Improved wear performance of ultra high molecular weight polyethylene coated with hydrogenated diamond like carbon

1 IMPROVED WEAR PERFORMANCE OF ULTRA HIGH MOLECULAR WEIGHT POLYETHYLENE COATED WITH HYDROGENATED DIAMOND LIKE CARBON 1,2*Puértolas J A; 1Martínez V; 1Martínez M J; 1Mariscal, M D; 2Medel, F J, 3López, C; 3Yubero, F 1Department of Materials Science and Technology, Instituto de Investigaciones en Ingeniería de Aragón, I3A, Universidad de Zaragoza, E-50018 Zaragoza, Spain. 2Department of Materials Science and Technology, Instituto de Ciencia de Materiales de Aragón, ICMA, Universidad de Zaragoza-CSIC, E-50012 Zaragoza, Spain. 3 Instituto de Ciencia de Materiales de Sevilla (CSIC- U. Sevilla). C/Américo Vespucio 49, E- 41092 Sevilla, Spain. *Correspondence to J.A. Puértolas, Full Professor Tel. : 34 976 762521; Fax. : 34976 761957; e-mail: [email protected] *Revised Manuscript Click here to view linked References 2 ABSTRACT Hydrogenated diamond like carbon (DLCH) thin films were deposited on medical grade ultra high molecular weight polyethylene (UHMWPE) by radio frequency plasma enhanced chemical vapour deposition. The DLCH-coating thicknesses ranged from 250 to 700 nm. The substrates were discs made of UHMWPEs typically used for soft components in artificial joints, namely virgin GUR 1050 and highly crosslinked (gamma irradiated in air to 100 kGy) UHMWPEs. Mechanical and tribological properties under bovine serum lubrication at body temperature were assessed on coated and uncoated polyethylenes by means of nanohardness and ball-on-disk tests, respectively. Morphological features of the worn surfaces were obtained by confocal microscopy and scanning electron microscopy. This study confirm an increase in surface hardness and good wear resistance for coated materials after 24 hours of sliding test compared to uncoated polyethylene. These results point out that to coat UHMWPE with DLCH films could be a potential method to reduce backside wear in total hip and knee arthroplasties. 3 INTRODUCTION Ultra high molecular weight polyethylene (UHMWPE) is the most relevant bearing material used in total hip and knee replacements due to its physical, chemical, and mechanical properties, as well as biocompatibility [1-2]. However, UHMWPE has been the weak material of total joint replacements throughout the years, because of wear-related problems [3] and mechanical degradation after long-term post-irradiation oxidation [4]. The former problem stems from the inevitable generation of polyethylene debris particles, which worn off the soft component after repeated cyclic bearing. These particles trigger a cascade of adverse biological reactions in the human body, which may cause the eventual loosening of the component, and therefore, revision of the prosthesis [3, 5]. Wear of articular surfaces has been recognized as the main responsible for the production of polyethylene particles. Particle production from other non-articular surfaces in modular implants has also been observed [6]. This secondary wear source relates to micromobility in modular junctions due to insufficiency of locking mechanisms during service [7]. Although first detected in hip replacements, this problem increasingly affects the knee, with polyethylene particle production estimations for non-articular surfaces being 2 to 100 times the femorotibial wear [8]. Moreover, retrieved knee implants showed non-articular wear 2 to 4 times higher than that of hip implants [9]. Changes in locking mechanisms to limit microdisplacements (from 5 to 311 microns for hip prostheses) and subsequent wear [10] are more difficult at the knee. Other options such as polishing the tibial tray may only be feasible in flat surfaces, but not in metal irregularities [11]. A potential solution for this problem might be to coat non-articular surfaces of polyethylene implants with low friction and wear resistant films. Highly crosslinked polyethylene (HXLPE) exhibits improved wear resistance due to an elevated crosslink density achieved by gamma or electron beam irradiation to doses ranging 50- 100 kGy [12]. However, the irradiation process also originates free radicals [13], whose thermal stabilization produces a negative effect on the fatigue and toughness properties of UHMWPE [14-15]. On the other hand, although HXLPE has shown greatly reduced wear volumes, 4 paradoxically it also exhibits decreased wear debris size, which may have the potential to increase biological activity and, then, to aggravate osteolytic reactions [16-17]. Hydrogenated diamond like carbon (DLCH) coatings have shown good performance regarding hardness, wear resistance, friction, chemical inertness and good biocompatibility, which make them a good option for implants in biomedical applications [18-19]. DLCH consists of amorphous carbon with a significant fraction of C-C sp3 bonds and H content in the 20-40 at. % [20]. In this field, DLCH coatings on metallic materials, which typically form a bearing couple with UHMWPE, have been developed to reduce the wear rate and corrosion of metallic components in prosthetic applications. However, while the second goal is achieved by DLCH coatings, the first one presents a strong controversy taken into account the results obtained in the laboratory and even in clinical trials. Both positive and negative wear behaviors have been reported in the last decade [18] from pin/ball-on-disks, simulator and ¨in vivo¨ studies for 316L, CoCrMo and Ti6Al4V materials as the metallic counterpart of UHMWPE in articulating components. Although some of the discrepancies are attributed to the use of different lubricants as well as the DLCH composition, structure and other experimental conditions, the fact is that some modern knee replacements have been recalled after clinical reports documenting DLCH coating failures. In consequence, new alternatives based on DLCH coated UHMWPE for nonarticulating surfaces must be promoted, maintaining the good performance of UHMWPE in total hip and knee replacements. DLCH coatings on UHMWPE have been prepared by pulsed arc plasma physical vapor deposition (PVD) [21], plasma source ion implantation [22] and plasma enhanced chemical vapor deposition (PECVD) [23-24]. In the latter case, good adhesion and high mechanical performances has been reported. The hypothesis of this work focuses on the use of DLCH coated UHMWPE to decrease the risk of osteolysis in modular implants, reducing the particle formation from non-articular surfaces. To accomplish these goals, pristine and highly cross linked UHMWPE were coated with DLCH layers with different thicknesses and their surface nanohardness and tribological behaviors were analyzed. 5 MATERIALS AND METHODS Raw and irradiated UHMWPE sample preparation The raw material used in this study was a compression-molded sheet of GUR 1050 UHMWPE (Perplas Medical Ltd., Lancashire UK). Discs 3 mm thick and 20 mm in diameter were machined from the compression-molded sheet. Before irradiation, all discs were ground and polished using SiC papers up to a surface roughness equal to 0.15  0.05 microns. Afterwards, they were gamma irradiated in air (I) to a final dose of 100 kGy (Aragogamma, Barcelona, Spain). DLCH coatings DLCH thin films were deposited on pristine and gamma irradiated UHMWPE by PECVD in a RF capacitive coupled reactor with plate parallel electrodes. DLCH layers of 250 and 700 nm were deposited on pristine (V) and irradiated (I) substrates using acetylene, Ar, and H2 gas mixture and 200 V bias voltage. Thus, the different material groups included in this study were referred to as V, V250, V700, I, I250 and I700, To achieve good adhesion between DLCH coatings and polyethylene substrates, an Ar plasma was applied to the UHMWPE samples to activate their surface, and a very thin (less than 20 nm) H rich DLCH accommodation layer was deposited using the same gas mixture and an increasing bias voltage between 100 V and 200 V, prior to the DLCH deposition process. Several techniques were used to characterize the DLCH coatings. Their H content (about 20 at. %) and density (about 1.8 gr/cm3) was evaluated using a combined analysis of X-ray photoelectron and reflection electron energy loss spectroscopies [25]. Other bonding characteristics of the final DLCH layers were evaluated using visible (514 nm) Raman spectroscopy. 6 Nano-hardness profiles. In-depth hardness profiles of uncoated and DLCH coated UHMWPE samples were obtained using a nanoindenter XP MTS system. The maximum loads used in the tests were 0.5 and 2.0 mN, depending on the samples. The depth rate was 5 nm/s. Wear experiments A commercial pin-on-disc tribometer (CSM instruments; Peseux, Switzerland) allowed assessing wear resistance and constantly monitoring the coefficient of friction for all materials. Wear tests were performed in at least three samples per material group working with a rotating vessel, which contained UHMWPE discs immersed in bovine serum (B-9433, Sigma Aldrich). The counterpart was a stationary ball, 6 mm in diameter, made of alumina with a roughness, Ra, of 0.05  0.002 m, whereas the load applied to the ball was 5.23 N. The radius of the circular track was 4 mm and the sliding speed was 0.05 m/s. The environment temperature was set to 37 °C, the body temperature. Assuming hertzian contact, the previous loading conditions gave an approximate contact pressure of 37 MPa, which is in the range of peak contact stresses (30 MPa) found for some contemporary polyethylene tibial inserts [26-27]. Nevertheless, the current experimental conditions did not represent an excessively severe tribological testing, as the global severity parameter proposed by da Silva and Sinatora lain out of the thermal failure regime for our test conditions [28]. Wear rates were measured after 12, 68 and 4400 m sliding distances. After each wear test, disks were removed from the tribometer and underwent a cleaning protocol following guidelines included in ASTM F2025. Confocal microscopy was performed using a SENSOFAR PLU 2300 optical imaging profiler and served to evaluate worn disc surfaces. Four line measurements diametrically opposed were carried out on the wear tracks of each sample. The average of the registered worn areas was used to calculate wear factors, k, according to the following equation: Ls rA km  2  (eq. 1) where r is the wear track radius, m A is the average worn area, L is the applied load, and s is the sliding distance. 7 The volume loss corresponding to the track profile was evaluated taking into account the plastic deformation suffered by the polyethylene discs under applied loads during wear tests. By virtue of the shape memory effect, UHMWPE and HXLPE heated to high enough temperature recover the original shape that has been previously lost due to mechanical deformation [29]. The recovery of plastic deformation after wear testing was achieved by heating samples at 120 °C for 36 hours. Preliminary tests confirmed that some DLCH coated samples demonstrated no removal of the DLCH layer after tribological testing, but they exhibited wear tracks exclusively due to permanent plastic deformation. This permanent plastic deformation was not completely recovered after the thermal treatment. Thus, the area Am introduced in equation (1) was the average worn area after the heating process and then corrected taking into account the nonrecoverable plastic deformation. Student´s t tests (or Wilcoxon tests when appropriate) served to detect significant differences between the friction coefficient and wear factor results of both uncoated and DLCH coated UHMWPE samples. P values less than 0.05 were considered indicative of significance. Microscopy observation Environmental scanning electron microscopy (SEM) equipment (HITACHI S-3400 N) was used to observe the morphology of the wear tracks in all the samples. RESULTS AND DISCUSSION Characteristics of the DLCH coatings Figure 1 shows the typical Raman spectrum registered for DLCH coatings. It showed the characteristic G peak at ~1530 cm-1 and a D peak at ~1328 cm-1, with a A(D)/A(G) intensity ratio of 0.63. This result is consistent with a DLCH with 30-40 % of C-C sp3 bonds, and H content of about 20 at. % and density of about 1.8 gr/cm3 [20]. Nano-hardness Figure 2 shows the nano-hardness profiles obtained for uncoated and DLCH coated UHMWPE samples. Qualitatively, a similar behavior was observed regardless of the substrate 8 materials, i.e. virgin or irradiated polyethylene, although nano-hardness in the latter was slightly higher than that of specimens corresponding to virgin substrates. Nano-hardness tests of uncoated samples showed a start-up value of ~1 GPa, which rapidly decreased until a constant value was reached. In virgin samples this value was 0.01 GPa and was reached at a depth of 60 nm, whereas in crosslinked polyethylene the value was 0.04 GPa and reached at 150 nm. These small hardness values are consistent with those obtained for UHMWPE commonly utilized for acetabular liners. Other general property deduced from nanoindentation profiles was that the overall hardness significantly rose with the thickness of the coatings. The V250 and I250 samples presented similar nano-hardness profiles with an initial decrease that finished in a well, which was, in turn, followed by a wide shoulder. In both coated materials the initial hardness was close to 2 GPa, the minimum appeared at 15 nm and the shoulder around 40 nm, although the well was more pronounced in V250 samples than in the I250 ones. From 300 nm in depth, hardness reached a steady value of 0.1 GPa, which was higher than the one displayed by uncoated samples. In samples with the thickest DLCH coatings, V700 and I700, the well disappeared, and after an initial and strong fall, the hardness followed a smooth decrease up to a value of ~0.3 GPa at 400 nm in depth. The present findings confirmed the strong influence of the substrate material, particularly when the coating and substrate stiffnesses are very different, as happens between DLCH and polyethylene. On the other hand, it is usual that the influence of the substrate becomes relevant at indentation depths higher than 15 % of the coating thickness, which coincides with the depth where the V250 and I250 samples underwent the shoulder. The other part of the profile where there was a significant improvement in hardness was the uppermost surface. However, this observed hardening was lower than the one obtained by Poliakov and colleagues for DLCH films with 250 nm thickness deposited on UHMWPE [21], although its shape profiles were closer to the ones depicted in our V700 and I700 samples. Probably, the reason for the relative softening is the lower sp3 content of the DLCH grown with PECVD compared with that obtained by pulse arc plasma PVD. 9 Friction The dependence of the friction coefficient f versus sliding time obtained during ball-on- disk testing is plotted in Figures 3a and 3b for the six materials. Uncoated virgin (V) and crosslinked (I) polyethylene samples exhibited the lowest friction coefficients at the beginning of the test, 0.10 and 0.12, respectively (Wilcoxon test; p ≤ 0.03 with respect to DLCH coated materials). Later on, the friction behavior remained practically constant for uncoated specimens during all the sliding distance. DLCH coated materials with a 700 nm layer displayed the highest friction coefficients (close to 0.20), which remained practically constant throughout the whole sliding distance. This friction coefficient was higher than those reported previously for DLCH coatings on steel substrates sliding against alumina counterpart in normal atmosphere, which ranged between 0.06 and 0.15 depending on the applied load and sliding velocity conditions [30-31]. The dynamic friction coefficient spectra of 250 nm coated materials revealed a different tribological behavior. Specimens with the thinnest DLCH layers started with a friction coefficient (~ 0.20) similar to that of 700 nm DLCH coated materials (Wilcoxon test; p ≥ 0.4). Then, the initial friction coefficient of 250 nm DLCH coated materials progressively reduced down to about 0.11 (the steady friction coefficient of uncoated materials), especially in the case of irradiated substrates (p = 0.24; I250 compared with I materials at the end of wear testing). These findings point out that the DLCH layer became unstable during sliding in a process that finished when the ball reached the substrate, and therefore the friction coefficient reflected that of the polyethylene/alumina bearing couple. Previous long-term tribological tests have also shown sudden jumps in the friction coefficient [30], although the comparison with the present study is not fully appropriate since the initial surface roughness of our polyethylene specimens was higher than that of the metallic substrates used in previous studies for DLCH coating. Other variables like liquid lubrication, temperature and DLCH composition could also affect the wear life of the coating. Wear Confocal microscopy images of the wear tracks were obtained for the six materials after 4400 m of sliding distance (Figure 4a-f). Uncoated virgin and crosslinked UHMWPE samples exhibited wear tracks with grooves parallel to the sliding direction, which are typical of an 16 FIGURE 1 17 FIGURE 2 18 FIGURES 3A-B 19 FIGURES 4A-F 20 FIGURES 5A-C 21 FIGURES 6A-F 22 FIGURES 7A-B 23 Table 1. Wear rate, k, of uncoated and DLCH coated materials after different sliding distances. WEAR RATE, k *10-6 (mm3 / N∙m) LENGTH(m) 12 68 4400 SAMPLE V 23±32 18.5±7 2.4 ± 0.1 V250 2.6 ± 0.5 V700 I 72±26 12.8±8.2 3.5 ± 0.2 I250 3.0 ± 0.2 I700 REVIEWER 1) NAOHIDE TOMITA Graduate School of Engineering, Kyoto University [email protected] He is working in tribology related to UHMWPE for joint prostheses 2) J. L. TIPPER Institute of Medical & Biological Engineering, University of Leeds UK [email protected] She is a relevant researcher in wear behaviour of bearing materials for total joint replacements. *Suggested Reviewers