SUKUMARAN - Revisiting-polymer-tribology-for-heavy-duty-application_2017_Wear

Prévia do material em texto

<p>Wear 376-377 (2017) 1321–1332</p><p>Contents lists available at ScienceDirect</p><p>Wear</p><p>http://d</p><p>0043-16</p><p>n Corr</p><p>journal homepage: www.elsevier.com/locate/wear</p><p>Revisiting polymer tribology for heavy duty application</p><p>Jacob Sukumaran a,n, Jan De Pauwa, Patric D. Neis b, Levente F. Tóth a, Patrick De Baets a</p><p>a Ghent University, Department of Electrical Energy, Systems & Automation, Soete Laboratory Technologiepark 903, B-9052 Zwijnaarde, Ghent, Belgium</p><p>b Laboratory of Tribology, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil</p><p>a r t i c l e i n f o</p><p>Article history:</p><p>Received 13 September 2016</p><p>Received in revised form</p><p>2 January 2017</p><p>Accepted 4 January 2017</p><p>Keywords:</p><p>Heavy duty testing</p><p>Polymer tribology</p><p>Transfer layer</p><p>Amorphous polymer</p><p>x.doi.org/10.1016/j.wear.2017.01.018</p><p>48/& 2017 Published by Elsevier B.V.</p><p>esponding author.</p><p>a b s t r a c t</p><p>Polymers are frequently considered as a candidate material for heavy duty tribological applications. Standard</p><p>tribological testing for such applications may result in unrealistic tribological properties in terms of wear me-</p><p>chanism, transfer film formation and friction coefficient etc. Thus the present research aims to explore the tri-</p><p>bological behaviour of currently available neat polymers under heavy duty loading condition. In this background,</p><p>nine different polymers (PAI, PEI, PC, PPSU, PA6, PET, PPS, PVDF, and UHMWPE) of amorphous and semi-crys-</p><p>talline groups were chosen from engineering and high performance grades. Large scale reciprocating slidingwear</p><p>tests with 50�50mm contact area were used. Considering heavy duty loading condition, the operational</p><p>parameter were idealised at 10 kN normal force and 50mm/s sliding speed. From the test results the polymers</p><p>are classified into three groups based on the wear mechanisms and the transfer layer characteristics observed</p><p>from the counterface. Group 1 did not persist any transfer layer (PC and PEI), whereas Group 2 materials (PET,</p><p>PPS, PPSU and PAI) has lumpy discontinuous transfer layer and finally Group 3 materials evidenced uniform thin</p><p>transfer layer. Clear signs of three body abrasionwere evidenced in Group 1material which is due to the absence</p><p>of transfer layer and also from the large loose debris. Poor adhesion of inhomogeneous lumpy transfer layer in</p><p>Group 2 showed an intermediate wear rate. Based on the wear and friction performance, the Group 3 materials</p><p>(PVDF, UHMWPE and PA6) are proposed as the candidate materials for heavy duty tribological application. In</p><p>regards to the transfer layer thickness and wear mechanisms at large scale testing, the present observations</p><p>showed incomparable results with the literature. This research concludes that the revisit to polymer tribology at</p><p>high load condition has indicated the difference in wear mechanism and the corresponding tribological char-</p><p>acteristics in large scale testing. As a future work multi-scale testing is an absolute must for the complete un-</p><p>derstanding of tribological characteristics w.r.t a specific material.</p><p>& 2017 Published by Elsevier B.V.</p><p>1. Introduction</p><p>Based on economic and ecological concern dry sliding contacts</p><p>(bearing application) based on polymers are preferred over oil</p><p>lubricated metallic contacts. This transition from metals and its</p><p>alloys towards polymer started since early 40s’ [1]. A range of</p><p>applications from small domestic appliances to heavily loaded</p><p>bearings use polymer as a tribological pair. This research focuses</p><p>more on heavily loaded plain bearings which is often critical while</p><p>considering the energy consumption and the service life. The</p><p>suitability of the polymers as plain bearing were investigated</p><p>earlier for applications such as guide vane bearings, bearing pads</p><p>in bridges, sea gates, ball bearing cage, heavily loaded ball joint</p><p>and liners in helicopter rotor control [2–9]. The focus of the pre-</p><p>sent research is based on two big questions concerning heavy load</p><p>application in polymer tribology i) lack of recent comparative</p><p>study from the view point of wear mechanism and materials and</p><p>ii) tribological property of amorphous grades at large scale.</p><p>Traditionally the tribological characterisation of polymers were</p><p>performed using standard tests such as block-on ring (ASTM G 176),</p><p>pin-on-disc (ASTM G 99) and pin abrasion test (ASTM G 132) [10–12].</p><p>Though several new results are published in the polymer tribology,</p><p>the results are based on standard testing. However, with standard</p><p>testing, the transfer layer characteristics and their role in altering their</p><p>tribological performance are questionable. In a previous investigation</p><p>on 18 different polymers using block-on ring configuration, no visible</p><p>transfer layer was observed (except for PPS) [13]. However, it is no-</p><p>teworthy to mention that the transfer layer play a significant role in</p><p>altering the tribological properties [14–19]. The few studies focused on</p><p>transfer layer formationwere also typically studied using standard test</p><p>methodologies. The observation of transfer layer was rather local and</p><p>qualitative, which may not be a good representative to study its ef-</p><p>fects. Moreover, limitation persist either in the area of contact or the</p><p>loading conditions where the forces are significantly low when com-</p><p>pared with real application [18–21]. Literature clearly points out that a</p><p>www.sciencedirect.com/science/journal/00431648</p><p>www.elsevier.com/locate/wear</p><p>http://dx.doi.org/10.1016/j.wear.2017.01.018</p><p>http://dx.doi.org/10.1016/j.wear.2017.01.018</p><p>http://dx.doi.org/10.1016/j.wear.2017.01.018</p><p>http://crossmark.crossref.org/dialog/?doi=10.1016/j.wear.2017.01.018&domain=pdf</p><p>http://crossmark.crossref.org/dialog/?doi=10.1016/j.wear.2017.01.018&domain=pdf</p><p>http://crossmark.crossref.org/dialog/?doi=10.1016/j.wear.2017.01.018&domain=pdf</p><p>http://dx.doi.org/10.1016/j.wear.2017.01.018</p><p>Table 1</p><p>Mechanical and thermal properties of the semi-crystalline and amorphous polymer.</p><p>Polymer Semi-crystalline Amorphous</p><p>Polymer type and commercial grade PET UHMWPE PA6 1023PVDF PPS PAI PPSU PEI PC</p><p>ERTALYTE TIVAR 1000 9 XAU PVDF 1000 PPS1000 T4203 PPSU U1001 PC 1000</p><p>Density (g/cm3) 1.39 0.93 1.15 1.78 1.35 1.41 1.29 1.27 1.2</p><p>Water absorption at saturation (at 23 °C) 0.5 o0.1 6.5 o0.10 0.1 4.4 1.1 1.3 0.4</p><p>Melting temperature (DSC, 10 °C/min) 245 135 215 175 280 NA NA NA (-)</p><p>Glass transition temperature (DSC, 20 °C/min) (-) (-) (-) (-) – 280 220 215 150</p><p>Thermal conductivity at 23 °C (W/(K m) 0.29 0.4 0.29 0.19 0.3 0.26 0.3 0.24 0.21</p><p>Max. allowable service temperature in air (for short periods) 160 120 180 160 260 267 210 200 135</p><p>Max. allowable service temperature in air (20,000 h) 100 80 105 150 252 250 180 170 120</p><p>Tensile strength (Mpa) 90 (-) 86 60 102 150 83 129 74</p><p>Tensile modulus of elasticity (Mpa) 3500 750 3500 2200 4000 4200 2450 3500 2400</p><p>Charpy impact strength - notched (kJ/m2) 2 115P 3 10 2 15 12 3.5 9</p><p>Ball indentation hardness(N/mm2) 170 33 165 110 205 200 95 165 120</p><p>As provided by the supplier.</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–13321322</p><p>completely different wear mechanism can be experienced between</p><p>low and high load conditions [22]. In standard testing, the evacuation</p><p>of debris from the contact zone is predominant than large scale testing</p><p>where the debris trapped in the contact zone forms a transfer film</p><p>[23]. This is due to the limitation in the scale difference between the</p><p>contact area and the size of the generated debris. Hence, the standards</p><p>may not fully correspond to the aforementioned applications. Test-</p><p>systems closely representing the real applications were also developed</p><p>for tribological testing at large scale [3,4,24,25].</p><p>Parallel to the scale of testing selecting an appropriate candidate</p><p>material is also important. In the 70s’ it was realised that the tribological</p><p>ranking of commercially available polymer grades would enable the</p><p>designers to make efficient choice in the material selection process [25].</p><p>The recent comparative and an explorative study on polymer friction</p><p>and wear characteristics dates back to the early 90s’ where Mens et al.</p><p>studied 18 different polymers.</p><p>The polymer research have grown in the</p><p>last few decades confirming to improved mechanical and thermal</p><p>properties [23–30]. The improvement in mechanical and thermal</p><p>properties will be well reflected in their tribological performance. In the</p><p>past, very few materials have been tested in large scale, often the</p><p>thermosets are the target groups [26]. Furthermore a decade ago few</p><p>reports on large scale testing were published on PA, PET, PI and</p><p>UHMWPE [3,4,13]. The recent tribology test programs frequently fol-</p><p>lows a small scale testing, hence the tribological output of recently</p><p>synthesised advanced polymer has not been studied in detail at large</p><p>scale. With the development of material synthesis in the past few</p><p>decades and limited reports on real scale studies, it is necessary to re-</p><p>implement the large scale testing for the recent polymers.</p><p>The present research attempts to extract the tribological ad-</p><p>vantage of the present day polymers and their potential in real</p><p>scale application. Understanding the difference in the governing</p><p>wear mechanism between the largescale and small scale testing</p><p>and considering the availability of literature in small scale testing</p><p>the present research will primarily focus on large scale testing. In</p><p>this background, commercially available thermoplastics pertaining</p><p>to the different groups such as engineering and high performance</p><p>plastics will be tested. Attempts were made to clearly reason the</p><p>performance of amorphous grades in real scale application. In</p><p>addition to their mechanical and thermal characteristics, the in-</p><p>ability of amorphous polymers on forming transfer layer has been</p><p>seen as a limitation. However considering the scarce reports on</p><p>amorphous grades the tribological characteristics at heavy loded</p><p>conditions will also be studied for amorphous polymers in the</p><p>present research. The results from the large scale testing will also</p><p>be compared with the existing results from literature available on</p><p>small scale tests.</p><p>2. Materials and methods</p><p>The scarcity of tribological data from large scale testing of</p><p>thermoplastics inspired the present research to primarily focuses</p><p>on friction and wear characteristics at heavy loads which are fre-</p><p>quently used as bearing materials. Though there are several ad-</p><p>vances in the additives, only neat polymers are considered in the</p><p>present research. This is due to the fact that a reference is required</p><p>w.r.t. the matrix before engaging any additives which may bring in</p><p>additional effects as uncertainties. In this background, nine dif-</p><p>ferent commercially available grades from Quadrant plastics were</p><p>used. Thermoplastics from different groups such as engineering</p><p>plastics and high performance plastics were considered for this</p><p>investigation. Apart from the functional classification, the poly-</p><p>mers were also chosen based on the crystallinity where both</p><p>amorphous and semi-crystalline grades are preferred. Among the</p><p>semi-crystalline grades Techtrons 4208 (Polyphenylene sulphide,</p><p>PPS), SYMALIT s 1000 (Polyvinylidene fluoride, PVDF), ERTALONs</p><p>6 XAU (Polyamide, PA6), TIVARs 1000 (Ultra high molecular</p><p>weight poly ethylene, UHMWPE) and Ertalytes PET-P (poly-</p><p>ethylene terephthalate, PET) are chosen. The amorphous grades</p><p>include Quadrants PC 1000 (polycarbonate, PC), Duratrons PEI</p><p>(polyetherimide, PEI), Quadrants PPSU (polyphenylsulfone, PPSU)</p><p>and Duratrons T4203 (Polyamide-imide, PAI). The polymers were</p><p>provided by Quadrant EPP Belgium [13]. The mechanical and</p><p>thermal properties of the polymers are given in Table 1. In the</p><p>current investigation, the polymers were tested against commer-</p><p>cially available steel 100 Cr6 (DIN 1.3505) counterface. Dynamic</p><p>mechanical analysis (DMA) and differential scanning calorimetry</p><p>(DSC) were employed to study the polymer glass transition tem-</p><p>perature. DMA measurements are performed on a TA Instruments</p><p>DMA Q800 with 3 °C/min temperature rate, 1 Hz frequency and</p><p>with a temperature range between �120 °C and 250 °C. Whereas,</p><p>DSC measurements were performed using TA Instruments DSC</p><p>Q2000 where 5 °C/min temperature rate was used within a tem-</p><p>perature range between �70 °C and 300 °C. Both DMA and DSC</p><p>were performed for untested and wear tested specimen to un-</p><p>derstand if the thermo-mechanical properties has changed as a</p><p>function of heavy loaded wear testing. In DSC, care was taken such</p><p>that the samples were collected from the contact surface.</p><p>In regards to the specimen preparation, blocks with dimension</p><p>50�50�7 mm were machined out from the raw material stock.</p><p>The area of contact is 2500 mm2 which is approximately a factor of</p><p>10 higher than the samples used in traditional pin on disc testing.</p><p>Earlier reports suggest that in specimen preparation, machining of</p><p>polymers maintains uniformity in morphology by eliminating the</p><p>non-spherulitic and transition region on the contact surface</p><p>[31,32], therefore the contact surface of all the polymers were</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–1332 1323</p><p>machined to a depth of 2 mm from the original surface. Further</p><p>physical operation were not performed on the contact surface of</p><p>the polymer. The steel counterface is the reciprocating component</p><p>which is machined to a rectangular block of size</p><p>200�80�20 mm. Subsequently, the contact surface was ground</p><p>perpendicular to the sliding direction. The average surface</p><p>roughness which is measured parallel to the sliding direction is</p><p>approximately Ra 0.20 mm. The roughness was scanned using a</p><p>stylus profilometer (Surfascan 3D roughness tester, Hommel so-</p><p>micronic) with a stylus S6T (radius 2 mm, angle 90°). Ra is calcu-</p><p>lated based on the ISO 4288 standard which has an assessment</p><p>length of lt¼4.00 mm and cut off λc¼0.80 mm for 0.1 μmoRa</p><p>r2 μm. The steel surfaces were cleaned with isopropanol and</p><p>subsequently with acetone in a ultrasonic cleaner prior to wear</p><p>testing.</p><p>2.1. Tribological characterization</p><p>The scale of the testing is an important factor in this research,</p><p>the experiments were designed to perform tests close to real scale.</p><p>The friction and wear characteristics of specimen materials were</p><p>quantified in a large-scale linear reciprocating sliding system</p><p>(large scale flat tribo meter). Simplification were made on the</p><p>aspect of geometry where a flat-on-flat configuration was chosen.</p><p>Fig. 1 depicts the schematic of the test configuration. The illus-</p><p>tration shows that two polymer specimens are forced against the</p><p>central sliding block which is bolted with two counterface material</p><p>on either side of the block. The tests were performed in a condi-</p><p>tioned chamber where a uniform temperature of 23 °C and a re-</p><p>lative humidity of 50% is maintained. More information about the</p><p>large scale flat tribometer can be found elsewhere [33,34]. The</p><p>conventional block-on-ring or pin-on-disc testing has a continuous</p><p>sliding where the static coefficient of friction is not the primary</p><p>focus. However, in design considerations the static coefficient of</p><p>friction is often used. This is resolved by the use of reciprocating</p><p>sliding system where precise observations on the static and dy-</p><p>namic coefficient of friction can be made. Operational parameters</p><p>with 50 mm/s sliding speed in reciprocating sliding motion and</p><p>the normal force 10 kN corresponding to 4 MPa contact pressure</p><p>was used. Data involving friction force, bulk temperature of the</p><p>steel counterface and dimensional change of the polymer were</p><p>acquired online. More information about the placement of sensors</p><p>for acquiring signals on friction force, temperature and dimen-</p><p>sional change can be found elsewhere [33,34]. All measurements</p><p>are recorded online using NI 6036E DAQ (National instruments</p><p>BNC 2100) in a LabVIEW and DEWESOFT platform.</p><p>The coefficient of friction (m) is evaluated from the friction force</p><p>(FFR) and the normal force (FN) according to Eq. (1). In order to</p><p>accurately acquire the static coefficient of friction, a high sampling</p><p>frequency at 20 kHz with continuous sampling was followed. A</p><p>more detailed description about the data processing for static and</p><p>Table 2</p><p>Tribological characteristics of the</p><p>amorphous and semi-crystalline polymer.</p><p>Polymer Wear mechanism Transfer layer</p><p>Amorphous polymer PEI Group 1 Abrasion No transfer layer</p><p>PC</p><p>PAI Group 2 Adhesion and</p><p>abrasion</p><p>Patchy discontinuous sec-</p><p>ondary layerPPSU</p><p>Semi-crystalline</p><p>polymer</p><p>PET</p><p>PPS</p><p>PVDF Group 3 Adhesion Smooth uniform primary</p><p>layerUHMWPE</p><p>PA6</p><p>dynamic coefficient of friction can be found elsewhere [34].</p><p>Equation for the coefficient of friction</p><p>( )μ = ( )F /2 F 1FR N</p><p>where FFR is the friction force [N] and FN is the normal force [N].</p><p>Archard model for the wear rate is universally accepted for all</p><p>materials, however, Shipway mentions that the hardness may not</p><p>be a critical parameter for polymer [32]. Therefore Lewis model</p><p>has been used in the present research [35]. The operating para-</p><p>meters was not considered for calculating the wear rate because,</p><p>the normal force and sliding speed were maintained constant for</p><p>all the tests. The wear rate (k) was calculated based on the ex-</p><p>pression given in Eq. (2). The specimen wear are measured offline</p><p>using the mass loss and thickness reduction. Tests were performed</p><p>to study the uncertainty from the test-setup where a deviation of</p><p>10% in COF and 20% in wear rate was observed. The polymers were</p><p>conditioned before and after testing to remove the moisture pre-</p><p>sent in the specimen. Hence the specimen were heated in an oven</p><p>at 70 °C for 20 hours before and after testing.</p><p>Equation for the specific wear rate</p><p>= Δ ρ ( )k m d. / 2</p><p>where</p><p>Δm – mass loss [g], ρ - density [g/m3], d – sliding distance</p><p>[mm].</p><p>3. Results and discussion</p><p>3.1. Friction and wear behaviour at large scale testing</p><p>The traditional exploratory research on finding a candidate</p><p>material for tribological application has two separate parts 1. the</p><p>tribological characteristics and 2. wear mechanism. The quantita-</p><p>tive values of friction, wear and bulk temperature gets the most</p><p>attention. Moreover, the qualitative information about wear me-</p><p>chanisms are often related to the physical and mechanical prop-</p><p>erties of the polymers to understand its correlation with tribolo-</p><p>gical behaviour. Having known that the rate of wear partly de-</p><p>pends on the wear mechanisms, the present research follows a</p><p>holistic approach where the wear characteristics are related to the</p><p>corresponding wear mechanism and the respective transfer layer</p><p>characteristics. Tribo-tests were performed for all nine materials</p><p>and their corresponding friction, wear, bulk temperature and</p><p>surface characteristics are discussed in detail. For better under-</p><p>standing, the friction and change in bulk temperature are plotted</p><p>as a function of number of cycles. A complete cycle represents a</p><p>full stroke with continuous forward and a backward motion in the</p><p>linear reciprocating system. A total sliding distance of 1 km which</p><p>corresponds to 5000 test cycles was used for measuring detectable</p><p>Static CoF Dynamic CoF Specific wear</p><p>rate</p><p>Bulk</p><p>temperature</p><p>ls [dimensionless] lD [dimensionless] k [m3 mm�1] Avg. [°C]</p><p>0.54 0.48 1.79E-11 148</p><p>0.50 0.49 2.36E-11 176</p><p>0.62 0.54 1.60E-13 171</p><p>0.38 0.27 2.09E-13 158</p><p>0.50 0.38 1.22E-13 195</p><p>0.43 0.44 3.20E-13 227</p><p>0.30 0.24 1.74E-14 155</p><p>0.27 0.21 2.53E-14 127</p><p>0.37 0.28 8.86E-15 169</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–13321324</p><p>wear of polymers. The total sliding distance of 1 km was identified</p><p>from the preliminary testing on PET samples. All the material did</p><p>not reach the planned 5000 cycles. Among the nine materials only</p><p>PC and PEI has failed during the initial period of testing. The point</p><p>of failure is indicated with the red dot in the in Fig. 2(b). The PC</p><p>and PEI are the failed amorphous thermoplastic polymer and did</p><p>not reach the steady state friction. However for comparison pur-</p><p>pose the average of last 100 cycles was considered for friction</p><p>calculation (only from these two materials). PAI and UHMWPE</p><p>have reached the 4703 and 4667 cycles and the tests was stopped</p><p>without any failure. Fig. 2(a) shows the tendency on static and</p><p>dynamic coefficient of friction for PVDF. It is evident that the static</p><p>coefficient of friction is higher than the dynamic coefficient of</p><p>friction and they both follow a similar trend. Except the failed</p><p>specimen, the rest of the seven materials (PAI, PET, PPS, PA6, PPSU,</p><p>PVDF and UHMWPE) reached the steady state friction at ap-</p><p>proximately 2000 cycles. The average static and dynamic coeffi-</p><p>cient of friction was calculated from last 1000 cycles for all the</p><p>materials (except PC and PEI). A common tendency for most semi-</p><p>crystalline grades is the linear increase in friction during the</p><p>running-in stage. However after achieving a peak value, the fric-</p><p>tion decreases to reach the steady state. Similar behaviour was</p><p>earlier reported for polyimide [36].</p><p>The overall averaged friction coefficient for the investigated</p><p>materials are presented in Fig. 3 the same is also tabulated in</p><p>Table 2. It is evident that most materials follow a similar trend</p><p>with a higher static coefficient of friction than the dynamic</p><p>Fig. 2. (a) static and dynamic coefficient of friction for</p><p>Fig. 1. Schematic representation of the test configur</p><p>coefficient of friction. Static coefficient of friction which is often</p><p>considered for design implication is deliberated in the present</p><p>research for the ranking of materials. It is evident that the semi-</p><p>crystalline group outperformed the amorphous grades, however</p><p>PPSU from the amorphous grade has low friction characteristics</p><p>when compared with the PET and PPS of the semi-crystalline</p><p>grades. This can be attributed to the interfacial behaviour in re-</p><p>sponse to the formation of transfer film which will be discussed in</p><p>the later section. Based on friction characteristics, among these</p><p>9 different polymers, UHMWPE has the lowest friction (0,27) fol-</p><p>lowed by PVDF (0,3), PA6 (0,37) and PPSU (0,38). Recently, Ka-</p><p>lácska has reported the frictional characteristics of several material</p><p>using the traditional pin on disc tests [11]. From his report it is</p><p>clear that the majority of the materials resulted with a CoF be-</p><p>tween 0.2 and 0.3. It is also interesting that a friction values of the</p><p>present research compared with the literature shows similar va-</p><p>lues for UHMWPE and PVDF and completely contradicting for PA6</p><p>and PPSU. Though such differences can be attributed to the ma-</p><p>terial characteristics and operating condition, the output factor</p><p>which is the transfer layer characteristics may also play a critical</p><p>role. The transfer layer characteristics may change substantially as</p><p>a function of the test scale. It is noteworthy to mention that au-</p><p>dible squealing noise was observed for PEI, PET, PAI and PPS which</p><p>can be attributed to the stick-slip behaviour. The squealing noise</p><p>appear during the initial period and is continued throughout the</p><p>testing. Beside the poor friction characteristics, the squealing noise</p><p>makes the material unfit for plain bearing applications.</p><p>PVDF (b) static COF tendency of all nine polymers.</p><p>ation and the large scale flat-on-flat tribotester.</p><p>Fig. 3. (a) average static and dynamic coefficient of friction for the amorphous and semi-crystalline grades (b) specific wear rate.</p><p>Fig. 4. Worn polymer specimen (50 � 50 mm) contact surface confirming to different wear modes (a) abrasive wear (b) combined abrasion adhesion and (c) interfacial wear</p><p>(adhesion).</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–1332 1325</p><p>In regards to the wear behaviour there is a clear tendency</p><p>where material groups can be classified in to categories (see</p><p>Fig. 2b). The wear behaviour of the investigated polymer reveals</p><p>relationship between the quantitative wear response and their</p><p>corresponding wear mechanisms. For a better understanding, the</p><p>wear characteristics of the investigated polymers are classified</p><p>under three groups based on the observed wear mechanism 1.</p><p>Group 1: Abrasive wear, 2. Group 2: Combined abrasive and ad-</p><p>hesive wear and 3. Group 3: adhesive wear. Abrasion was found to</p><p>be the dominant wear mechanism in amorphous grades particu-</p><p>larly for PC and PEI which has higher wear rate (see Fig. 3b). This</p><p>can be attributed</p><p>to the inability to form transfer layer formation</p><p>which is typical for the amorphous grades [37]. However, the</p><p>formation of transfer layer may not be fully beneficial if there is an</p><p>increased adhesion between the transfer film and the parent</p><p>material. Existing literature clearly indicates that despite the</p><p>transfer layer formation in PPS, a high friction values can be ex-</p><p>pected [13]. In the present investigation, PAI, PPSU, PET and PPS</p><p>were found with significant amount of transfer layer on the</p><p>polymer surface itself. Moreover the transfer layer thickness was</p><p>inconsistent and also did not have a uniform distribution. This</p><p>addresses the fact that the transfer layer acts as a protective agent.</p><p>However due to increased adhesion between the transfer film and</p><p>the parent material there is a dynamic effect in peeling and re-</p><p>plenishment of the same. This enables new surface to be exposed</p><p>to abrasive action where debris from increased wear tends to form</p><p>new transfer layers. This group of materials which consist PAI,</p><p>PPSU, PET and PPS can be classified under Group 2. The third case</p><p>is the adhesive wear or merely gliding, where the transfer layer</p><p>persist but uniformly distributed over the polymer contact surface</p><p>which promotes merely the interfacial wear. UHMWPE, PVDF and</p><p>PA6 falls under this group, similar wear behaviour with better</p><p>ranking for UHMWPE were reported earlier [25]. An example of</p><p>the macrograph from these three groups is given in Fig. 4. All the</p><p>three mechanisms, abrasion, combined abrasion and adhesion and</p><p>adhesive wear can be clearly observed from Fig. 4.</p><p>Frictional heating plays a vital role in altering the surface</p><p>characteristics thus the wear mechanism is partly dependent on</p><p>the bulk temperature which is lower than the flash temperature.</p><p>The measured bulk temperature of the counter surface is shown in</p><p>Fig. 5(a), the temperature curve also reveals that the steady state</p><p>has been reached at 2000 cycles. The friction curve and the tem-</p><p>perature profile are in good agreement for confirming the steady</p><p>state period. With the low coefficient friction, the frictional heat-</p><p>ing is significantly reduced thus the PVDF, UHMWPE and PPSU</p><p>tends to show lower bulk temperature. It is noteworthy to men-</p><p>tion that the amorphous grades did not reach the glass transition</p><p>temperature where most semi-crystalline grades has passed the</p><p>glass transition temperature. This enables the contact surface of</p><p>the semi-crystalline grades to move around the steel asperity and</p><p>thus promoting low friction characteristics. Earlier reports con-</p><p>firms the decrease in static friction with increase in temperature</p><p>particularly within the range of (175–200 °C) [37,38]. It is inter-</p><p>esting that in Fig. 2b the friction increases to a maximum value</p><p>during the running-in and gradually reduces. Such a tendency can</p><p>be addressed from the viewpoint of flash temperature for which</p><p>the temperature is plotted against the friction curve. All the ma-</p><p>terials follow an increasing static coefficient of friction with the</p><p>Fig. 5. (a) bulk temperature as a function of time (b) friction curve vs bulk temperature.</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–13321326</p><p>increasing temperature caused by frictional heating. On reaching a</p><p>peak position, the friction decreases and this can be clearly seen in</p><p>Fig. 5(b). Such characteristic feature can be attributed to the flow</p><p>ability of polymer that allows the material to flow around the steel</p><p>asperity without significant resistance and hence the reduction in</p><p>coefficient of friction. Thermal degradation of the surface is visible</p><p>in certain cases, particularly with PAI with black coloration (see</p><p>Fig. 4b). However in our case, PAI is operating at a temperature</p><p>lesser than Tg and also the prescribed long duration operation</p><p>temperature. Thus the effects can be related to the increased flash</p><p>temperature at the interface which may have affected the transfer</p><p>layer. It is also interesting to point out that the slower the time</p><p>period to the steady state results in lower friction coefficient. The</p><p>overall tribological behaviour is given in Table 2.</p><p>3.2. Wear mechanisms</p><p>The earlier mentioned three groups of polymer wear categor-</p><p>ized primarily based on the wear performance and its corre-</p><p>sponding wear mechanisms. Hence this section attempts to de-</p><p>liberate the relation between both through mass loss and surface</p><p>characterization technique which is typically used for wear me-</p><p>chanism briefing. A reflected light bright field optical microscopy</p><p>was used to study the wear mechanisms. For ease of under-</p><p>standing, all micrographs were acquired with the same orienta-</p><p>tion, an example of sliding direction is indicated for PC in Fig. 6 (a).</p><p>From the micrographs it is evident that abrasion is prevalent in the</p><p>amorphous grades. Also irrespective to the crystallinity all the</p><p>materials will experience abrasion (2 body) during the initial</p><p>stage. This is due to the absence of transfer layer where the</p><p>Fig. 6. Group 1 materials with a bra</p><p>polymer surface are exposed to steel asperities. PC and PEI was</p><p>found with severe wear, as a consequence of abrasion. The other</p><p>amorphous grades (PAI and PPSU) also experiences abrasionwhich</p><p>can be concluded from the scratch patterns present in the contact</p><p>surface (see Fig. 7a and b). Beside the generic wear mechanisms it</p><p>is necessary to study micro-mechanism thereby the wear process</p><p>can be fully understood. It is clearly observed from the micro-</p><p>graphs that abrasion is one of the dominant mechanism and</p><p>prevalent in all material grades It is well known from literature</p><p>that abrasion may have different micro-mechanism such as micro-</p><p>cutting, micro-ploughing and micro-fatigue. Characteristics of in-</p><p>dividual debris will be explained later, however for clarity and to</p><p>correlate with micro-mechanism the debris as observed from the</p><p>polymer contact surface are discussed in detail. In PC, the abrasion</p><p>mechanism can be identified from the embedded polymer wear</p><p>debris. These debris are formed as a consequence of micro-cutting</p><p>from the steel asperities, such phenomena has also been earlier</p><p>reported for PC [39]. The flattened roll morphology as seen in</p><p>Fig. 6(a) is similar to the debris morphology observed by Aharoni</p><p>for PA6 [40]. The sliding direction and the axis of the roll and the</p><p>machine marks from the counterface parallel to each other can be</p><p>clearly seen in Fig. 6. It is typical for thermoplastics to have roll-</p><p>formation as a consequence of abrasion. Several authors [39–42]</p><p>have reported roll-formation and the present research also ap-</p><p>pears to show signs of roll-formation for PC, PPS, UHMWPE and</p><p>PET (see Fig. 7(c)(d) and Fig. 8(a)).The debris show parallel lines</p><p>which are perpendicular to the sliding direction. Such lines are</p><p>typically the grinding marks from the counterface material. This</p><p>means the debris has been transferred back to the parent polymer</p><p>surface. In the present study the roll formation were present for all</p><p>sion mechanism (a) PC (b) PEI.</p><p>Fig. 7. Combine abrasion and adhesive wear mechanism prevalent in Group 2 materials (a) PPSU, (b) PAI, (c) PPS and PET (d) PET.</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–1332 1327</p><p>three groups however, their morphology and geometry between</p><p>the groups vary significantly.</p><p>Group 1 materials did not undergo thermal softening and was</p><p>observed with rough profile on the edges with large flattened rolls</p><p>of width approximately 150–200 mm.</p><p>Group 2 materials (PET and PPS) was also found with flattened</p><p>rolls on the polymer contact surface. A common feature among the</p><p>two materials in this group is the flattened rolls appears thin</p><p>(410 mm) and smooth profile connected to the transfer layer.</p><p>Group 3 materials has flattened rolls on the polymer contact</p><p>surface which has a morphology similar to Group 2 materials</p><p>where smoothened profiles can be observed. However the size of</p><p>these ribbon like feature are significantly larger than that of the</p><p>Group 2 materials.</p><p>Having hard steel asperity engaging the soft polymer surface</p><p>may have caused by micro-cutting however,</p><p>the difference in</p><p>contact temperature in relation to the thermal characteristics may</p><p>have reflected on the morphology of rolls. The difference in wear</p><p>mechanisms between the amorphous and the semi-crystalline</p><p>grades is also evident from the presented figures. The rolls like</p><p>structure embedded in the PC contact surface appears to have ir-</p><p>regular rough profiles on the edges. But for the semi-crystalline</p><p>grades (PPS, PET and UHMWPE) rather a smooth profiles was</p><p>observed. (see Figs. 7c, d and 8a). In semi-crystalline grades the</p><p>thermal softening of the debris may yields to the formation of</p><p>smooth profiles. However in case of PC and PEI the debris were</p><p>generated from the beginning of testing where the temperature is</p><p>lower than the Tg. Thus thermal softening is less likely to occur in</p><p>case of the amorphous grades where abrasion through micro-</p><p>cutting follows throughout the wear process. Whereas in semi-</p><p>crystalline grades the flattened rolls after thermal softening yields</p><p>better and aids to the formation of transfer layer thus the initially</p><p>present micro-cutting transforms to interfacial wear. It is note-</p><p>worthy to mention that the flattened rolls on the polymer contact</p><p>surface acts as a protective agent in the wear process. The differ-</p><p>ence in surface morphology between the embedded debris and</p><p>the rest of the polymer contact surface indicates role of debris</p><p>introducing an interfacial wear. This difference in surface mor-</p><p>phology clearly points out that the rest of the polymer contact</p><p>surface is not affected by the counterface material. However the</p><p>main difference reflecting on the wear rate is from the formation</p><p>of transfer layer as this may cover a significant area of the contact</p><p>surface.</p><p>In regards to the PEI, micro-voids were observed together with</p><p>the fibrillation. From the observations on mass loss and wear</p><p>mechanism, it can be stated that the Group 1 of thermoplastics</p><p>was dominated by abrasion mechanism. The second group of</p><p>polymer grades which has experienced severe adhesion has also</p><p>shown sign of abrasion in all four material by means of roll for-</p><p>mation and also scratch pattern in the direction of sliding direction</p><p>(see Fig. 7). Scratch pattern was dominant for the amorphous</p><p>grades PPSU and PAI.</p><p>It is noteworthy to mention that all the four materials in Group</p><p>2 has chunk of transfer layer in the worn contact surface. Except</p><p>PPSU, the rest of the three polymers (PAI, PPS and PET) in this</p><p>group has experienced the maximum temperature which also</p><p>indicates the hindrance for heat transfer caused by the polymer</p><p>transfer layer on the steel counterface. The poor adhesion of</p><p>transfer layer to the counterface as observed from the PC contact</p><p>surface (see Fig. 6(a)) will facilitate new abrasion process to have</p><p>increased wear of the polymer. Having both amorphous and semi-</p><p>crystalline polymer in the Group 2 may allow us to draw conclu-</p><p>sions on the negligible influence of thermo-mechanical properties</p><p>on transfer later characteristics. On the other hand, the influence</p><p>Fig. 8. Group 3 materials indicate interfacial wear for (a) UHMWPE, (b) PVDF and (c) PA6.</p><p>Fig. 9. Worn counterface contact surface of Group 1 materials (a) PC and (b) PEI.</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–13321328</p><p>of van der Waal forces and electrostatic forces which may promote</p><p>the adhesion characteristics are still to be explored in the large</p><p>scale testing of polymers. It is also clear that the wear process is</p><p>cyclic with repeated abrasion and adhesion of the polymer contact</p><p>surface. The transfer layer facilitated by adhesion of debris to the</p><p>counterface and subsequent back transfer to the parent material</p><p>will allow the surface being partly exposed to the abrasive action</p><p>of counterface asperities.</p><p>The Group 3 material UHMWPE, PVDF and PA6 experiences the</p><p>lowest wear rate. Fig. 8 clearly shows the uniform distribution of</p><p>transfer layer without any lumpy spots. Moreover, the debris being</p><p>an integral part of the transfer layer is also seen (see arrow mark</p><p>on Fig. 8a). The wave like pattern in Fig. 8a and b clearly points out</p><p>the adhesive nature of the transfer film which is sufficient to form</p><p>pattern on the contact surface rather than removing the parent</p><p>material itself. In all three cases abrasion was prevalent during the</p><p>initial stage of wear, this can be observed from the roll formation</p><p>characteristics and the scratch pattern present in the polymer.</p><p>However the capability of forming transfer layer tends to change</p><p>the wear mechanism which will be experienced in the steady</p><p>stage. Materials from the Group 1 undergoes severe abrasion due</p><p>to single and repeated cycle deformation and hence the increased</p><p>rate of wear. In Group 2, the materials undergoes combined</p><p>abrasion and adhesion process. The wear particle generated by</p><p>abrasion and the adhesion mechanism facilitates formation of</p><p>transfer layer. Though the abrasion may induce increased wear, the</p><p>transfer film acts as a protective agent in decreasing the wear rate.</p><p>Group 3 materials merely undergoes adhesion facilitated by</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–1332 1329</p><p>uniform transfer layer which can be observed in the polymer (see</p><p>Fig. 8). Considering the used operative condition and based on the</p><p>combined results from friction and wear rate, the Group 3 material</p><p>are well suited for plain bearing application.</p><p>3.3. Transfer layer formation and debris characteristics</p><p>One of the critical parameters dictating the tribological char-</p><p>acteristics of polymer-metal contact is the transfer layer. Briscoe</p><p>et al. classifies the polymer transfer film based on their mor-</p><p>phology and distribution which in-fact plays an important role</p><p>[43]. In the present research, observations from post-mortem</p><p>analysis of contact surface for PC and PEI (Group 1 material) shows</p><p>embedded debris. However, the major portion of the generated</p><p>wear particles was evacuated from the contact zone. This is typical</p><p>among the amorphous material where the generated wear particle</p><p>has a poor chance of forming a thin layer transfer layer [37]. Since</p><p>the debris does not contribute to the transfer layer formation, the</p><p>same is rolled into a significantly thick bundle which acts as a third</p><p>body in inducing abrasion. Abrasion scratches were also found in</p><p>the steel counterface which may indicate the change in debris</p><p>characteristics as a consequence of wear (see Fig. 9).</p><p>In regards to the Group 2 materials, transfer layer as observed</p><p>from the counterface surface are predominantly present for PPSU</p><p>and PET. The lumpy deposit in these two cases were referred as</p><p>secondary transfer layer in the literature [45]. Primary layer which</p><p>is the deposit between the steel asperity can also be observed in</p><p>the microscopy as a discoloration of the contact surface. In regards</p><p>to the secondary transfer layer, a step by step process of deposition</p><p>with several layers was evidenced (see Fig. 10a). The total thick-</p><p>ness of the transfer layer gradually increases as a function of time</p><p>and it may peel out when the adhesion between the transfer layer</p><p>and the counterface material is weaker than the adhesion between</p><p>the transfer layer and the parent polymer surface. Similar char-</p><p>acteristics was reported earlier in the literature for polymer-</p><p>polymer contact [44]. This means the adhesion between the dif-</p><p>ferent layer also contribute to the formation of the transfer layer</p><p>which is dependent on the molecular and supramolecular struc-</p><p>tures. Having a lumpy formation and inhomogeneous distribution</p><p>of transfer layer does not allow for a quantitative evaluation of</p><p>transfer layer characterization. This is evidenced from the profi-</p><p>lometry measurement where the thickness of the transfer layer</p><p>varies as the function of measured location (see Fig. 11b). In re-</p><p>gards to the Group 3 material the counterface showed signs of</p><p>discoloration from the original surface and this can be attributed</p><p>to the debris accumulation between the roughness valleys. The</p><p>wear track appears to be uniform throughout the surface and very</p><p>limited mild abrasion scars can be found</p><p>on the contact surface.</p><p>Fig. 10. Counterface wear track and transfer film form</p><p>Very often the transfer layer characteristics in small scale</p><p>testing are discussed from the view point of the counterface ma-</p><p>terial [18–20]. However, few reports clearly mentions the dynamic</p><p>characteristics of transfer layer deposition and removal which is an</p><p>ongoing process [44,46]. Also having such dynamicity, the in-</p><p>vestigation from port-mortem may introduce uncertainty if in-</p><p>formation is merely accounted at the end of testing and also from</p><p>counterface. Additionally, the inhomogeneous distribution may</p><p>not allow to characterize and reason its effect on tribological</p><p>performance. However, with a more qualitative approach as per</p><p>the classification of Brisco et al. the influence can be broadly dif-</p><p>ferentiated between the three groups for the given material. Hence</p><p>in the present research both the counterface and the polymer</p><p>surface were accounted for the transfer layer characteristics. To</p><p>have a complete overview, the presence of debris and its char-</p><p>acteristics were also studied in the present investigation. Litera-</p><p>ture refers to a lumpy transfer film as a normal behaviour [14]</p><p>which is also observed in Group 2 material. Moreover the thick</p><p>layer are meant for affecting the tribological properties and hence</p><p>the studies are often focused on the same. In the conventional</p><p>standard testing, the reported thickness of the transfer layer ran-</p><p>ges between 0,1 and 1.0 µm [14,44,47,48]. However, in the present</p><p>case, lumpy layer up to 25 mm thickness were observed on the</p><p>counterface material (PET and PPS). Literature indicates that me-</p><p>chanical interlocking of the transfer film by counterface asperity is</p><p>the dominant mechanism to hold the transfer layer [19]. Never-</p><p>theless, the adhesive characteristics of the transfer film based on</p><p>van der Waals force or electrostatic force are to be considered in</p><p>the study as well. From the present investigation it is clear that the</p><p>adhesion characteristics plays a dominant role. This is due to the</p><p>fact that even with layered deposition, the peeling of the transfer</p><p>layer from the counterface is often at the junction of transfer film</p><p>and the steel counterface. This can be clearly seen from the contact</p><p>surface of PET where the grinding marks from the steel surface are</p><p>clearly seen. The same also fits well for the PPSU debris where the</p><p>transfer layer comprising multiple layer fails at the interface be-</p><p>tween the counterface asperity and the transfer film (see Fig. 12d).</p><p>Mergler et al. reports that the thickness of the transfer layer</p><p>does not play an important role in altering the frictional char-</p><p>acteristics [19]. This is also the case with the present research</p><p>where there is no correlation between transfer layer character-</p><p>istics and friction, hence the classification based on the wear</p><p>mechanism is not in line with the friction characteristics. This</p><p>contradicts to the statement of Sviridyonok where friction char-</p><p>acteristics is affected by the monomolecular layer [44]. A common</p><p>characteristics feature for the Group 2 material in relation to the</p><p>transfer layer is that all materials in this group experiences a high</p><p>frequency variation in the friction characteristics. This can be</p><p>ation from Group 2 materials (a) PPSU, (b) PET.</p><p>Fig. 11. Secondary layer thickness measured from counterface material (a) schematic of the measurement and (b) profilometry measurements.</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–13321330</p><p>attributed to the dynamicity involved in the formation and peeling</p><p>of transfer layer due to poor adhesion (see Fig. 2b). Group 2 ma-</p><p>terials shows increased wear rate when compared to the Group</p><p>3 materials, and this can be attributed to the lumpy layer de-</p><p>position. Similar results were earlier reported where nylon with</p><p>patchy layer having increased wear rate when compared with the</p><p>uniform deposition [14].</p><p>The advantages of uniform transfer film over discontinuous</p><p>film has been already reported by Samyn et al. [49]. Comparing the</p><p>observation from the present investigation and the literature data</p><p>on small scale result may be trivial from the view point of transfer</p><p>layer characteristics. This is due to the difference in the thickness</p><p>Fig. 12. Debris characteristics of amorphous g</p><p>of the transfer layer between the two scales. Moreover, large rolled</p><p>out debris (PC) introduces three body abrasion. Whereas in tra-</p><p>ditional testing the debris do not grow far enough due to the</p><p>limitation in the area of contact. In the small scale testing, either</p><p>the debris stays small or the debris may evacuate the contact zone.</p><p>Considering the importance of large scale testing, recent tests on</p><p>large scale were performed for pure matrix, however their ap-</p><p>proach lacks information on the wear mechanism [50]. The re-</p><p>ported friction values are several factors smaller than the small</p><p>scale testing results. Previous research on large scale testing</p><p>showed a decreasing trend with increasing contact pressure</p><p>[49,50]. Comparing the literature, the large scale tests have</p><p>rades (a) PEI, (b) PC, (c) PAI and (d) PPSU.</p><p>Fig. 13. DSC curve before after wear for the investigated specimen and the debris.</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–1332 1331</p><p>provided deviations from the existing reports. Though variations</p><p>already persists between the small and large scale tests it can also</p><p>be attributed to operating variables. In a comparative study by</p><p>Mens et al. the small scale test reports shows the poor formation</p><p>of transfer film [13]. Depending upon the adhesion strength be-</p><p>tween the transfer layer and the counterface or the parent mate-</p><p>rial the friction force changes. Since debris characteristics changes,</p><p>the wear mechanism tend to change which may result in a dif-</p><p>ferent friction behaviour.</p><p>In regards to the debris, all the amorphous polymers generated</p><p>significant amount of debris. This also confirms the dynamicity in</p><p>the transfer of the polymer film between the counterface surface</p><p>and the worn polymer surface. PC and PEI had different debris</p><p>morphology when compared with the PAI and PPSU. In regards to</p><p>PC and PEI the debris appears to be chunks of rolled sheet without</p><p>any orientation (se Fig. 12a and b). However, the PAI and PPSU</p><p>sheets of transfer layer can be seen in Fig. 12(c) and (d). Debris</p><p>from all four material had the roughness pattern from the coun-</p><p>terface (see Fig. 12a and d) and hence the lack of primary transfer</p><p>layer which may be the primary cause for repeated abrasion. Black</p><p>coloration of the PAI debris indicate the thermal degradation of</p><p>polymer debris. In PAI the major portion of the debris consist of</p><p>powder like debris instead of chunks. Unlike the amorphous</p><p>grades, the semi-crystalline polymer were found with no notice-</p><p>able debris. Hence the generated debris were contained within the</p><p>interface. This also points out the increased adhesion character-</p><p>istics between the transfer layer and the contacting bodies for</p><p>Group 2 and 3 polymer grades. On comparing with the literature,</p><p>it is interesting to point out the size difference of the debris and</p><p>the rolls formed in the wear process. At small scale due to the</p><p>limitation in the size of the contact zone the debris evacuate even</p><p>before they are grown substantially. However in real practise</p><p>having large contact zone the debris roll sufficiently to large size</p><p>which may act as an abrasive in developing three body abrasive</p><p>wear. This can be clearly seen in PC and PEI.</p><p>The combined inference from the polymer worn surface, steel</p><p>counterface and the debris characteristics reveal the same where</p><p>the Group 1 material undergoes abrasion and the Group 2 and</p><p>3 material experiences sever and mild adhesive wear. The debris</p><p>were further investigated in the DSC which will be discussed in</p><p>the later section. It is noteworthy to mention that the maximum</p><p>bulk temperature experienced by the amorphous grade is always</p><p>lesser than the glass transition temperature. This in-fact tends to</p><p>deliberate the brittle nature of the material and hence playing a</p><p>role in particle</p><p>generation process and as a consequence plate like</p><p>debris were formed for the amorphous grades. Similar phenom-</p><p>enon was also reported elsewhere [48]. Thermal degradation was</p><p>not experienced from the debris of PEI and PC. However small</p><p>signs of thermal degradation on the surface of the PAI and PPSU</p><p>debris were noticed as black coloration. This can be attributed to</p><p>the impedance of the heat dissipation due to the formation of</p><p>lumpy transfer layer. Having seen the different types of transfer</p><p>layer characteristics, one has to remember that there is no set of</p><p>boundary condition to classify these layers in to a particular group</p><p>and hence the results may not be universal. The results also pro-</p><p>poses the requirement of methodology and strictly defined</p><p>boundary condition for transfer layer characteristics.</p><p>3.4. Thermo-mechanical analysis</p><p>Literature clearly reports the pivotal role of temperature on dic-</p><p>tating the tribological properties of polymers in sliding [37,38]. This</p><p>can be associated with the physio-mechanical changes as a con-</p><p>sequence of frictional heating. The DMA analysis before and after did</p><p>not show with changes in the Tg, however the storage modulus of the</p><p>amorphous grades appear to be increased after the wear. The main</p><p>objective of the DSC is not to point out the relationship between</p><p>mechanical behaviour and its wear response to the changing tem-</p><p>perature. Such a study may require a different approach to explicate</p><p>the thermal advantage between materials. In the current study the</p><p>authors attempt to merely indicate the difference in thermal char-</p><p>acteristics before and after wear for pointing out the negligible dif-</p><p>ference in thermal properties of the bulk material. Thereby the future</p><p>research can be focused on the merely studying the thermal char-</p><p>acteristic of the surface. In the DSC, a heating cycle (1) up to 250 °C</p><p>followed by a cooling cycle to �70 °C (2) and subsequent heating to</p><p>250 °C was made in the DSC measurement. The thermographs are</p><p>shifted over a fixed value along the y-axis for presentation purpose</p><p>(see Fig. 13). Comparing the heat flow signals before and after wear</p><p>testing minor changes were observed in the cold crystallization peaks</p><p>(exothermic) for the semi-crystalline material. Moreover widening of</p><p>the melting region is observed for PET and shifting in the melting</p><p>peaks were observed for PET, PA6, PVDF and UHMWPE. Widening of</p><p>the melting peaks can be attributed to the distribution of oriented</p><p>surface as a consequence of sliding [36]. Amongst the amorphous</p><p>material, no significant change was observed between the measure-</p><p>ments made before and after wear. The glass transition temperature of</p><p>the debris remained the same as that of the bulk material. It is also</p><p>noteworthy to point out that the recrystallization and the melting</p><p>peak position remains similar with minor shifts for the semi-crystal-</p><p>line material. Also having the melting peak in both material it can be</p><p>concluded that the crystallinity of the material remains unchanged.</p><p>4. Conclusions</p><p>In the heavy duty testing of nine different polymers from both</p><p>the amorphous and semi-crystalline grades the following conclu-</p><p>sions can be drawn based on the friction characteristics, specific</p><p>wear rate, wear mechanism, transfer layer characteristics and</p><p>debris morphology:</p><p>J. Sukumaran et al. / Wear 376-377 (2017) 1321–13321332</p><p>� The polymer wear and friction characteristics in high load</p><p>condition primarily depend upon the formation of transfer layer</p><p>which is partly controlled by the material types.</p><p>� Among the tested materials three groups can be formed based</p><p>on the wear mechanism which results in corresponding wear</p><p>rate. Group 1 consist of amorphous grades with PC and PEI</p><p>which are prone to severe abrasion and hence resulting with</p><p>highest wear rate. Group 2 material (PAI, PPS, PET and PPSU)</p><p>experiences a combined adhesion and abrasion process. Severe</p><p>adhesive wear is the major reason for the increased wear rate of</p><p>Group 2 material. Both amorphous and semi-crystalline grades</p><p>were present in the group 2. However the wear rates are rela-</p><p>tively lesser than the Group 1 material which is attributed to the</p><p>transfer layer formation. Finally the Group 3 materials which</p><p>are UHMWPE, PA6 and PVDF experiences mild adhesive wear</p><p>and more suitable for the heavy duty wear application.</p><p>� Both Group 1 and 2 materials experiences stick-slip behaviour and</p><p>thus cannot be considered for plain bearing applications.</p><p>� In regards to the friction characteristics, UHMWPE and PVDF</p><p>have superior performance which can be considered for heavy</p><p>duty applications.</p><p>� In heavy duty testing, the severe wear owing to three body abrasion</p><p>is due to the inability of the material (PC and PEI) to form transfer.</p><p>� In heavy duty testing it is not completely appropriate to take in</p><p>account of the quantitative characteristics of the transfer layer</p><p>due to their inhomogeneous distribution. Moreover the transfer</p><p>layer characteristics needs to be investigated in both polymer</p><p>and steel counterface material.</p><p>Certainly more research is needed to understand the process</p><p>involved in the transfer layer formation from the view point of</p><p>molecular interaction and this proposes a multi-scale tribological</p><p>investigation.</p><p>Acknowledgements</p><p>This investigation was supported and funded by Fonds We-</p><p>tenschappelijk Onderzoek – Vlaanderen (FWO) 12S1915N. La-</p><p>boratory Soete acknowledges the material support received</p><p>through Hendrik Vandenbruaene of Quadrant EPP Belgium N. V.</p><p>The authors are grateful to Angelot Noel Fodem Kenfack and</p><p>Wouter Ost for their support on experimental activities.</p><p>References</p><p>[1] A.B. Strong, History ofComposite Materials, ed: July, 2002.</p><p>[2] R. Gawarkiewicz, M. Wasilczuk, Wear measurements of self-lubricating bear-</p><p>ing small oscillatory movement, Wear 263 (2007) 458–462.</p><p>[3] P. Samyn, L. Van Schepdael, J.S. Leendertz, A. Gerber, W. Van Paepegem, P. De</p><p>Baets, J. Degrieck, Large-scale friction and wear tests on a hybrid UHMWPE-</p><p>pad/primer coating combination used as bearing element in an extremely</p><p>high-loaded ball-joint, Tribol. Int. 39 (2006) 796–811.</p><p>[4] P. De Baets, P. De Baets, Comparison of the wear behaviour of six bearing materials</p><p>for a heavily loaded sliding system in seawater, Wear 180 (1995) 61–72.</p><p>[5] C. Ren, J. Chen, H. Pan, X. Huang, H. Zhu, Wear mechanism of PA6 roller used in</p><p>cold rolling process, Eng. Fail. Anal. 28 (2013) 311–317.</p><p>[6] B. Carsten, Plastic bearings prove their mettle, Mach. Des. (1996) 61–63, 68-20.</p><p>[7] PTFE bearing pad for olympic stadium, Tribology (1968) 199.</p><p>[8] R.L. Fusaro, Self-lubricating polymer composites and polymer transfer film</p><p>lubrication for space applications, Tribol. Int. 23 (2) (1990) 105–122.</p><p>[9] J.K. Lancaster, Accelerated wear testing as an aid to failure diagnosis and</p><p>materials selection, Tribol. Int. (1982) 323–329.</p><p>[10] H. Rajashekaraiah, S. Bheemappa, S.H. Yang, S. Mohan, Abrasive wear beha-</p><p>viour of thermoplastic copolyester elastomer composites: a statistical ap-</p><p>proach, Int. J. Precis. Eng. Manuf. 17 (2016) 755–763.</p><p>[11] G. Kalácska, An engineering approach to dry friction behaviour of numerous</p><p>engineering plastics with respect to the mechanical properties, Express Polym.</p><p>Lett. 7 (2) (2013) 199–210.</p><p>[12] X. Zhang, X. Pei, Q. Wang, Friction and wear behaviour of basalt-fabric-</p><p>reinforced/solid-lubricant-filled phenolic composites, J. Appl. Polym. Sci. 117</p><p>(2010) 3428–3433.</p><p>[13] J.W.M. Mens, A.W.J. de. Gee, Friction and wear behaviour of 18 polymers in contact</p><p>with steel in environments of air and water, Wear 149 (1991) 255–268.</p><p>[14] S. Bahadur, The development of transfer layers and their role in polymer tri-</p><p>bology, Wear 245 (2000) 92–99.</p><p>[15] J. Ye, A.C. Moore, D.L. Burris, Transfer film tenacity: a case study using ultra-</p><p>low-wear alumina–PTFE, Tribol. Lett. (2015) 50–59.</p><p>[16] V.K. Jain, S. Bahadur, Material transfer in polymer-polymer sliding, Wear 46</p><p>(1978) 177–188.</p><p>[17] S.H. Rhee, K.C. Ludema, Mechanisms of formation of polymeric transfer films,</p><p>Wear 46 (1978) 231–240.</p><p>[18] A. Kapoor,</p><p>S. Bahadur, Transfer film bonding and wear studies on CuS-bylon</p><p>composite sliding against steel, Tribol. Int. 27 (1994) 323–329.</p><p>[19] Y.J. Mergler, R.P. Schaake, A.J. Huis in’t Veld, Material transfer of POM in sliding</p><p>contact, Wear 256 (2004) 294–301.</p><p>[20] J. Vande Voort, S. Bahadur, The growth and bonding of transfer film and the</p><p>role of Cu Sand PTFE in the tribological behaviour of PEEK, Wear 181–183</p><p>(1995) 212–221.</p><p>[21] S. Bahadur, D. Gong, J.W., Anderegg, he investigatin of the action of fillers by</p><p>XPS studies of the transfer filsm of PEEK and its composites containing CuS</p><p>and CuF2, Wear 160 (1) (1993) 131–138.</p><p>[22] P. Samyn, P. De Baets, Friction and wear of acetal: a matter of scale, Wear 259</p><p>(2005) 697–702.</p><p>[23] J.K. Lancaster, Dry bearings: a survey of materials and factors affecting their</p><p>performance, Tribology (1973) 219–251.</p><p>[24] J.K. Lancaster, Accelerated wear testing of PTFE composite bearing materials,</p><p>Tribol. Int. (1979) 65–75.</p><p>[25] J.C. Anderson, Wear of commercially available plastic materials, Tribol. Int.</p><p>(1982) 255–263.</p><p>[26] P. Samyn, W. Van, Paepegem, J.S. Leendertz, A. Gerber, L. Van Schepdael, J. Degrieck,</p><p>P. De Baets, Large-scale evaluation of constrained bearing elements made of</p><p>thermosetting polyester resin and polyester fabric, J. Tribol. 128 (2006) 681–696.</p><p>[27] H. Jang, J. Ha, J. Yoo, J. Pyo, K. Choi, C. Lee, T. Ryu, S. Lee, W. Kim, Synthesis and</p><p>characterization of polycarbonate copolymers containing benzoyl groups on</p><p>the side chain for scratch resistance, Int. J. Polym. Sci. (2016) 1–6.</p><p>[28] 〈http://www.quadrantplastics.com/eu-en/products/machinable-plastics.html〉,</p><p>August 2016.</p><p>[29] F. Stempfle, P. Ortmann, S. Mecking, Long-chain aliphatic polymers to bridge</p><p>the gap between semicrystalline polyolefins and traditional polycondensates,</p><p>Chem. Rev. 116 (7) (2016) 4597–4641.</p><p>[30] P. Samyn, G. Schoukens, Experimental extrapolation model for friction and wear of</p><p>polymers on different testing scales, Int. J. Mech. Sci. 50 (2008) 1390–1403.</p><p>[31] S. Apichartpattanasiri, J.N. Hay, S.N. Kukureka, A study of the tribological be-</p><p>haviour of polyamide 66 with varying injection-moulding parameters, Wear</p><p>251 (1–12) (2001) 1557–1566.</p><p>[32] P.H. Shipwayand, N.K. Ngao, Microscale abrasive wear of polymeric materials,</p><p>Wear 5 (1–6) (2003) 742–750.</p><p>[33] V. Rodriguez, J. Sukumaran, A. Schlarb, P. DeBaets, Influence of solid lubricants on</p><p>tribological properties of polyetheretherketone(PEEK), Tribol. Int. 103 (2016) 45–57.</p><p>[34] V.Rodriguez, J.Sukumaran, Y.Perez, P.De Baets, A.Igartua, B.Fernandez, A.K.</p><p>Schlarb, Tribological behaviour of the low and high viscosity PEEK at various</p><p>testing scales. Eibar, Spain, Proceedings of the Lubrication, maintenance and</p><p>tribotechnology, 145–149, 2012.</p><p>[35] R.B. Lewis, Predicting the Wear of Sliding Plastic Surfaces, Mech. Eng. 86</p><p>(1964) 33.</p><p>[36] P. Samyn, Tribophysical Interpretation of Scaling Effects in Friction and Wear</p><p>for Polymers (PhD Dissertation), Ghent University, Belgium, 2007.</p><p>[37] V.A. Bely, A.I. Sviridenok, M.I. Petrokovets, Friction andWear in Polymer-Based</p><p>Materials, Pergamon press 1982, p. 111.</p><p>[38] G.M. Bartenev, A.I. Elkin, The friction of polymers at the initial stage of sliding</p><p>in various temperature ranges, Wear 2 (1968) 393–403.</p><p>[39] J.H. Byett, C. Allen, Dry sliding wear behaviour of polyamide 66 and poly-</p><p>carbonate composites, Tribol. Int. 25 (1992) 237–246.</p><p>[40] S.M. Aharoni, Wear of polymers by roll-formation, Wear 25 (1973) 309–327.</p><p>[41] Y.K. Chen, S.N. Kukureka, C.J. Hooke, M. Rao, Surface topography and wear me-</p><p>chanisms in polyamide 66 and its composites, J. Mater. Sci. 35 (2000) 1269–1281.</p><p>[42] L. Chang, Z. Zhang, H. Zhang, A.K. Schlarb, On the sliding wear of nanoparticle</p><p>filled polyamide 66 composites, Compos. Sci. Technol. 66 (2006) 3188–3198.</p><p>[43] B.J. Briscoe, S.K. Sinha, Wear of polymers, J. Eng. Tribol. 216 (2002) 401–413.</p><p>[44] A.I. Sviridyonok, V.A. Bely, V.A. Smurugov, V.G. Savkin, A study of transfer in</p><p>firctional interaction of polymers, Wear 25 (3) (1973) 301–308.</p><p>[45] J. Sukumaran, Vision Assisted Tribography of Rolling-Sliding Contact of Poly-</p><p>mer-Steel Pairs (PhD dissertation), Ghent University, Belgium, 2014.</p><p>[46] J. Sukumaran, V. Rodriguez, T.D. Nguyen, J. De Pauw, A. Matyas, P. De Baets,</p><p>Tansfer layer dynamicity in roll-slip of polymer metal pairs, Int. J. Sustain.</p><p>Constr. Des. 4 (2) (2013) 1–8.</p><p>[47] K. Tanaka, Transfer of semi-crystalline polymers sliding against a smooth steel</p><p>surface, Wear 75 (1982) 183–199.</p><p>[48] Z. Zalisz, A running-in model for the reciprocating sliding of Nylon 6.6 against</p><p>stainless streel, Wear 121 (1988) 71–93.</p><p>[49] P. Samyn, P. De Baets, G. Schoukens, A.P. Van Peteghem, Large-scale tests on</p><p>friction and wear of engineering polymers for material selection in highly</p><p>loaded sliding systems, Mater. Des. 27 (2006) 535–555.</p><p>[50] V. Quaglini, P. Dubini, Friction of polymers sliding on smooth surfaces, Adv.</p><p>Tribol. (2011) 8.</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref1</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref1</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref1</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref2</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref2</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref2</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref2</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref2</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref3</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref3</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref3</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref4</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref4</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref4</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref5</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref5</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref6</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref7</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref7</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref7</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref8</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref8</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref8</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref9</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref9</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref9</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref9</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref10</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref10</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref10</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref10</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref11</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref11</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref11</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref11</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref12</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref12</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref12</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref13</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref13</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref13</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref14</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref14</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref14</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref15</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref15</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref15</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref16</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref16</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref16</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref17</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref17</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref17</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref18</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref18</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref18</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref19</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref19</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref19</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref19</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref20</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref20</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref20</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref20</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref21</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref21</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref21</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref22</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref22</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref22</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref23</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref23</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref23</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref24</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref24</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref24</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref25</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref25</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref25</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref25</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref26</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref26</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref26</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref26</p><p>http://www.quadrantplastics.com/eu-en/products/machinable-plastics.html</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref27</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref27</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref27</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref27</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref28</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref28</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref28</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref29</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref29</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref29</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref29</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref30</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref30</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref30</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref31</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref31</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref31</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref32</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref32</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref33</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref33</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref34</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref34</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref35</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref35</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref35</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref36</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref36</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref36</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref37</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref37</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref38</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref38</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref38</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref39</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref39</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref39</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref40</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref40</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref41</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref41</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref41</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref42</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref42</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref43</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref43</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref43</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref43</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref44</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref44</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref44</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref45</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref45</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref45</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref46</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref46</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref46</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref46</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref47</p><p>http://refhub.elsevier.com/S0043-1648(17)30075-3/sbref47</p><p>Revisiting polymer tribology for heavy duty application</p><p>Introduction</p><p>Materials and methods</p><p>Tribological characterization</p><p>Results and discussion</p><p>Friction and wear behaviour at large scale testing</p><p>Wear mechanisms</p><p>Transfer layer formation and debris characteristics</p><p>Thermo-mechanical analysis</p><p>Conclusions</p><p>Acknowledgements</p><p>References</p>