15 10 min 0.15 10 min µ 55 min µ 55 min 0.10 115 min 0.10 115 min 175 min 175 min 0.05 235 min 0.05 235 min

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

0.30 0.30 PET PET 0.25 Ra=0.02 µm 0.25 Ra=0.4 µm

10 min 0.20 10 min 0.20 55 min 55 min 0.15 0.15 115 min µ 115 min µ 175 min 0.10 175 min 0.10 235 min 235 min 0.05 0.05

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

0.16 0.16 PTFE PTFE 0.14 Ra=0.02 µm 0.14 Ra=0.4 µm 0.12 0.12 0.10 10 min 0.10 10 min 55 min 55 min 0.08 0.08 µ 115 min µ 115 min 0.06 175 min 0.06 175 min 0.04 235 min 0.04 235 min 0.02 0.02 0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

0.10 0.10 UHMWPE UHMWPE Ra=0.02 µm Ra=0.4 µm 0.08 0.08

0.06 10 min 0.06 10 min µ 55 min µ 55 min 0.04 115 min 0.04 115 min 175 min 175 min 0.02 235 min 0.02 235 min

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

Figure 3: Dynamic friction behavior of tested materials at different time stages of the running-in process (specific bearing pressure= 0.3 MPa, test duration=4 h, shaft Ra roughness=0.02 and 0.4 µm)

92 shafts with Ra surface roughness of 0.02 µm exhibited a lower initial friction followed by and 0.4 µm. The hump in friction observed 50-75% reduction in friction during running- at sliding speed of ~0.55 m/s is due to the in when using the smoothest shaft. A vibrations caused by approaching natural marginal reduction in friction during frequency of the test rig at specific rotational running-in (approx. 10%) was obtained speed of the motor. when the experiments were carried out with In general, all the materials showed a trend the roughest shaft. of decreasing friction coefficient during the After the running-in period, frictional running-in period and this trend was behavior of the materials obtained with consistent with shafts of different surface smooth and rough shafts was significantly roughness values. different. Figure 4 shows the average However, a considerably larger reduction in dynamic friction curves obtained for each friction during running-in was obtained material after four hours of running-in using when experiments were carried out using the shafts of different surface roughness. smoothest shaft (Ra=0.02µm) compared to UHMWPE bearings provided almost the case with the roughest shaft unchanged friction with variation in sliding (Ra=0.4µm). PEEK, PTFE and PET speed or shaft surface roughness.

0.20 0.30 PEEK PET 0.16 0.25 0.20 0.12 0.02 µm 0.02 µm µ 0.1 µm 0.15 0.1 µm µ 0.08 0.2 µm 0.2 µm 0.10 0.4 µm 0.4 µm 0.04 0.05

0.00 0.00 0.06 0.13 0.26 0.51 1.02 0.06 0.13 0.26 0.51 1.02 Sliding Speed [m/s] Sliding Speed [m/s]

0.16 0.10 PTFE UHMWPE 0.14 0.08 0.12

Figure 4: Dynamic friction behavior of materials after four hours of running-in (specific bearing pressure=0.3 MPa)

However, it was a different case for the trend or a significant change in dynamic remainder of the materials; a decreased shaft friction could be observed with further roughness not only reduced friction at high variations in pressure or shaft surface sliding speeds but also led to a considerable roughness. At breakaway, PEEK exhibited a reduction in friction in the boundary/mixed different frictional response to variations in lubrication regime, simultaneously shifting shaft roughness and contact pressure with a the friction curves down and to the left. trend of decreasing friction with increasing As shown in Figure 4, a considerable shaft roughness and bearing pressure. reduction in friction was obtained by reducing the shaft surface roughness from UHMWPE. UHMWPE showed similar 0.4 µm to 0.1 µm; however, a further trends in dynamic and breakaway friction reduction in friction by using a smoother maps. Variations in shaft roughness led to shaft (Ra=0.02 µm) was marginal and less marginal changes in friction while a trend of significant. decreasing friction with increasing pressure could be observed with more pronounced C.3.2. Friction Maps effects occurring at breakaway compared to Breakaway and dynamic friction maps were dynamic friction. obtained for all materials after the four hour running in period. More than 300 tests were PTFE. PTFE exhibited a reduction in carried out to investigate the influence of friction with reduced shaft surface bearing pressure and shaft surface roughness roughness. More pronounced effects of on frictional behavior of polymers in water. pressure on friction were observed with high Figure 5 shows the results of average shaft roughness values; however, in general breakaway and dynamic friction coefficients a trend of decreasing friction was found with of different polymeric materials obtained increasing pressure for PTFE. with variation in specific bearing pressure and shaft surface roughness. The friction PET. The dynamic friction map of PET is trends observed for different materials are not included in Figure 5 due to the large described as follows: scatter in the results. However, in accordance with the breakaway friction PEEK. At the lowest specific bearing found for PEEK, PTFE and UHMWPE, a pressure, PEEK exhibited a trend of trend of decreasing friction coefficient could decreasing friction with decreasing shaft be observed with increasing pressure for surface roughness. The reduced combined PET. surface roughness increases film parameter In general, a trend of decreasing friction was and enhances lubrication of the surfaces obtained with increasing contact pressure for when using a smoother shaft. However, with all materials while frictional response to an increase in load, a considerably higher variations in shaft surface roughness varied friction was obtained; while no consistent from one material to another.

Figure 5: Average dynamic and breakaway friction coefficient of materials after four hours of running-in versus specific bearing pressure and shaft surface roughness. The results are averaged from at least three test runs (300 start/stop cycles) for each combination. Error bars show one standard deviation at the measured points. The dynamic friction map of PET is not shown due to the large scatter in results.

A trend of increasing friction was observed found on metal surfaces, [8] due to their low with increasing shaft roughness for PTFE coverage and discrete distribution they do whereas an opposite trend was obtained with not seem to hinder direct metal and polymer PEEK at breakaway. No significant contact. While many polymers are generally dependence of friction on shaft roughness regarded as chemically resistant, their was found with UHMWPE or PET. The interaction with metallic counter-parts in a various trends observed in friction of the tribological system cannot be neglected. polymers with variations in shaft roughness Frictional heat and pressure at asperity- may be explained by the contribution of asperity contact can enhance the conditions adhesive and deformation components of for chemical reaction between polymers and friction for each material considering their metal counter-faces in water. vast differences in elastic modulus and In earlier studies on tribo-chemical reaction surface free energy. of PTFE with metal counter-faces [11, 12, These results indicate that although using a 13], formation of metal fluorides were shaft of reduced surface roughness can detected using XPS when PTFE was rubbed favorably decrease the dynamic friction of against aluminum, stainless steel and nickel polymer bearings; it may adversely affect counter-parts. Deposition of the reaction the materials’ breakaway friction and products and formation of such tribo-films thereby increase the torque required for on metal surfaces can influence the machine start-up; a critical issue in some wettability and surface free energy of the applications such as pumped storage metal surfaces and thus may alter the hydropower plants. lubricating and frictional behavior of the surfaces. To investigate the latter, screening tests C.4. Discussions were carried out to examine the possible C.4.1. Running-in behavior changes in wettability of Inconel counter- In dry polymer/metal contacts, the reduction faces when sliding against the polymeric in friction during running-in is mainly materials. attributed to the formation of a polymer Due to the practical limitations for contact transfer layer on metal counter-faces. The angle measurements on curved shaft transfer layer partially inhibits direct contact surfaces, these measurements were carried between hard metal asperities and soft out on flat Inconel plates which were pre- polymer surfaces and generally results in rubbed against the polymers in a simplified reduction of friction and wear rate as sliding test configuration. progresses. In lubricated conditions, Cylindrical polymer pins were rubbed however, formation of such continuous against Inconel plates using a reciprocating transfer layers is affected by the presence of Cameron-Plint tribo-meter. Prior to rubbing, lubricant in the contact. Although some the metal plates were polished to Ra surface micro-patches of polymer transfer may be roughness of 0.03 µm in order to minimize

100 95 90 85 Unworn Inconel 80 Unworn Polymer

Degrees 75 Inconel worn by polymer 70 65 60 PEEK PET PTFE UHMWPE

Figure 6: Water contact angles of unworn Inconel, unworn polymer and worn Inconel surfaces on the wear track

the effect of surface roughness on contact Figure 6 shows the results of water contact angle measurements [14]. The experimental angle measurements for unworn and worn conditions for preparation of Inconel plates Inconel plates and polymer materials. for contact angle measurements are The results show a significant reduction in summarized in Table 7. hydrophobicity of Inconel surfaces when After tribo-testing, the Inconel plates were sliding against the polymers in water. gently wiped by a wet cloth to remove the The considerably lower water contact angle loose wear particles from the wear track. of worn Inconel surfaces from that of Contact angle measurements were carried unworn Inconel or polymers suggests out on polymer and Inconel surfaces using formation of a reaction layer on Inconel distilled water. Sessile drops of 4µL were surfaces rather than deposition of a physical used and contact angles were recorded after polymer transfer layer. Formation of a one second of drop deposition on the reaction layer on the Inconel counter-face surfaces. At least six measurements were could have partially contributed to the performed for each sample and the results reduction of friction during running-in. were averaged from these measurements. However the significant differences in the magnitude of friction reduction when using shafts of different surface roughness suggest Table 7: Experimental conditions to prepare the that the dominant mechanism for the trend is Inconel plates for contact angle measurements Load 20 N not only the result of formation of a reaction Temperature R.T. (21-23 ºC) layer and possible changes in surface free Stroke length 5 mm energy of Inconel counter-faces. Frequency 7 Hz It is well known that one of the major Test Duration 4 hours surface characteristics which can affect the Total Sliding Distance 1008 m tribological behavior of tribo-pairs is lnconel Ra Roughness 0.03 µm topography of the contacting surfaces [15, Lubricant Distilled Water 16]. Roughness measurements of polymer

5 Unworn

Worn by shaft Ra=0.02 µm 4 Worn by shaft Ra=0.1 µm

3 Worn by shaft Ra=0.2 µm

Worn by shaft Ra=0.4 µm 2 Ra Ra [µm] Roughness 1

0 PTFE UHMWPE PEEK PET

Figure 7: Average Ra roughness of bearing surfaces before and after sliding against shafts with different roughness values surfaces in the contact region of the bearings Film parameter (λ), which is defined as the before and after the tests showed a ratio of minimum fluid film thickness and considerable reduction in the surface the combined roughness of the rubbing roughness of PEEK, PTFE and PET. Figure surfaces, is generally considered as one of 7 shows the average Ra roughness of the the most relevant parameters to characterize polymer bearings before and after sliding lubrication of sliding surfaces. Although the against shafts of different surface roughness magnitude of the film parameter may not values. It can be seen that no significant specifically be associated with a lubrication change in roughness characteristics of regime it serves as a strong comparative tool UHMWPE could be observed due to its to assess the lubrication status and severity relatively high wear resistance. of the mechanical interaction of contacting At low sliding speeds, when the load is surfaces in lubricated conditions. mainly supported by asperity-asperity Smoothening of the bearing surfaces during contacts, smoothening of the polymer running-in leads to a reduction in combined bearings can reduce the mechanical surface roughness and thus results in an interaction between the asperities of shaft increase in the film parameter at a specific and bearing surfaces. In addition this leads operating condition. This promotes to a reduction in real contact pressure separation of the shaft and bearing surfaces between the asperities of the mating surfaces at lower rotational speeds; shifting the which allow for water to remain between the friction curves to the left. contacting asperities and act as a boundary As shown in Figure 3, PEEK, PTFE and lubricant. PET exhibited considerable reduction in At higher sliding speeds, the influence of the friction at high sliding speeds but no such surface roughness is even more pronounced reduction in friction could be observed in due to its key role in determining the the case of UHMWPE. UHMWPE is well lubrication regime between the contacting known for its abrasion resistance among surfaces. thermoplastic polymers. This is an

98 advantage for its application as a bearing a more severe mechanical interaction material but can adversely affect its running- between the shaft and bearing surfaces in behavior. The initially rough machined occurs. This results in friction levels known surfaces of UHMWPE bearings as well as for the polymers in the boundary lubrication its wear resistance are likely to hinder regime. This is clearly observed in the smoothening of the surface asperities. This dynamic friction map of PEEK, where an will considerably reduce the hydrodynamic increase in pressure from 0.3 MPa to 0.5 effects leading to rather unchanged friction MPa results in a sudden jump in friction for of UHMWPE bearings even at high sliding low shaft roughness values. In the case of speeds. PTFE this transition was not readily This highlights the importance of the as- observed due to low friction levels known made surface finish of the polymer liners of for this material in the boundary lubrication sliding bearings especially for wear resistant regime. bearing materials. Low friction and quick According to the generally accepted friction separation of the shaft and bearing surfaces models, two main mechanisms contribute to at the very first start-up is crucial to avoid the friction of polymer-metal sliding excessive frictional heating, melting and contacts. These mechanisms include the possible catastrophic bearing failure. adhesion component which stems from shearing of the interfaces formed by the C.4.2. Influence of shaft roughness The friction curves obtained at low bearing pressures (Fig. 4) showed a considerably lower friction when using a smoother shaft. This is attributed to the increased film parameter resulting from reduced combined roughness of shaft and bearing surfaces. Although characterization of worn polymer surfaces after running-in showed no significant difference in surface roughness of polymers when sliding against different shafts, a lower combined roughness is obviously obtained with a smoother shaft. This allowed bearings to operate with low friction when a combination of low shaft roughness and low pressure was applied. Figure 8: Roughness profiles of Figure 8 shows roughness profiles of PEEK a) Shaft Ra=0.02 µm b) Shaft Ra=0.4 µm surfaces before and after sliding against c) Worn PEEK sliding against shaft of Ra=0.02 µm shafts of Ra=0.02 µm and Ra=0.4 µm. d) Worn PEEK sliding against shaft of Ra=0.4 µm At high pressures, however the lubricant is e) Unworn PEEK more readily squeezed out of the contact and

99 asperity contacts, and the deformation C.4.3. Influence of pressure component resulting from ploughing action While materials’ frictional response to for tribo-pairs with significantly different variations in surface roughness was varied; elastic moduli. Using shafts of different in general, a trend of decreasing friction was surface roughness can influence both observed with increasing pressure for all the deformation and adhesion components materials. through changes of the real area of contact Variation in applied load also affects both between the mating surfaces. A reduction in deformation and adhesion components of the real area of contact increases the asperity friction. With an increase in normal load the contact pressure and favors the deformation real area of contact increases due to the component whereas an increased real area of increased deformation of the surface contact favors the adhesion component of asperities. However, as the asperities friction. The opposite trends observed in undergo deformation, their surface friction behavior of PEEK and PTFE as a characteristics may change depending on the function of shaft roughness may be contact pressure and the type of the explained by the contribution of these materials. components to the friction of each material, In an earlier study, Maeda et al. showed the considering the large difference in elastic significant influence of bulk deformation of moduli and surface free energies of PEEK 2 rubber on its adhesion and surface free (E=3.6 GPa, γ≈42 mJ/m ) and PTFE (E=0.4 energy characteristics [17]. Contact pressure 2 GPa, γ≈19 mJ/m ). Similar trends were also can also affect the shear strength of reported earlier by Quaglini et al. [3] in dry polymeric materials [18, 19]. Benabdallah sliding contacts showing that polymers with [20] experimentally investigated the low elastic moduli exhibited better sliding influence of real contact pressure on shear behavior against smooth counter faces strength of some thermoplastics identifying whereas for high-modulus polymers lower three distinctive regions for UHMWPE. The friction could be obtained when sliding first of which was characterized by a against rougher counter surfaces. reduction in shear strength with increasing While a clear trend in breakaway friction of pressure. The second was a transition region PEEK could be observed with variations in where real contact pressure was equal to the shaft roughness; no such trend was found in yield strength of the material, and the last dynamic friction of PEEK. This may be region was characterized by increasing shear attributed to the presence of a layer of water strength with increasing pressure occurring molecules between the contacting surfaces at real contact pressures larger than the yield during continuous sliding. This layer can strength of UHMWPE. This behavior disturb short range molecular adhesive however was not consistently observed with forces between the mating surfaces and lead other polymeric materials e.g. PA66 showed to a reduction in the contribution of the an increase in shear strength with increasing adhesive component of friction in the real contact pressure even at pressures below dynamic friction map of PEEK. the yield strength.

100

The various surface and bulk responses of the smoothest shaft. A higher initial friction different polymers to variations in load and marginal reduction in friction during make it almost impossible to attribute the running-in (~10%) was observed with the trend observed to a single mechanism. roughest shaft. However, one should not neglect the possible influence of pressure on formation 3. Contact angle measurements revealed a of tribo-films on Inconel counter-surfaces. significant increase in wettability of Inconel Increased pressure may favorably enhance counter faces. This suggests formation of a formation of tribo-films on metal surfaces reaction layer on worn Inconel surfaces which can in turn alter their surface free when sliding against the polymers. energy and thus influence the frictional response of the tribo-pairs. However 4. The friction maps showed a trend of determination of the extent and significance decreasing friction with increasing pressure; of the latter on frictional behavior of however the materials’ frictional response to polymers requires further investigation. changes in counter surface roughness were varied.

C.5. Conclusions 5. A trend of increasing friction was In this study, the frictional behavior of four observed with increasing shaft roughness for unfilled thermoplastic polymers has been PTFE whereas an opposite trend was investigated using a water lubricated journal observed for PEEK at breakaway. No bearing. The influence of shaft surface significant dependence of friction on shaft roughness and contact pressure on frictional roughness was found for UHMWPE and behavior of these materials was studied. PET. Dynamic friction curves as well as dynamic and breakaway friction maps have been 6. A shaft of lower surface roughness may obtained for different shaft roughness and be beneficial for low dynamic friction of pressure combinations. The mechanisms polymers in all lubrication regimes; involved were discussed. The major findings however, it can adversely affect frictional of this work can be summarized as follows: response at breakaway.

1. In general, all the materials showed a trend of decreasing friction coefficient Acknowledgements during the running-in period and this trend The research presented in this paper has was consistent for the shafts of different been funded by “StandUp for Energy”. surface roughness. StandUp for Energy is a collaborative initiative between Uppsala University, The 2. For PEEK, PTFE and PET, a lower initial Royal Institute of Technology (KTH), The friction followed by 50-75% reduction in Swedish University of Agricultural Sciences friction was observed during running-in with (SLU) and Luleå University of Technology

101

(LTU). It arose as a result of the Wear, 1994, Vol. 175, Issue 1-2, pp. 219- Government’s commitment to high quality 225. research in areas of strategic importance to [8] Golchin A., Simmons G., Glavatskih S., society and the business sector. Prakash B., Tribological Behaviour of Polymeric Materials in Water Lubricated Contacts, Proceedings of the Institution of Reference Mechanical Engineers, Part J: Journal of [1] Kraker A., Ostayen R., Rixen D., Engineering Tribology, in press. [9] D Landolt and S Mischler, Calculation of Stribeck Curves for (Water) Tribocorrosion of Passive Metals and Lubricated Journal Bearings, Tribology Coatings, Woodhead Publishing, International, 2007, Vol. 40, pp. 459-469. Cambridge, 28 Dec. 2011. [2] Mofidi M., Prakash B., Influence of [10] Peels J.A., Meesters C.J.M., Ontwerp counterface topography on sliding friction en constructieve aspecten van een 2kN- and wear of some elastomers under dry glijlageropstelling, De constructeur, 1996, sliding conditions, Proceedings of the No. 3, pp. 40-43. Institution of Mechanical Engineers, Part J: [11] Gong D., Xue Q., Wang H., ESCA Journal of Engineering Tribology, 2008, Study on Tribochemical Characteristics of Vol. 222, pp. 667-673. [3] Quaglini V. et al., Influence of Filled PTFE, Wear, 1991, Vol. 148, pp. 161- Counterface Roughness on Friction 169. Properties of Engineering Plastics for [12] Cadman P., Gossedge G. M., The Bearing Applications, Materials and Design, Chemical Interaction of Metals with 2009, Vol. 30, pp. 1650-1658. Polytetrafluoroethylene, Journal of Material [4] Zsidai L. et al., The tribological Science, 1979, Vol. 14, pp. 2672-2678. [13] Cadman P., Gossedge G. M., The behaviour of engineering plastics during Chemical Nature of Metal- sliding friction investigated with small-scale Polytetrafluoroethylene Tribological specimens, Wear, 2003, Vol. 253, pp. 673- Interactions as Studied by X-ray 688. Photoelectron Spectroscopy, Wear, 1979, [5] Santner E., Czichos H., Tribology of Vol. 54, pp. 211-215. polymers, Tribology International, 1989, [14] Busscher H. J., Van Pelt A.W. J., De Vol. 22, Issue 2, pp. 103-109. Boer P., De Jong H. P., Arends J., The [6] Lloyd A.I.G., Noel R.E.G., The effect of Effect of Surface Roughenning of Polymers counterface surface roughness on the wear on Measured Contact Angles of Liquids, of UHMPWE in water and oil-in-water Colloinds and Surfaces, 1984, Vol. 9, pp. emulsion, Tribology International, 1988, 319-331. Vol. 21, Issue 2, pp. 83-88. [15] A.I.G. Lloyd, R.E.G. Noel, The Effect [7] Fisher J. et al., The effect of sliding of Counterface Surface Roughness on the velocity on the friction and wear of Wear of UHMWPE in Water and Oil-in- UHMWPE for use in total artificial joints,

102

Water Emulsion, Tribology International, Vol. 21, Issue 2. Apr. 1988, pp. 83-88. [16] C. G. Clarke, C. Allen, The Water Lubricated, Sliding Wear Behaviour of Polymeric Materials against Steel, Tribology International, Vol. 24, issue 2, Apr. 1991, pp. 109-118. [17] Maeda K., Bismarck A., Briscoe B., Effect of Bulk Deformation on Rubber Adhesion, Wear, May 2007, Vol. 263, pp. 1016-1022. [18] Briscoe B. J., Tabor D., Shear Properties of Thin Polymeric Films, Journal of Adhesion, 1978, Vol. 9, pp. 145-155. [19] Briscoe B. J., Tabor D., The effect of Pressure on the Frictional Properties of Polymers, Wear, 1975, Vol. 34, pp. 29-38. [20] Benabdallah S. H., Shear Strength Resulting from Static Friction of Some Thermoplastics, Journal of Material Science, 1993, Vol. 28, pp. 3149-3154.

103