Tribology International 44 (2011) 1032–1046 Contents lists available at ScienceDirect Tribology International journal homepage: www.elsevier.com/locate/triboint Scratch resistance of high performance polymers K. Friedrich a,b,n, H.J. Sue c, P. Liu c, A.A. Almajid b a b c Institute for Composite Materials (IVW GmbH), Technical University of Kaiserslautern, 67663 Kaiserslautern, Germany CEREM, College of Engineering, King Saud University, Riyadh, Saudi Arabia Polymer Technology Center, Texas A&M University, College Station, Texas, USA a r t i c l e i n f o a b s t r a c t Article history: Received 11 January 2011 Received in revised form 4 April 2011 Accepted 11 April 2011 Available online 20 April 2011 Scratch tests were carried out on various high performance polymers, including (1) polybenzimidazole (PBI), (2) polyparaphenylene (PPP), (3) polyetheretherketone (PEEK), and (4) polyimide (PI). The scratch damage features were characterized using laser confocal and scanning electron microscope. Scratch resistance at room temperature decreased in the same order as the materials are listed above. It was attempted to correlate the scratch depth with basic mechanical properties, such as Young’s modulus, tensile strength, and scratch hardness. Also, the scratch coefﬁcient of friction was considered as a possible measure to differentiate between the various materials tested. & 2011 Elsevier Ltd. All rights reserved. Keywords: Scratch High performance polymers Elastic modulus 1. Introduction The scratch process can be deﬁned as a mechanical deformation process where a controlled force or displacement is exerted on a hard tip to indent onto another material substrate and move across its surface at a prescribed speed . In terms of wear as a sub-discipline of tribology, scratch can be considered as a single path, single asperity wears process with high abrasion characteristics (Fig. 1). The typical mechanisms of abrasive action can be categorized as microplowing, microcutting, and microcracking . The scratch performance of polymers has caught signiﬁcant attention in the past few years because of their greatly expanded usage in the electronic, optical, household, and automotive applications, where long term esthetics is important. Unlike ceramics and metals, polymers are particularly susceptible to visible surface deformation and damage, even under low contact loads. Various scratch induced damage features, such as mars, ﬁsh-scales, parabolic cracks, and material removals have been observed from a wide variety of polymeric materials [3–6]. Majority of these observations have been made, however, only for commodity or engineering polymers, such as polyethylene, polypropylene, polystyrene, polycarbonate, epoxy, etc.; whereas, the investigation of the scratch behavior of existing or newly developed high performance polymers is still lacking. It was therefore the objective of this study to carry out a series of n Corresponding author at: Institute for Composite Materials (IVW GmbH), Technical University of Kaiserslautern, 67663 Kaiserslautern, W-67653 Kaiserlautern, Germany. E-mail address: [email protected] (K. Friedrich). 0301-679X/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.triboint.2011.04.008 scratch tests on various high performance polymers, of which some of them have already been used for other tribological applications, such as bearings or sliding elements. 2. Experimental details 2.1. Materials The polymers chosen for the present scratch tests are listed in Table 1. All materials, except PEEK, were received from the manufacturers as compression molded or extruded plates. PEEK was prepared in-house in the form of 3 mm thick dog bone samples using an Arburg injection molding machine. All the material properties listed in Table 1 were taken from data sheets published by the manufacturers [7–10]. Regarding the special characteristics of the polymers chosen, the following short statements can be made: (a) Polybenzimidazole (PBI) PBI is an imidized thermoplastic, being the highest performing engineering plastic currently available. It offers the highest heat resistance (heat deﬂection temperature of 427 1C) and mechanical property retention over 205 1C of any unﬁlled plastic. Also, its wear resistance and load carrying capabilities at extreme temperatures is better than that of many other reinforced or unreinforced engineering plastics. PBI is also an excellent thermal insulator. Other plastics in melt do not stick to PBI. These characteristics make it ideal for contact seals and insulator bushings in plastic production and molding equipment . K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 1033 Tribology Lubrication Wear Friction (Surface Damage with /without Remarkable Material Removal) Abrasion Adhesion (SharpTips Scratching) (Sliding against Smooth Surface) Surface Fatigue Tribo-Corrosion (Rolling Counterpart) (Chemical Interactions) Single Pass Multiple Pass (Repeatedly against the Same Counterpart) Many Asperities Single Asperity (e.g. Sand Paper) (e.g. Tip of Nail) Reduction of Abrasiveness with Time Many Scratches Simultaneously Single Scratch (Material Sees always a Fresh Counterpart) Micro-Plowing Micro-Cutting Micro - Cracking (e.g. Mars) (e.g. Chips) (e.g. Fish Scale Pattern) (w/wo Material Removal) Fig. 1. Schematic relationship between the terms tribology, wear, abrasion, and scratch. Table 1 List of materials investigated, including manufacturer and important properties (E¼ elastic modulus; sB ¼strength; anI ¼notched Izod impact energy; eB ¼strain at break; and r ¼density). anl (J/m) r (g/cm3) 3 28 1.301 207 5 43 1.210 3.55 2.95 100 80 34 7.6 54 46 1.301 1.401 2.76 61 8.5 53 1.427 Short name Manufacturer company code E (GPa) sB (MPa) PBI PBI performance prod. Celazole U60 SD Solvay adv. polymers PrimoSpire PR 120 Victrex PEEK 450 G DuPont (Vespel) ST 2002 (2% Gr) DuPont (Vespel) ST 2010 (10% Gr) 5.86 159 8.30 PPP PEEK PI 2 PI 10 (b) Polyparaphenylene (PPP) PPP can be considered as a self-reinforced polymer, which at room temperature is one of the stiffest and strongest unreinforced plastics to-date. Even without ﬁber reinforcement, PPP delivers tensile properties that are comparable to those of many reinforced plastics. Additional beneﬁts are their lighter weight and no loss of ductility. These properties combine with high compressive strength—one of the highest among plastics—makes PPP an excellent candidate for weight-sensitive applications that have historically relied on composites and specialty metals for superior mechanical performance . (c) Polyetheretherketone (PEEK) PEEK is a semicrystalline thermoplastic with excellent mechanical and chemical resistance properties that are retained at high temperatures. It exhibits Young’s modulus eB (%) of 3.6 GPa and tensile strength of 90–100 MPa. PEEK has a glass transition temperature at around 143 1C and melts at around 343 1C. It is highly resistant to thermal degradation as well as attacked by both organic and aqueous environments. Because of its robustness, PEEK is used to fabricate items used in demanding applications, including bearings, piston parts, pumps, compressor plate valves, and cable insulation. It is one of the few plastics suitable for ultra-high vacuum applications. PEEK is considered an advanced biomaterial used in medical implants. It is extensively used in the aerospace, automotive, electronic, and chemical process industries . (d) Polyimide (PI) The chemical structure of a typical polyimide is called aromatic, as all of the carbons in the polymer chain are part of either the imide or the benzene (aromatic) rings. This 1034 K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 arrangement leads to a higher glass transition temperature (if one can be detected) and greater thermal and oxidative stability than the majority of organic polymers. Thermosetting polyimides are also known for their good chemical resistance and excellent mechanical properties. They exhibit very low creep and high tensile strength. The polyimide materials are lightweight, ﬂexible, and resistant to heat and chemicals. Therefore, they are used in the industry for ﬂexible cables, as insulating ﬁlms, as bearings, and for medical tubing . 2.2. Hardness testing The Martens hardness HM (see Ref. ) of different materials was determined by the use of a Dynamic Ultra Micro Hardness Tester, DUH-202, Shimadzu Corp., Japan. A Vickers diamond indenter was pressed onto the ﬂat surface of each sample, until a complete load-indentation depth-curve was achieved. From this, the Martens hardness HM could be calculated as HM ¼ F=ð26:43h2max Þ, ð1Þ where F is the test load in (N), and hmax is the indentation depth under applied test load in (mm). HM is therefore a simpliﬁed hardness value (ignoring the viscoelastic recovery of the polymer) compared to the classical indentation hardness, in which the maximum load is related to an area function. The latter is calculated from the product of a constant c (dependent on the indenter geometry) and the depth hc as the intercept of the tangent line drawn from the upper part of the unloading curve (describing the elastic deformation effect) and the displacement axis . Scratch Direction and Speed Normal Load Polymer Sample Stainless Steel Indenter Indenter: Tip Diameter 1 mm Roughness 60–100 nm Scratch Length 15 mm Scratch Speed 10 mm/s Load Increase from 5 to 90 N Test Load [mN] Fig. 2. Photo of the scratch testing device (a), and schematic illustration of scratch test and corresponding testing parameters (b). 1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 PBI PPP 0 1 2 3 4 5 6 PEEK 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Indentation Depth [μm] Fig. 3. Test load vs. indentation depth during Martens hardness test of PBI, PPP, and PEEK. K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 2.3. Scratching machine and test conditions In close relationship to the ASTM and ISO standard for scratch testing of polymers [13,14], a custom-built scratch machine (Surface Machine Systems, LLC, Texas, USA) was utilized to perform the scratch tests at ambient condition (Fig. 2). The machine is capable of recording tangential and normal forces as well as scratch distance and instantaneous depth experienced by the stylus. A stainless steel spherical tip with a diameter of 1 mm and a surface roughness Ra of 60–100 nm was used for all polymers tested. Scratching was conducted on the smooth surfaces of the samples, which were pre-cleaned by an air duster. Since the hardness of the indenter was always much higher than that of the corresponding polymers, it can be considered as a rigid body for simplicity. Due to the various dimensions of the as-received plates, the scratch length was set in all cases at 15 or 20 mm. A scratch velocity of 10 mm/s was employed for all polymers, and the normal load in the progressive load scratch tests increased from 5 to 90 N. 2.4. Scratch damage analysis Test Load [mN] After testing, the depth of the scratches h, at the ﬁnal position of the indenter was measured (relative to the height level of the 1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 1035 normal specimen surface), using in some cases a violet laser scanning confocal microscope (Keyence VK 9700), but in most cases a white light proﬁlometer (FTR GmbH) having a lateral resolution of 1 mm and a vertical resolution of 3 nm. An average of at least two measured scratch depths is reported in this study. An ultra-high resolution ﬁeld emission scanning electron microscope (FESEM) (Carl Zeiss SMT AG, SUPRATM 40VP, Germany) was used to investigate further details of the scratch surface characteristics, operating with 10 kV acceleration voltage, a working distance of less than 10 mm, and a secondary electron detector. Therefore, the respective surfaces were sputter-coated for 70 s with gold. The scratch width, which was also measured from the laser proﬁlometer micrographs for the calculation of the scratch hardness Hs  is Hs ¼ FN =As ¼ ð4FN Þ=pw2 , ð2Þ where FN is the normal load, As is the sliding contact area, and w is the measured residual width at the end of the individual scratches (on the height level of the normal specimen surface). The actual width from hill top to hill top, wHH is actually wider and was also measured from the scratch pictures for comparison. Also considering the tangential force FT that is continuously measured during the scratch process, it is possible to deﬁne a PBI PI 2 0 1 2 3 4 5 6 7 8 PI 10 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Indentation Depth [μm] Fig. 4. Test load vs. indentation depth during Martens hardness test of PBI, PI 2, and PI 10. 500 450 Martens Hardness [MPa] 400 390 3.4 + 34 389 + 5.0 350 300 197 + 4.2 250 200 150 130 + 2.3 130 + 2.2 PI 10 PI 2 100 50 0 PBI PPP PEEK Fig. 5. Martens hardness values of the different polymers tested. K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 scratch coefﬁcient of friction SCOF as follows : SCOF ¼ FT =FN : ð3Þ A plot of the latter against the sliding distance should approach an equilibrium value (after a short scratching-in period) as long as no drastic changes in the scratch mechanisms during the course of the scratch occur. If, on the other hand, a transition from one scratch feature (e.g. microcracking) to another (e.g. plowing associated with material removal) takes place, this should also be associated with a change in SCOF (as it was shown in Ref. ). An estimation of the speciﬁc scratch wear rate wss, can also be determined from the ﬁnal parts of the scratches when considering the loss in volume of the material over an average short distance of the scratch lave, under an average normal load FNave, that acted over this distance wss ¼ DV =ðFNave lave Þ ¼ CSs lave =ðFNave lave Þ, where CSs refers to the cross-section of the scratch shortly before its end. The latter can be assumed to have the shape of a segment of a circle, with a chord length equal to w (residual width) and Scratch Coefficient of Friction  0.50 PI10_Sample9_Scratch2 90 0.40 80 0.30 70 Scratch Coefficient of Friction 0.20 0.10 ð4Þ 60 50 Normal Force 0.00 40 -0.10 Force [N] 1036 30 Tangential Force -0.20 20 -0.30 10 -0.40 0 5 10 Scratch Distance [mm] 0 20 15 Fig. 6. Typical course of the scratch loads and the scratch coefﬁcient of friction vs. scratch distance. Top View μm Topography [μm] z Width [mm] Length Profile x-x z μm Distance [mm] Distance [mm] Depth Profile z-z Topography [μm] wHH 580 μm 54 μm Distance [mm] Fig. 7. White light proﬁlometer results of a typical scratch on a PI 10 surface: (a) top view of the scratch with indication of proﬁles taken; (b) length proﬁle x–x of the scratch; and (c) depth proﬁle z–z, with clear evidence of the scratch valley in the middle and hills at both edges of the scratch (wHH ¼ 756 mm). K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 height equal to the scratch depth h: 2 2 CSs ¼ hð3h þ 4w Þ =ð6wÞ: ð5Þ 1037 produced. However, no remarkable differences were found for the two different interval times. Assuming further that FNave is nearly equal to FN at the end of the scratch, the speciﬁc scratch wear rate reduces to 3. Results and discussion wss ¼ CSs =FN , 3.1. Martens hardness of different polymers ð6Þ 3 ð7Þ In the present case, some of the residual post-scratch depths were measured with the laser confocal microscope one day after scratch testing, but the majority of the scratches was evaluated by white light proﬁlometry eight months after the scratches were 50 Penetration Depth Recovery PI10_Sample9_Scratch2 Scratch Depth [μm] 0 Residual Depth after Scratch -50 100 100 90 90 80 80 70 70 60 50 -100 40 Normal Force Penetration Depth during Scratch -150 -200 -250 0 5 10 15 30 60 50 40 30 20 20 10 10 Penetration Depth Recovery [%] PDR ¼ ðPDRPSDÞ=PD: Fig. 3 shows the typical hardness curves of three polymers (PBI; PPP; and PEEK), whereby each curve represents an average of 10 different measurements. The scatter in each case was very small. While PBI and PPP have almost the same shape, the curve for PEEK is much ﬂatter, representing a lower hardness and modulus of this polymer. However, PEEK is still clearly harder than the two graphite ﬁlled PIs, as shown in Fig. 4. In both cases, PBI can be considered as a reference curve. An evaluation of the Martens hardness values from the different curves results in the following order: PBI¼PPP 4PEEK 4PI 10 ¼PI 2 (Fig. 5). The very narrow scatter of the measured data is also indicated in this diagram. Normal Force [N] expressed in the units [mm /Nm]. The penetration depth recovery (PDR) in polymers  can also be quantiﬁed via the progressive load scratch test. For this, the penetration depth (PD) during the test must be recorded, and the residual post-scratch depth (RPSD) sometime after the test needs to be measured. As a result, the penetration depth recovery during the scratch process can be obtained by 0 0 20 Scratch Distance [mm] Fig. 8. Relationship between penetration depth and penetration depth recovery. PI 2 As Received Surface x x Scratch Direction μm μm Fig. 9. Laser confocal microscope picture of the scratch proﬁle on material PI 2: (a) top view, (b) 2D picture of the end part of the scratch surrounded by the normal, and as-received surface; and (c) length proﬁle through the middle of the scratch. (For interpretation of the references to colour in this ﬁgure, the reader is referred to the web version of this article.). 1038 K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 3.2. General scratch behavior A typical testing curve for polymers under the progressive load scratch test is shown in Fig. 6. Both the normal and the tangential (frictional) forces do not increase as perfectly linear, but slightly progressive. However, there is no point in the curves that can identify a clear onset of failure. When the normal load has reached its maximum level and the scratch test comes to a complete stop, there is always some movement of the indenter in the backwards direction, which has a minor inﬂuence on the penetration depth at this point (which was neglected in further considerations). The corresponding scratch coefﬁcient of friction PI 2 μm w = 726 μm h =67 μm μm wHH = 793μm Fig. 10. Depth proﬁle through the scratch on material PI 2. PI 2 741μm wHH =741 Scratch Surface SthSf Graphite Spalling Fig. 11. SEM pictures of the scratch on the surface of PI 2: (a) top view with height to height width wHH indicated (the latter is of similar size as measured from the laser confocal microscope scan); and (b)–(d) magniﬁcations of the surface inside the scratch, with local spots of graphite spalling. K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 (SCOF), also shown in Fig. 6 exhibit a large magnitude of ﬂuctuation at the beginning of the test, due to the inertial effects of instantaneously accelerating the scratch head to the designated scratch speed. After this, the SCOF increases slightly, smoothly and in a digressive manner to its ﬁnal level, again inﬂuenced at the end by the back-movement of the indenter. An evaluation of the scratch produced in this way (here: material PI 10), using white light proﬁlometry leads to the proﬁles shown in Fig. 7. The length proﬁle x–x of the residual scratch shows a non-linear course, with less material deformation (and scratch visibility) at the beginning and a progressively increasing 1039 residual scratch depth towards the end of the scratch (Fig. 7b). At this position the scratch depth is also slightly deeper than expected from the normal scratch course, due to the back-movement of the scratch head (as mentioned above). Besides, more or less material was pushed upwards at the edges of the scratch and in front of the indenter. The depth proﬁle z–z (Fig. 7c) allows obtaining more precise information on the remaining (residual) scratch width and depth at the end of the scratch. In the case of PI 10, the remaining crosssection has a segment of circle shape with a chord length of 580 mm and a height of 54 mm. The CSs appears, however, in a Normal Surface PI 10 Height Top wHH = 735μm Scratch Surface Graphite Spalling Fig. 12. SEM pictures of the scratch on the surface of PI 10: (a) top view with height to height.width wHH indicated (the latter is of similar size as measured from the white light proﬁlometer scan); and (b)–(d) magniﬁcations of the surface inside the scratch, with local spots of graphite spalling. Fig. 13. Laser confocal microscope picture of the scratch proﬁle on material PEEK: (a) top view, (b) 2D picture of the end part of the scratch; and (c) depth proﬁle through the scratch, with an indication of the actual scratch depth h, the width w, and the height to height width wHH. 1040 K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 different shape in the proﬁle shown, due to the different scales of the topography vs. distance axis. Fig. 8 shows a comparison of the measured penetration depth (PD) during scratching with the residual depth (RPSD) eight months after scratch. Both curves possess a slightly non-linear shape, with a progressive increase in scratch depth over the scratch distance. Using Eq. (7), the penetration depth recovery (PDR) is of the order 80% for the material used in this particular test (PI 10). 3.3. Scratch details of materials tested 3.3.1. Polyimide The two variants of polyimide (PI 2 and PI 10) differ only in the amount of graphite incorporated into the polyimide matrix. This results in smaller differences in the mechanical properties, so the appearance of the scratches on both variants is also rather similar. Fig. 9 shows the end region of the scratch at which the 90 N load level was reached. The different colors shown in Fig. 9a refer to different height levels, with the blue area (inside of the scratch) being below the normal surface level (orange), and the red rim around the scratch lying above the level of the specimen’s surface. Fig. 9b shows a 2D view of the scratch, following the original scratch direction, so that the deeper valley and the pushed-up hills at the rim of the scratch are more visible, relative to the as-received surface of the sample. The length proﬁle x–x through the scratch and its edges (Fig. 9c) indicates a slightly rougher appearance of the scratch surface shortly before the indenter came to a ﬁnal stop that is associated with a little backwards movement. Here the proﬁle indicates a slightly deeper scratch depth than shortly before this point was reached. It is followed by a steep reduction in scratch depth up to the heightened rim region, before the normal plateau of the plate became evident. A corresponding 2D plot of the scratch with an indicated depth proﬁle z–z through the deepest point of the scratch is shown in Fig. 10a. The measured values here (depth h¼67 mm; and width w¼726 mm) were used for further calculations (Fig. 10b). The segment of circle shape of the CSs is much more visible in the present case, as compared to Fig. 7c. SEM micrographs of the scratches on the material PI 2 (Fig. 11) and PI 10 (Fig. 12) illustrate that the surfaces inside of the scratches are relatively smooth. Only higher magniﬁcations of PEEK Normal Surface wHH = 721 μm Height Top Scratch Surface Fig. 14. SEM pictures of the scratch on the surface of PEEK: (a) top view with height to height width wHH indicated (the latter is of similar size as measured from the laser confocal microscope scan); and (b) magniﬁcation of the heightened edge of the scratch, in relation to the scratch and normal surface. PPP Scratch Direction Extrusion Direction Sharp Features μm w = 410 μm h = 11 μm μm wHH = 466μm Fig. 15. Laser confocal microscope picture of the scratch proﬁle on material PPP: (a) top view; (b) 2D picture of the end part of the scratch; and (c) depth proﬁle through the scratch, indicating the actual scratch depth h, the width w, and the height to height width wHH. K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 these surfaces show local sites of material ﬂaking and these features are more frequently observed for PI 10 than for PI 2 (cf. Figs. 11c and 12c). Therefore, it seems to be reasonable to assume that these spalling effects are associated with the graphite ﬂakes ﬁlled into the PI matrix for better friction behavior, i.e., lower SCOF. When comparing the width of the scratch end regions (from hill top to hill top) shown in Figs. 11a and 12a, it becomes obvious that the PI 2 possesses a slightly lower scratch resistance than PI 10. 1041 3.3.2. Polyetheretherketone The corresponding scratch pattern on the PEEK surface is shown in Fig. 13. Both, the width and the depth of the scratch are clearly smaller than that in case of the PI-samples. Otherwise, the features look very similar. This is also true when the SEM of the PEEK scratches is observed (Fig. 14). No spectacular surface damage features are visible within the scratch region. Only some pushed-up material hills clearly indicate the scratch contour. Fig. 16. SEM pictures of the scratch on the surface of PPP: (a) top view with height to height width.wHH indicated (the latter is of similar size as measured from the white light proﬁlometer scan); (b) sharp crack features on both side of the scratch; and (c) and (d) magniﬁcations of the periodic cracks, starting from the inside of the scratch and pointing opposite to the scratch direction. Fig. 17. Laser confocal microscope picture of the scratch proﬁle on material PBI: (a) top view; (b) 2D picture of the end part of the scratch; and (c) depth proﬁle through the scratch, indicating the actual scratch depth h, the width w, and the height to height width wHH. 1042 K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 3.3.3. Polyparaphenylene When it comes to the PPP-samples, the contours of the scratches are even ﬂatter than that of PEEK, indicating a higher scratch resistance of this material. Fig. 15 shows that only a small amount of material is pushed upwards at the edges of the scratch and in front of it. Some sharp features are visible in these regions, which may be a result of periodic, and parabolic cracks pointing opposite the scratch direction (as it has been observed in previous studies on epoxy and polycarbonate ). But this conjecture can only be validated after higher resolution SEM analysis. In fact, the SEM micrographs give a clear evidence for the existence of such cracks on both sides of the scratch (Fig. 16). It seems that each pair of the cracks on both sides belongs to one parabolic crack, of which the middle part was smeared over by a thin layer of material as the stylus moved passed it in the scratch direction. observed, which might have been the reason for the roughened features seen with the laser confocal microscope. 3.4. Comparison between different materials 3.4.1. Scratch depth, wear rate and SCOF When considering the residual scratch depth and the measured scratch coefﬁcient of friction, both seem to be related to some degree. However, this relationship is not linear, but seems to be a matter of the material’s characteristic. PBI and PPP, both possessing high modulus, high strength, but a rather brittle behavior exhibit low scratch depth and low SCOF (Fig. 19). On the other hand, the more ductile polymers (here PEEK, PI 10, and PI 2) possess a higher SCOF that is associated with a higher level 90 Scratch Depth h [μm] 80 3.3.4. Polybenzimidazole It is even more difﬁcult to recognize the scratches on the PBI samples (Fig. 17). Within the roughness environment on the surface of the machined samples, the scratch contour is only visible with the laser confocal microscope due to some pushed-up material chunks at the edges. Surprisingly, the scratches are clearly detectable when the samples are observed with the naked eye. Also by SEM, the scratch contour on the PBI surface is hardly visible (dotted line in Fig. 18a). Outside this border line, the specimen surface is characterized by a granular topography, which seems to be due to the sintering process of the ﬁne PBI powder during the compression molding of the plates. Within the scratch, the surface structure looks a bit smoother, obviously a result of the ironing effect by the moving stylus. But no deep scratch effects are otherwise visible. Only at the edges some small cracking phenomena (pointing in the scratch direction) can be 70 60 PI 10 PI 2 50 PEEK 40 30 20 PPP 10 PBI 0 0 0.10 0.15 0.20 0.25 0.30 0.35 0.40 Scratch Coeff. of Friction SCOF  0.45 Fig. 19. Relationship between scratch depth and scratch coefﬁcient of friction for the different materials tested. Fig. 18. SEM pictures of the scratch on the surface of PBI: (a) top view with height to height width wHH indicated (the latter is of similar size as measured from the white light proﬁlometer scan); and (b)–(d) different magniﬁcations of the boundary between scratch and surrounding, normal material surface, with some indication of short cracks, starting from the edge into the scratch interior. K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 of scratch depth. Between the two PI’s, the one with the higher content of graphite (PI 10) shows a slightly lower SCOF. This conﬁrms an earlier statement by others, suggesting that a lower 90 80 PI 2 Scratch Depth h [μm] 70 PI 10 60 50 PEEK 40 30 20 coefﬁcient of friction is desirable for polymers to be more scratch resistant . The depth of a scratch is also associated with a ﬁnal removal of material from the scratched surface. The latter can be expressed by the speciﬁc scratch wear rate. Plotting the residual scratch depth against this wear rate yields an almost linear relationship between the two quantities (Fig. 20). This is not surprising, since the scratch depth was also used for the calculation of the scratch wear rate (according to Eqs. (4)–(6)). In the future we will compare this scratch wear rate with wear rates obtained from different tests, such as sliding against smooth steel or abrasive paper in order to see if any good correlations between these tests exist. At least from a previously published study by Sinha et al.  it seems that the projected wear debris area after scratching can be correlated with the speciﬁc wear rate found in pin-on-disc experiments with different polymers. PPP 10 PBI 0 0 100 150 200 300 250 350 50 „ Spec. “Scratch Wear Rate Wss [10-3mm3/ Nm] 400 Fig. 20. Relationship between scratch depth and speciﬁc scratch wear rate for the different materials tested. 100 Penetration Depth Recovery [%] 1043 90 80 70 3.4.2. Penetration depth recovery Polymers are viscoelastic and viscoplastic in nature. A portion of the material at the rear edge of the scratcher recovers almost instantaneously after the scratcher moves forward . It is obvious that the more elastic recovery, the better it becomes to minimize the scratch visibility. The recovery of polymers can be estimated by the penetration depth recovery upon removal of the applied scratch stress . A typical testing curve for polymers under the progressive load scratch test is shown in Fig. 8, along with the calculated penetration depth recovery. In general, a relatively high viscoelastic recovery (70%–98%) was found for all the polymers investigated (Fig. 21). The material parameters that inﬂuence the level of recovery appear to be complex. They are being subjected to further investigation. 60 50 3.5. Correlations with other material properties 40 30 20 10 0 PBI PPP PEEK PI 10 PI 2 Fig. 21. Penetration depth recovery for the different materials tested. σB h E [μm] [GPa] [MPa] 90 225 9 80 8 200 70 7 175 An attempt to correlate the scratch results with the mechanical properties of the materials (as listed in Table 1) did not show any simple relationship (Fig. 22). This was also found out by Sue et al.  who ﬁnally concluded that different sets of material parameters and other relevant factors may be needed to properly correlate with different scratch damages. This statement was based on experiments with polymers of different speciﬁc εB anI [J/m] [%] 90 90 σB anI E h 80 80 70 70 60 60 60 6 150 50 5 125 50 50 40 4 100 40 40 30 3 75 30 30 20 2 50 20 20 10 1 25 10 10 0 0 0 0 0 εB PBI PPP PEEK PI 10 PI 2 Fig. 22. Comparison between the residual scratch depth (h) and different mechanical properties (E¼ elastic modulus; sB ¼ strength; anI ¼notched Izod impact energy; and eB ¼strain at break) of the different materials tested. 1044 K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 90 Table 2 Calculation of the apparent strain e within the ﬁnal stage of scratching, using a modiﬁed Tabor relationship  of the form e ¼0.2a/(R2 a2)1/2, where a¼ w/2 (see Eq. (2)), and R¼ radius of indenter (0.5 mm). Average values were used for the calculation. Scratch Depth h [μm] 80 70 60 PI 2 PI 10 50 PEEK 40 PPP PBI 0 0 2 3 4 5 6 7 Elastic Modulus E [GPa] 8 PBI performance prod. Celazole U60 SD Solvay adv. polymers PrimoSpire PR 120 Victrex PEEK 450 G DuPont (Vespel) ST 2002 (2% Gr) DuPont (Vespel) ST 2010 (10% Gr) 90 80 70 60 PI 2 PI 10 50 PEEK 40 30 20 PBI 10 PPP 0 0 10 40 50 60 20 30 Notched Izod Impact Energy [J/m] 70 80 90 80 70 60 PI 2 PI 10 50 40 PEEK 30 20 PPP PBI 0 0 5 10 15 20 25 Strain at Break [%] 30 35 PI 10 h (mm) w (mm) e (%) 5.5 242 4.5 10.6 375 8.0 48.3 57.0 538 650 12.7 17.1 52.0 573 14.0 9 Fig. 23. Relationship between scratch depth and elastic modulus for the different materials tested. Scratch Depth h [μm] PBI PEEK PI 2 20 10 Scratch Depth h [μm] Manufacturer company code PPP 30 10 Short name 40 Fig. 24. Relationship between scratch depth and notch Izod impact energy (a) and strain at break (b). performance: (i) ductile and strong; (ii) ductile and weak; (iii) brittle and weak; and (iv) brittle and strong. In spite of this fact, a ﬁrst indication for the scratch resistance of a polymeric material is the elastic modulus. Sue et al.  showed that for a variety of polymers, a higher modulus was also associated with reduction in the residual scratch depth. This is also evident in the results of this study, as shown in Fig. 23. A similar tendency became visible when plotting the scratch depth vs. the tensile strength of these materials. On the other hand, a comparison of the scratch depth with the two properties representing the notch sensitivity (anl) and the ductility (eB) tested in this study yielded trends, which hardly allow any serious conclusions to the materials’ scratch resistance (Fig. 24a and b). A better correlation can be found, on the other hand, when comparing the scratch depth with the apparent strains during scratching. The calculation of which is given in Table 2 . A plot of the scratch depth and the SCOF against the scratch hardness, calculated according to Eq. (1) leads to a similar shape of the curve as it was seen for h vs. E (Fig. 25). Also, this is in good agreement with previous ﬁndings made by Sue et al. . It must be mentioned that the scratch hardness values are a function of the geometry and material type of the scratch stylus, but is also dependent on the scratch velocity and testing temperature. Briscoe et al. [21,22] measured values in the range between 70–800 MPa for various polymers, including polytetraﬂuoroethylene (PTFE), polyethylene (UHMW-PE), polymethylmethacrylate (PMMA), PEEK and polycarbonate (PC). This means that the values found here are in a reasonable level, especially when considering that here we are dealing with the hardest polymers available so far (i.e., PPP; and PBI). Finally, it is noted that in the present case a relatively linear relationship is found when plotting the scratch depth against the Martens hardness (Fig. 26). Since HM is an easily measurable value, this parameter can be used to estimate the polymeric materials’ scratch resistance in terms of residual scratch depth. However, other works  have shown that HM is only useful for simple surfaces that exhibit simple damage features, such as ironing. If the surface exhibits skin-core characteristics or possesses a complex morphology, HM would no longer be a good parameter to correlate with scratch deformation and damage. This also means if events like cracking and ductile drawing are present, HM may not be a good parameter to consider (although in the present case, in spite of crack features in PPP or ductile plowing events in case of PI and PEEK), the relationship did not look bad. Nevertheless, signiﬁcant work is still needed to establish the fundamental structure-property relationships for the scratch behavior of high performance polymers. 4. Conclusion From this study it can be concluded that the scratch behavior of various high performance polymers follows, in principle, the K. Friedrich et al. / Tribology International 44 (2011) 1032–1046 1045 90 0.45 PI 2 0.40 Scratch Depth h [μm] 70 0.35 PI 10 60 0.30 PEEK 50 0.25 40 0.20 30 0.15 PPP 20 0.10 PBI 10 0 0 250 500 750 1000 1250 1500 Scratch Hardness HS [MPa] 1750 Scratch Coeff. of Friction SCOF  80 0.05 0 2000 Fig. 25. Plot of scratch depth and scratch coefﬁcient of friction against scratch hardness for the different materials tested. However, this relationship is more reliable if all materials involve the same damage features and do not have any effects due to additives 90 80 Scratch Depth h [μm] 70 PI 2 60 PI 10 PEEK 50 40 30 20 PBI 10 PPP 0 0 50 100 150 200 250 Martens Hardness HM [MPa] 300 350 400 Fig. 26. Relationship between scratch depth and Martens hardness for the different materials tested. same scheme as it has been observed for other commodity or engineering plastics. However, since the polymers investigated here possess a relatively high elastic modulus, the scratch recovery was higher, and in some cases the scratches were more difﬁcult to detect than in previous studies, where lower performance polymers were tested. A simple correlation of the scratch resistance with other material properties could not be found here, i.e., it is a combination of different properties and other factors, which determines how deep the scratches on a particular polymeric material can be. Although one can state, in general, that a higher modulus or a higher Martens hardness of the polymer are favorable for a lower scratch sensitivity. The relative role of modulus or hardness vs. ductility or impact toughness still remains unclear. The most scratch resistant material in this study was polybenzimidazole. Acknowledgments One of us, K. Friedrich, is grateful to the German Research Foundation (DFG FR 675/51-1) for initiating the collaboration with the colleagues at Texas A&M University, USA. The authors would also like to acknowledge the Distinguished Scientist Fellowship Program at the King Saud University, Saudi Arabia, for the ﬁnancial support of this project. Special thanks are also given to Polymer Scratch Consortium at Texas A&M University for the access and usage of its research facility. References  Jiang H, Browning R, Sue H-J. Understanding of scratch induced damage mechanisms in polymers. Polymer 2009;50:4056–65. 1046 K. Friedrich et al. / Tribology International 44 (2011) 1032–1046  Zum Gahr K-H. Microstructure and wear of materials. Amsterdam: Elsevier; 1987.  Wong M, Lim GT, Moyse A, Reddy JN, Sue H-J. A new test methodology for evaluating scratch resistance of polymers. Wear 2004;256:1214–27.  Wong M, Moyse A, Lee F, Sue H-J. Study of surface damage of polypropylene under progressive loading. J Mater Sci 2004;39:3293–308.  Xiang C, Sue H-J, Chu J, Masuda K. Roles of additives in scratch resistance of high crystallinity polypropylene copolymers. Polym Eng Sci 2001;41:23–31.  Chu J, Xiang C, Sue H-J, Hollis RD. Scratch resistance of mineral-ﬁlled polypropylene materials. Polym Eng Sci 2000;40:944–55.  /http://www.celazolepbi.com/ﬁles/_u60/CELAZOLE-U-60-PBI-Perf_002.pdfS.  /http://www.solvayadvancedpolymers.com/static/wma/pdf/7/9/8/2/Primos pire_PR120.pdfS.  /http://www.victrex.com/docs/datasheets-docs/450G.pdfS.  /http://www2.dupont.com/Vespel/en_US/assets/downloads/vespel_s/ H74544.pdfS.  Grellmann W, Seidler S, editors. Munich, Germany: Hanser; 2007.  Briscoe BJ, Fiori L, Pelillo E. Nano-indentation of polymeric surfaces. J Phys D: Appl Phys 1998;31:2395–405.  ASTM International, ASTM D7027-05; 2005.  International Organization for Standardization, ISO 19252:2008; 2008.  Xiang C, Sue H-J, Chu J, Coleman B. Scratch behavior and material property relationship in polymers. J Polym Sci B: Polym Phys 2001;39:47–59.  Browning, R, Sue, H-J: Minkwitz, R, Charoensirisomboon, P. Effect of acrylonitrile content on the scratch behavior of styrene–acrylonitrile random copolymers. Polym Eng Sci, in press. doi:10.1002/pen.22003.  Sinha SK, Chong WLM, Lim S-C. Scratching of polymers—modeling abrasive wear. Wear 2007;262:1038–47.  Briscoe BJ, Evans PD, Biswas SK, Sinha SK. The hardnesses of poly(methylmethacrylate). Tribol Int 1996;29:93–104.  Jardret V, Zahouani H, Loubet JL, Mathia TG. Understanding and quantiﬁcation of elastic and plastic deformation during a scratch test. Wear 1998;218:8–14.  Pelletier H, Le Houerou V, Gauthier C, Schirrer R. Scratch experiments and ﬁnite element simulation: friction and nonlinearity effects. In: Sinha SK, Briscoe BJ, editors. Polymer Tribology. London, UK: Imperial College Press; 2009. p. 108–40. Chapter 4.  Stuart BH, Briscoe BJ. Scratch hardness studies of poly (ether ether ketone). Polymer 1996;37:3819–24.  Briscoe BJ, Pelillo E, Sinha SK. Scratch hardness and deformation maps for polycarbonate and polyethylene. Poly Eng Sci 1996;36:2996–3005.  Browning R, Lim G, Moyse A, Sun L, Sue H-J. Effects of slip agent and talc surface treatment on the scratch behavior of TPOs. Poly Eng Sci 2006;46: 601–608.