Original Article
The influence of nominal stress on wear factors of carbon fibre–reinforced polyetheretherketone (PEEKOPTIMAÒ Wear Performance) against zirconia toughened alumina (BioloxÒdelta ceramic)
Proc IMechE Part H: J Engineering in Medicine 2014, Vol. 228(6) 587–592 Ó IMechE 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0954411914538783 pih.sagepub.com
Andrew Evans1, Henrietta Horton2, Anthony Unsworth3 and Adam Briscoe4
Abstract Carbon fibre–reinforced polyetheretherketone is an attractive alternative to ultra-high-molecular-weight polyethylene in artificial joints, but little has been published on the influence of stress on the wear factor. We know that in ultra-highmolecular-weight polyethylene, the wear factor reduces as the normal stress increases, which is counter-intuitive but very helpful in the case of non-conforming contacts. In this study, carbon fibre–reinforced polyetheretherketone (PEEKOPTIMAÒ Wear Performance) has been investigated in a pin-on-plate machine under steady loads and under stresses typical of hip and knee joints. At stresses below about 6 MPa, wear factors are between 10 and a 100 times lower than for ultra-high-molecular-weight polyethylene but at higher stresses the wear factors increase substantially.
Keywords Wear analysis/testing, bio-composites, biomaterials, tribology of materials
Date received: 13 March 2014; accepted: 14 May 2014
Introduction The demands of patients for replacement joints that allow the resumption of more athletic or more demanding work-based activities have led to the design of hip joints with larger and larger diameters. In order to cope with the potential extra wear rates generated by these activities, new material combinations have been investigated including the use of carbon fibre–reinforced polyetheretherketone (CFR-PEEK) cups and both alumina and zirconia toughened alumina (ZTA) (BioloxÒdelta) femoral head replacements. These have been shown to have extraordinarily low wear rates,1 which are only one-hundredth of a Charnley hip (22 mm diameter metal on ultra-high-molecular-weight polyethylene (UHMWPE)) or one-tenth of highly cross-linked polyethylene (HXLPE) hip joints. CFR-PEEK wear debris is well tolerated in the body,2 and CFR-PEEK components can be moulded or machined to the finished size and shape. Deliberate fibre orientation can reduce the wear even more.3
If CFR-PEEK is to be used more widely to replace UHMWPE or HXLPE, then a detailed knowledge of how the wear rate is affected by the stress level imposed is important to evaluate. For example, UHMWPE has been shown to have very high wear factors at low stresses. These reduce rapidly with increasing stress level4 and while this seems counter-intuitive, it has been confirmed in other laboratories5 though some have failed to observe this phenomenon.6,7 From a design point of view, it is important to know the wear factor characteristics of any proposed material 1
Mott MacDonald, London, UK Bristol University, Bristol, UK 3 School of Engineering and Computing Sciences, Durham University, Durham, UK 4 Invibio Ltd, Thornton-Cleveleys, UK 2
Corresponding author: Anthony Unsworth, School of Engineering and Computing Sciences, Durham University, Durham DH1 3LE, UK. Email: [email protected]
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
588
Proc IMechE Part H: J Engineering in Medicine 228(6)
Figure 1. Diagram of pin-on-plate apparatus. 1: weight; 2: lever arm; 3: drive motor; 4: 12-V DC motor; 5: pin holder; 6: level sensor; 7: heated lubricant bath; 8: CFR-PEEK test pin; 9: ceramic test plate; DC: direct current; CFR-PEEK: carbon fibre–reinforced polyetheretherketone.
combination since these may affect which joint (hip, knee and ankle) or joint design can utilise this material fully. This article describes the use of a pin-on-plate machine to measure the wear factors of CFR-PEEK against ZTA ceramic as stress levels increase.
Apparatus A four-station pin-on-plate machine was used for these experiments (Figure 1). Four rectangular ZTA ceramic plates were fixed into the reciprocating sledge that was heated by embedded thermostatically controlled electrical heaters. Pins made from CFR-PEEK were held in rotating pin holders each of which was rotated by an individual 12-V direct current (DC) motor. The frequencies of the rotation of the pins and reciprocation of the plates were independently variable; however, throughout this work, they were both set to 1 Hz. The pins were subjected to uni-directional rotation while loaded against the plates that were undergoing reciprocation. Loading of the pins on the plates was provided by ‘dead-weights’, which were placed at different lever arm positions to increase or decrease the loads (and stresses) on the specimens. The transmitted loads were measured by a transducer to take account of any friction in the pin holders and lever arm bearings rather than relying on the moment-arm calculations alone. The reciprocating sledge incorporated a lubricating bath containing a liquid, in this case new-born calf serum in de-ionised water. In order to maintain the lubricant level when evaporation was taking place, a system comprising a level sensor, a de-ionised water reservoir and a small electric pump was used to top-up the bath to a set level automatically. The whole machine was covered by a Perspex cover to minimise dust penetration into the lubricant.
Materials and methods CFR-PEEK pins The pins were made from CFR-PEEK (pitch based), 30% fibre fill by weight (PEEK-OPTIMAÒ Wear Performance) supplied by Invibio Ltd, ThorntonCleveleys. Pins with a wearing face of 4.0, 5.0, 6.0, 7.5 and 9.5 mm diameter and total length of 18 mm were used. The 4.0-mm-diameter pins also had a central section of 7 mm diameter to help prevent buckling when carrying the higher loads or bending under combined loading and sliding (Figure 2).
ZTA ceramic plates Biolox delta plates were supplied by CeramTec and measured 48 3 24 3 5 mm. Each had a surface finish of Ra = 0.002 mm. They were aluminium oxide matrix composite ceramic comprising approximately 75% alumina (Al2O3), 24% zirconia (ZrO2) and other trace elements largely to resist crack propagation.
Lubricant The lubricant was de-ionised water with 25% bovine serum added to give a protein content of 18 g/L. Sodium azide (0.2%) and 20 mM ethylenediaminetetraacetic acid (EDTA) were added to reduce bacterial growth and calcium deposition, respectively.
Methods The pin-on-plate machine was thoroughly cleaned and the rotating pin holders lubricated to enable free movement of the rotating pins. A pre-calibrated load transducer was fitted to the reciprocating carriage and the
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
Evans et al.
589 every 300,000 cycles the experiment was halted; the specimens were removed, cleaned, dried and weighed according to the protocol in Appendix 1. Four pins and four plates were actively loaded and subjected to motion during each experiment while a fifth pin and plate acted as a control for fluid absorption. All specimens were inserted into the same lubricant at the same temperature, but the control pins were not subjected to any wearing process. All specimens were cleaned and weighed in exactly the same way using a Mettler-Toledo Balance sensitive to 0.01 mg. A minimum of three measurements were recorded for each sample, and the average weight loss was converted to volumetric wear using the density of CFR-PEEK of 1.42 mg/mm3.8
Figure 2. Pin shape to resist buckling.
Imaging pins were loaded against the transducer to check for any friction in the loading mechanism. This ensured that the load that was applied to the pins was transmitted to the plates. This calibration was used to calculate the actual stresses on the pins. The desired loading was obtained by adding weights at various moment arms on the levers. Table 1 shows the loads, the cross-sectional areas of the wear face of the pins and the contact stresses used in these experiments. The loads were applied and the arms adjusted until they were horizontal, assuring the true load was being applied. As the pins wore, the arms were adjusted to maintain their horizontal position. Fresh lubricant was added to the heated bath and the temperature brought to 37 °C. The frequencies of reciprocation and rotation were both set to 1 Hz and the magnetic cycle counter was set to 0. The stroke length of the reciprocating sledge was set to 25 mm. Lubricant was replenished daily and approximately
The surfaces were examined using both non-contacting white light interferometry (NewView 100; Zygo) and light microscopy. Ten readings were taken for each surface and the worn surfaces were compared with the control surfaces.
Results Figures 3 and 4 show the graphs of volumetric wear against sliding distance for specimens 1–8. These were the higher wear factor specimens obtained at higher stresses. There was little obvious running-in wear, although three specimens, which were from those with lower wear factors (pins 14, 15 and 16; Table 1), did show signs of running-in wear.
Wear factors Table 1 shows the wear factors obtained from all the experiments together with the final roughness of the
Table 1. Stresses, wear factors, roughness and skewness for all the pins. Pin number
Contact stress (MPa)
Steady-state wear factor (31026 mm3/N m)
1 6.85 0.123 2 5.53 0.417 3 5.83 0.064 4 6.85 1.39 5 5.99 1.54 6 8.73 1.10 7 9.36 3.68 8 6.98 2.52 9 0.533 0.0282 10 1.227 0.0311 11 3.346 0.0075 12 1.917 0.035 13 9.746 1.58 14 7.477 0.0167 15 3.346 0.035 16 5.065 0.059 Average initial roughness before the wear tests began
Final roughness, Sq (mm)
Final skewness, Ssk (mm)
Load (N)
Diameter of the pin (mm)
0.130 0.124 0.145 0.110 0.119 0.100 0.196 0.096 0.67 0.48 0.368 0.929 0.419 0.255 0.716 0.468 2.29
22.231 23.39 23.10 20.85 26.61 22.73 28.3 20.45 210.36 210.61 211.06 25.61 27.44 22.84 29.35 27.67 20.455
86.2 69.5 73.2 86.1 75.3 109.8 117.6 87.8 37.8 54.2 65.7 54.2 191.4 146.8 65.7 99.5
4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 9.5 7.5 5.0 6.0 5.0 5.0 5.0 5.0
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
590
Proc IMechE Part H: J Engineering in Medicine 228(6)
Figure 3. Wear of CFR-PEEK pins (1–4) sliding against ceramic plates.
Figure 5. Pin 5; 5.99 MPa after 2 million cycles.
higher stresses, the wear factors increase dramatically to way above the values for UHMWPE.
Discussion
Figure 4. Wear of CFR-PEEK pins (5–8) sliding against ceramic plates.
pins (Sq) and the final skewness (Ssk) of the surface topography. It is seen that compared with the roughness and skewness prior to starting the wear experiments, the roughness reduced by an order of magnitude while the skewness became much more negatively distributed.
Surface topography Figure 5 illustrates the worn surface of the CFR-PEEK where the cross sections of the carbon fibres can be seen together with some small areas of fibre pull-out when the fibres were lying axially along the surface.
Wear factors against contact stress Figure 6 shows all the wear factors plotted against the contact stress, and it is clear that below about 6 MPa, the wear factors are exceptionally small at about 1028 mm3/N m. This is about two orders of magnitude lower than UHMWPE against stainless steel.9,10 However, at
The first and very striking observation is how very low the wear factors are for CFR-PEEK when the stress levels are small (Figure 6). Indeed, typically, the wear factor at 2 MPa would be 0.035 3 1026 mm3/N m compared with 1.1 3 1026 mm3/N m for UHMWPE against stainless steel. However, at a stress level of 9 MPa, the wear factor can exceed 3 3 1026 mm3/N m. Others have also found this characteristic for CFRPEEK including Wang et al.,11 whose results were consistent with these. This becomes more startling when the results from Vassiliou and Unsworth4 are considered on the same axes (Figure 6). These comprehensive studies of the wear of UHMWPE against stainless steel show the unusual properties of UHMWPE under different levels of stress. At low levels of stress, the wear factors are very high but fall by two orders of magnitude as the stress increases. This is counter-intuitive but is seen to be true in practice. UHMWPE, when used in hip replacements where the stress is relatively low due to the high conformity of the contact, is found to have a wear rate of typically 50 mm3 per million cycles (depending on radius of the joint). On the other hand, most knee joints are less conforming and hence have higher stresses; yet, their wear rate is typically 3 mm3 per million cycles.12 The exception is for conforming knee designs that have wear rates that are twice this level, but still considerably less than in hips. So UHMWPE showed lower wear under high-stress applications such as knee joints, while CFR-PEEK against alumina in large diameter hip applications1 showed extremely low wear rates compared with UHMWPE. In knees, while a useful reduction in wear factor was seen using CFR-PEEK, this was much less of an improvement than in hips.12 While a great deal of consistency was observed between the wear factors of CFR-PEEK below a stress
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
Evans et al.
591
Figure 6. Wear factors of CFR-PEEK against ceramic and UHMWPE against stainless steel.4 UHMWPE: ultra-high-molecular-weight polyethylene.
level of 6 MPa, above this the variation was much greater. As an example in Figure 3, consider pins 1 and 4. Here, the loads were equal on both pins as were the cross-sectional areas of their contact faces; yet, the wear factors were an order of magnitude different. At low contact stresses, pins 10 and 12, for example, even though the diameter changed from 9.5 to 6 mm, the wear factors were almost identical. As with the UHMWPE experiments,4 the CFR-PEEK failed to demonstrate any relationship between wear factors and radius or load though it should be remembered that wear factor itself includes the load, so load is included in both the ordinate and abscissa (stress). There are two other concerns expressed by some surgeons regarding carbon fibre–filled polymers, one of which is whether the carbon will scratch and wear the ceramic counter-face. The other is whether the fibres will dislodge from the matrix and cause irritation of the surrounding soft tissues. The variations in mass of the active ceramic plates are comparable to those of the control plates, which have done no sliding against the PEEK pins. If the maximum wear volumes are calculated for the plates, then these are generally below 0.07 mm3, which translate to wear factors of the order of 3 3 1029 mm3/N m, which are three orders of magnitude lower than the pins. This suggests that this wear has no significance whatever so this worry would seem to be unfounded based on these limited experiments. Turning to the issue of dislodged carbon fibres, Figure 5 indicates that it can happen (see the dark areas) but it seems to be relatively small and associated with fibres that are almost completely worn through with only a small amount of surface area still in contact with the matrix. There are many examples of the fibres wearing gradually at the same rate as the PEEK matrix rather than exhibiting fibre pull-out13 and this was generally seen in this work. Wear of CoCrMo alloy counter-faces has also proved to be of little concern since it is generally two to
three orders of magnitude lower than the wear of the CFR-PEEK components,14 which are themselves much lower than the wear of UHMWPE. It should, however, be stated that we did not systematically study the scratching of the metal counter-faces in our experiments to obtain the wear factors of CFR-PEEK.
Conclusion The wear factors of CFR-PEEK against ZTA ceramic at stresses below about 6 MPa are extremely low and between about one-hundredth and one-tenth of the wear of UHMWPE against stainless steel, depending on the stress level. However, at higher stress, the wear rate was shown to increase dramatically. This suggests that CFR-PEEK against ZTA would be very good in low-stress situations such as hip joints but would be less effective in higher stress applications. However, CFR-PEEK was not seen to be subject to fibre pullout, or to excessive scratching of the ceramic plates. Acknowledgements This work was carried out as part of the SHIELD project and the authors would like to thank TSB for the funding as well as Invibio Ltd and CeramTec for supplying the specimens. Declaration of conflicting interests Dr Adam Briscoe is a paid employee of Invibio Ltd. Funding This work was carried out as part of the TSB SHIELD Project. Research specimens were provided by Invibio Ltd and CeramTec. References 1. Scholes SC, Inman IA, Unsworth A, et al. Tribological assessment of a flexible carbon-fibre-reinforced
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
592
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Proc IMechE Part H: J Engineering in Medicine 228(6) poly(ether-ether-ketone) acetabular cup articulating against an alumina femoral head. Proc IMechE, Part H: J Engineering in Medicine 2008; 222: 273–283. Toth JM. Chapter 7. Biocompatibility of polyaryletheretherketone polymers. In: Kurtz SM (ed.) PEEK biomaterials handbook. Oxford: Elsevier, 2012, pp.81–92. Scholes Susan C, de Jong S and Anthony U. Effect of carbon fibre orientation on the wear of CFR-PEEK/ CFR-PEEK bearing couples. In: Proceedings of 21st annual congress of the International Society for Technology in Arthroplasty, Seoul, Korea, 1–4 October 2008, p.190. Vassiliou K and Unsworth A. Is the wear factor dependent on nominal contact stress in UHMWPE contacts? Proc IMechE, Part H: J Engineering in Medicine 2004; 218: 101–107. Barbour PSM, Barton DC and Fisher J. The influence of contact stress on wear of UHMWPE for total replacement hip prostheses. Wear 1995; 181–183: 250–257. Rostoker W and Galante JO. Contact pressure dependence of wear rates of ultra high molecular weight polyethylene. J Biomed Mater Res 1979; 13(6): 957–964. Rose RM, Goldfarb HV, Ellis E, et al. On the pressure dependence of the wear of ultra high molecular weight polyethylene. Wear 1983; 92(1): 99–111. Invibio Biomaterial Solutions. MOTIS polymer typical material properties, http://www.invibio.com (2009, accessed 23 March 2012). Joyce T, Unsworth A, Cartwright T, et al. Preliminary comparison of the wear of silane XLPE with that of UHMWPE, both polymers rubbing against hard counterfaces (abstract). J Bone Joint Surg Br 81(Supp. 1): 85. Joyce T, Monk D, Scholes SC, et al. A multi-directional wear-screening device and preliminary results of UHMWPE articulating against stainless steel. Bio-Med Mater Eng 2000; 10: 241–249. Wang A, Lin R, Stark C, et al. Suitability and limitations of carbon fiber reinforced PEEK composites as bearing surfaces for total joint replacements. Wear 1999; 225–229: 724–727.
12. Scholes SC and Unsworth A. Pitch-based carbon-fibrereinforced poly (ether-ether-ketone) OPTIMA assessed as a bearing material in a mobile bearing unicondylar knee joint. Proc IMechE, Part H: J Engineering in Medicine 2009; 223: 13–25. 13. Kurtz SM and Nevelos J. Chapter 16. Arthroplasty bearing surfaces. In:Kurtz SM (ed.) PEEK biomaterials handbook. Oxford: Elsevier, 2012, pp.261–276. 14. Scholes SC and Unsworth A. Wear studies on the likely performance of CFR-PEEK/CoCrMo for use as artificial joint bearing materials. J Mater Sci: Mater Med 2009; 20(1): 163–170.
Appendix 1 Cleaning protocol
Rinse in de-ionised water Wipe with lint-free tissue (all surfaces) Place in ultrasonic bath in de-ionised water for 10 min at 18 °C Rinse in distilled de-ionised water Place in ultrasonic bath in 1% NeutraconÒ solution for 10 min Rinse in de-ionised water Place in ultrasonic bath in de-ionised water for 10 min Rinse in distilled de-ionised water (do not wipe) Place in ultrasonic bath in de-ionised water for 3 min Dry with air duster Rinse in isopropanol for 3 min Dry with an air duster Place in vacuum oven at room temperature for 30 min to dry Leave to acclimatise for 30 min Weigh to achieve three consecutive readings that agree to within 0.1 mg
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
The influence of nominal stress on wear factors of carbon fibre–reinforced polyetheretherketone (PEEKOPTIMAÒ Wear Performance) against zirconia toughened alumina (BioloxÒdelta ceramic)
Proc IMechE Part H: J Engineering in Medicine 2014, Vol. 228(6) 587–592 Ó IMechE 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0954411914538783 pih.sagepub.com
Andrew Evans1, Henrietta Horton2, Anthony Unsworth3 and Adam Briscoe4
Abstract Carbon fibre–reinforced polyetheretherketone is an attractive alternative to ultra-high-molecular-weight polyethylene in artificial joints, but little has been published on the influence of stress on the wear factor. We know that in ultra-highmolecular-weight polyethylene, the wear factor reduces as the normal stress increases, which is counter-intuitive but very helpful in the case of non-conforming contacts. In this study, carbon fibre–reinforced polyetheretherketone (PEEKOPTIMAÒ Wear Performance) has been investigated in a pin-on-plate machine under steady loads and under stresses typical of hip and knee joints. At stresses below about 6 MPa, wear factors are between 10 and a 100 times lower than for ultra-high-molecular-weight polyethylene but at higher stresses the wear factors increase substantially.
Keywords Wear analysis/testing, bio-composites, biomaterials, tribology of materials
Date received: 13 March 2014; accepted: 14 May 2014
Introduction The demands of patients for replacement joints that allow the resumption of more athletic or more demanding work-based activities have led to the design of hip joints with larger and larger diameters. In order to cope with the potential extra wear rates generated by these activities, new material combinations have been investigated including the use of carbon fibre–reinforced polyetheretherketone (CFR-PEEK) cups and both alumina and zirconia toughened alumina (ZTA) (BioloxÒdelta) femoral head replacements. These have been shown to have extraordinarily low wear rates,1 which are only one-hundredth of a Charnley hip (22 mm diameter metal on ultra-high-molecular-weight polyethylene (UHMWPE)) or one-tenth of highly cross-linked polyethylene (HXLPE) hip joints. CFR-PEEK wear debris is well tolerated in the body,2 and CFR-PEEK components can be moulded or machined to the finished size and shape. Deliberate fibre orientation can reduce the wear even more.3
If CFR-PEEK is to be used more widely to replace UHMWPE or HXLPE, then a detailed knowledge of how the wear rate is affected by the stress level imposed is important to evaluate. For example, UHMWPE has been shown to have very high wear factors at low stresses. These reduce rapidly with increasing stress level4 and while this seems counter-intuitive, it has been confirmed in other laboratories5 though some have failed to observe this phenomenon.6,7 From a design point of view, it is important to know the wear factor characteristics of any proposed material 1
Mott MacDonald, London, UK Bristol University, Bristol, UK 3 School of Engineering and Computing Sciences, Durham University, Durham, UK 4 Invibio Ltd, Thornton-Cleveleys, UK 2
Corresponding author: Anthony Unsworth, School of Engineering and Computing Sciences, Durham University, Durham DH1 3LE, UK. Email: [email protected]
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
588
Proc IMechE Part H: J Engineering in Medicine 228(6)
Figure 1. Diagram of pin-on-plate apparatus. 1: weight; 2: lever arm; 3: drive motor; 4: 12-V DC motor; 5: pin holder; 6: level sensor; 7: heated lubricant bath; 8: CFR-PEEK test pin; 9: ceramic test plate; DC: direct current; CFR-PEEK: carbon fibre–reinforced polyetheretherketone.
combination since these may affect which joint (hip, knee and ankle) or joint design can utilise this material fully. This article describes the use of a pin-on-plate machine to measure the wear factors of CFR-PEEK against ZTA ceramic as stress levels increase.
Apparatus A four-station pin-on-plate machine was used for these experiments (Figure 1). Four rectangular ZTA ceramic plates were fixed into the reciprocating sledge that was heated by embedded thermostatically controlled electrical heaters. Pins made from CFR-PEEK were held in rotating pin holders each of which was rotated by an individual 12-V direct current (DC) motor. The frequencies of the rotation of the pins and reciprocation of the plates were independently variable; however, throughout this work, they were both set to 1 Hz. The pins were subjected to uni-directional rotation while loaded against the plates that were undergoing reciprocation. Loading of the pins on the plates was provided by ‘dead-weights’, which were placed at different lever arm positions to increase or decrease the loads (and stresses) on the specimens. The transmitted loads were measured by a transducer to take account of any friction in the pin holders and lever arm bearings rather than relying on the moment-arm calculations alone. The reciprocating sledge incorporated a lubricating bath containing a liquid, in this case new-born calf serum in de-ionised water. In order to maintain the lubricant level when evaporation was taking place, a system comprising a level sensor, a de-ionised water reservoir and a small electric pump was used to top-up the bath to a set level automatically. The whole machine was covered by a Perspex cover to minimise dust penetration into the lubricant.
Materials and methods CFR-PEEK pins The pins were made from CFR-PEEK (pitch based), 30% fibre fill by weight (PEEK-OPTIMAÒ Wear Performance) supplied by Invibio Ltd, ThorntonCleveleys. Pins with a wearing face of 4.0, 5.0, 6.0, 7.5 and 9.5 mm diameter and total length of 18 mm were used. The 4.0-mm-diameter pins also had a central section of 7 mm diameter to help prevent buckling when carrying the higher loads or bending under combined loading and sliding (Figure 2).
ZTA ceramic plates Biolox delta plates were supplied by CeramTec and measured 48 3 24 3 5 mm. Each had a surface finish of Ra = 0.002 mm. They were aluminium oxide matrix composite ceramic comprising approximately 75% alumina (Al2O3), 24% zirconia (ZrO2) and other trace elements largely to resist crack propagation.
Lubricant The lubricant was de-ionised water with 25% bovine serum added to give a protein content of 18 g/L. Sodium azide (0.2%) and 20 mM ethylenediaminetetraacetic acid (EDTA) were added to reduce bacterial growth and calcium deposition, respectively.
Methods The pin-on-plate machine was thoroughly cleaned and the rotating pin holders lubricated to enable free movement of the rotating pins. A pre-calibrated load transducer was fitted to the reciprocating carriage and the
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
Evans et al.
589 every 300,000 cycles the experiment was halted; the specimens were removed, cleaned, dried and weighed according to the protocol in Appendix 1. Four pins and four plates were actively loaded and subjected to motion during each experiment while a fifth pin and plate acted as a control for fluid absorption. All specimens were inserted into the same lubricant at the same temperature, but the control pins were not subjected to any wearing process. All specimens were cleaned and weighed in exactly the same way using a Mettler-Toledo Balance sensitive to 0.01 mg. A minimum of three measurements were recorded for each sample, and the average weight loss was converted to volumetric wear using the density of CFR-PEEK of 1.42 mg/mm3.8
Figure 2. Pin shape to resist buckling.
Imaging pins were loaded against the transducer to check for any friction in the loading mechanism. This ensured that the load that was applied to the pins was transmitted to the plates. This calibration was used to calculate the actual stresses on the pins. The desired loading was obtained by adding weights at various moment arms on the levers. Table 1 shows the loads, the cross-sectional areas of the wear face of the pins and the contact stresses used in these experiments. The loads were applied and the arms adjusted until they were horizontal, assuring the true load was being applied. As the pins wore, the arms were adjusted to maintain their horizontal position. Fresh lubricant was added to the heated bath and the temperature brought to 37 °C. The frequencies of reciprocation and rotation were both set to 1 Hz and the magnetic cycle counter was set to 0. The stroke length of the reciprocating sledge was set to 25 mm. Lubricant was replenished daily and approximately
The surfaces were examined using both non-contacting white light interferometry (NewView 100; Zygo) and light microscopy. Ten readings were taken for each surface and the worn surfaces were compared with the control surfaces.
Results Figures 3 and 4 show the graphs of volumetric wear against sliding distance for specimens 1–8. These were the higher wear factor specimens obtained at higher stresses. There was little obvious running-in wear, although three specimens, which were from those with lower wear factors (pins 14, 15 and 16; Table 1), did show signs of running-in wear.
Wear factors Table 1 shows the wear factors obtained from all the experiments together with the final roughness of the
Table 1. Stresses, wear factors, roughness and skewness for all the pins. Pin number
Contact stress (MPa)
Steady-state wear factor (31026 mm3/N m)
1 6.85 0.123 2 5.53 0.417 3 5.83 0.064 4 6.85 1.39 5 5.99 1.54 6 8.73 1.10 7 9.36 3.68 8 6.98 2.52 9 0.533 0.0282 10 1.227 0.0311 11 3.346 0.0075 12 1.917 0.035 13 9.746 1.58 14 7.477 0.0167 15 3.346 0.035 16 5.065 0.059 Average initial roughness before the wear tests began
Final roughness, Sq (mm)
Final skewness, Ssk (mm)
Load (N)
Diameter of the pin (mm)
0.130 0.124 0.145 0.110 0.119 0.100 0.196 0.096 0.67 0.48 0.368 0.929 0.419 0.255 0.716 0.468 2.29
22.231 23.39 23.10 20.85 26.61 22.73 28.3 20.45 210.36 210.61 211.06 25.61 27.44 22.84 29.35 27.67 20.455
86.2 69.5 73.2 86.1 75.3 109.8 117.6 87.8 37.8 54.2 65.7 54.2 191.4 146.8 65.7 99.5
4.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 9.5 7.5 5.0 6.0 5.0 5.0 5.0 5.0
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
590
Proc IMechE Part H: J Engineering in Medicine 228(6)
Figure 3. Wear of CFR-PEEK pins (1–4) sliding against ceramic plates.
Figure 5. Pin 5; 5.99 MPa after 2 million cycles.
higher stresses, the wear factors increase dramatically to way above the values for UHMWPE.
Discussion
Figure 4. Wear of CFR-PEEK pins (5–8) sliding against ceramic plates.
pins (Sq) and the final skewness (Ssk) of the surface topography. It is seen that compared with the roughness and skewness prior to starting the wear experiments, the roughness reduced by an order of magnitude while the skewness became much more negatively distributed.
Surface topography Figure 5 illustrates the worn surface of the CFR-PEEK where the cross sections of the carbon fibres can be seen together with some small areas of fibre pull-out when the fibres were lying axially along the surface.
Wear factors against contact stress Figure 6 shows all the wear factors plotted against the contact stress, and it is clear that below about 6 MPa, the wear factors are exceptionally small at about 1028 mm3/N m. This is about two orders of magnitude lower than UHMWPE against stainless steel.9,10 However, at
The first and very striking observation is how very low the wear factors are for CFR-PEEK when the stress levels are small (Figure 6). Indeed, typically, the wear factor at 2 MPa would be 0.035 3 1026 mm3/N m compared with 1.1 3 1026 mm3/N m for UHMWPE against stainless steel. However, at a stress level of 9 MPa, the wear factor can exceed 3 3 1026 mm3/N m. Others have also found this characteristic for CFRPEEK including Wang et al.,11 whose results were consistent with these. This becomes more startling when the results from Vassiliou and Unsworth4 are considered on the same axes (Figure 6). These comprehensive studies of the wear of UHMWPE against stainless steel show the unusual properties of UHMWPE under different levels of stress. At low levels of stress, the wear factors are very high but fall by two orders of magnitude as the stress increases. This is counter-intuitive but is seen to be true in practice. UHMWPE, when used in hip replacements where the stress is relatively low due to the high conformity of the contact, is found to have a wear rate of typically 50 mm3 per million cycles (depending on radius of the joint). On the other hand, most knee joints are less conforming and hence have higher stresses; yet, their wear rate is typically 3 mm3 per million cycles.12 The exception is for conforming knee designs that have wear rates that are twice this level, but still considerably less than in hips. So UHMWPE showed lower wear under high-stress applications such as knee joints, while CFR-PEEK against alumina in large diameter hip applications1 showed extremely low wear rates compared with UHMWPE. In knees, while a useful reduction in wear factor was seen using CFR-PEEK, this was much less of an improvement than in hips.12 While a great deal of consistency was observed between the wear factors of CFR-PEEK below a stress
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
Evans et al.
591
Figure 6. Wear factors of CFR-PEEK against ceramic and UHMWPE against stainless steel.4 UHMWPE: ultra-high-molecular-weight polyethylene.
level of 6 MPa, above this the variation was much greater. As an example in Figure 3, consider pins 1 and 4. Here, the loads were equal on both pins as were the cross-sectional areas of their contact faces; yet, the wear factors were an order of magnitude different. At low contact stresses, pins 10 and 12, for example, even though the diameter changed from 9.5 to 6 mm, the wear factors were almost identical. As with the UHMWPE experiments,4 the CFR-PEEK failed to demonstrate any relationship between wear factors and radius or load though it should be remembered that wear factor itself includes the load, so load is included in both the ordinate and abscissa (stress). There are two other concerns expressed by some surgeons regarding carbon fibre–filled polymers, one of which is whether the carbon will scratch and wear the ceramic counter-face. The other is whether the fibres will dislodge from the matrix and cause irritation of the surrounding soft tissues. The variations in mass of the active ceramic plates are comparable to those of the control plates, which have done no sliding against the PEEK pins. If the maximum wear volumes are calculated for the plates, then these are generally below 0.07 mm3, which translate to wear factors of the order of 3 3 1029 mm3/N m, which are three orders of magnitude lower than the pins. This suggests that this wear has no significance whatever so this worry would seem to be unfounded based on these limited experiments. Turning to the issue of dislodged carbon fibres, Figure 5 indicates that it can happen (see the dark areas) but it seems to be relatively small and associated with fibres that are almost completely worn through with only a small amount of surface area still in contact with the matrix. There are many examples of the fibres wearing gradually at the same rate as the PEEK matrix rather than exhibiting fibre pull-out13 and this was generally seen in this work. Wear of CoCrMo alloy counter-faces has also proved to be of little concern since it is generally two to
three orders of magnitude lower than the wear of the CFR-PEEK components,14 which are themselves much lower than the wear of UHMWPE. It should, however, be stated that we did not systematically study the scratching of the metal counter-faces in our experiments to obtain the wear factors of CFR-PEEK.
Conclusion The wear factors of CFR-PEEK against ZTA ceramic at stresses below about 6 MPa are extremely low and between about one-hundredth and one-tenth of the wear of UHMWPE against stainless steel, depending on the stress level. However, at higher stress, the wear rate was shown to increase dramatically. This suggests that CFR-PEEK against ZTA would be very good in low-stress situations such as hip joints but would be less effective in higher stress applications. However, CFR-PEEK was not seen to be subject to fibre pullout, or to excessive scratching of the ceramic plates. Acknowledgements This work was carried out as part of the SHIELD project and the authors would like to thank TSB for the funding as well as Invibio Ltd and CeramTec for supplying the specimens. Declaration of conflicting interests Dr Adam Briscoe is a paid employee of Invibio Ltd. Funding This work was carried out as part of the TSB SHIELD Project. Research specimens were provided by Invibio Ltd and CeramTec. References 1. Scholes SC, Inman IA, Unsworth A, et al. Tribological assessment of a flexible carbon-fibre-reinforced
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
592
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Proc IMechE Part H: J Engineering in Medicine 228(6) poly(ether-ether-ketone) acetabular cup articulating against an alumina femoral head. Proc IMechE, Part H: J Engineering in Medicine 2008; 222: 273–283. Toth JM. Chapter 7. Biocompatibility of polyaryletheretherketone polymers. In: Kurtz SM (ed.) PEEK biomaterials handbook. Oxford: Elsevier, 2012, pp.81–92. Scholes Susan C, de Jong S and Anthony U. Effect of carbon fibre orientation on the wear of CFR-PEEK/ CFR-PEEK bearing couples. In: Proceedings of 21st annual congress of the International Society for Technology in Arthroplasty, Seoul, Korea, 1–4 October 2008, p.190. Vassiliou K and Unsworth A. Is the wear factor dependent on nominal contact stress in UHMWPE contacts? Proc IMechE, Part H: J Engineering in Medicine 2004; 218: 101–107. Barbour PSM, Barton DC and Fisher J. The influence of contact stress on wear of UHMWPE for total replacement hip prostheses. Wear 1995; 181–183: 250–257. Rostoker W and Galante JO. Contact pressure dependence of wear rates of ultra high molecular weight polyethylene. J Biomed Mater Res 1979; 13(6): 957–964. Rose RM, Goldfarb HV, Ellis E, et al. On the pressure dependence of the wear of ultra high molecular weight polyethylene. Wear 1983; 92(1): 99–111. Invibio Biomaterial Solutions. MOTIS polymer typical material properties, http://www.invibio.com (2009, accessed 23 March 2012). Joyce T, Unsworth A, Cartwright T, et al. Preliminary comparison of the wear of silane XLPE with that of UHMWPE, both polymers rubbing against hard counterfaces (abstract). J Bone Joint Surg Br 81(Supp. 1): 85. Joyce T, Monk D, Scholes SC, et al. A multi-directional wear-screening device and preliminary results of UHMWPE articulating against stainless steel. Bio-Med Mater Eng 2000; 10: 241–249. Wang A, Lin R, Stark C, et al. Suitability and limitations of carbon fiber reinforced PEEK composites as bearing surfaces for total joint replacements. Wear 1999; 225–229: 724–727.
12. Scholes SC and Unsworth A. Pitch-based carbon-fibrereinforced poly (ether-ether-ketone) OPTIMA assessed as a bearing material in a mobile bearing unicondylar knee joint. Proc IMechE, Part H: J Engineering in Medicine 2009; 223: 13–25. 13. Kurtz SM and Nevelos J. Chapter 16. Arthroplasty bearing surfaces. In:Kurtz SM (ed.) PEEK biomaterials handbook. Oxford: Elsevier, 2012, pp.261–276. 14. Scholes SC and Unsworth A. Wear studies on the likely performance of CFR-PEEK/CoCrMo for use as artificial joint bearing materials. J Mater Sci: Mater Med 2009; 20(1): 163–170.
Appendix 1 Cleaning protocol
Rinse in de-ionised water Wipe with lint-free tissue (all surfaces) Place in ultrasonic bath in de-ionised water for 10 min at 18 °C Rinse in distilled de-ionised water Place in ultrasonic bath in 1% NeutraconÒ solution for 10 min Rinse in de-ionised water Place in ultrasonic bath in de-ionised water for 10 min Rinse in distilled de-ionised water (do not wipe) Place in ultrasonic bath in de-ionised water for 3 min Dry with air duster Rinse in isopropanol for 3 min Dry with an air duster Place in vacuum oven at room temperature for 30 min to dry Leave to acclimatise for 30 min Weigh to achieve three consecutive readings that agree to within 0.1 mg
Downloaded from pih.sagepub.com at University of Bath - The Library on June 30, 2015
