J Mater Sci: Mater Med DOI 10.1007/s10856-014-5252-y
Treating orthopedic prosthesis with diamond-like carbon: minimizing debris in Ti6Al4V Luciane Y. S. Oliveira • Neide K. Kuromoto Carlos J. M. Siqueira
•
Received: 13 December 2013 / Accepted: 3 June 2014 Springer Science+Business Media New York 2014
Abstract Prostheses are subject to various forms of failing mechanisms, including wear from ordinary patient motion. Superficial treatments can improve tribological properties of the contact pair, minimizing wear and increasing prostheses lifetime. One possibility is the diamond-like carbon (DLC) coating, where Carbon is deposited with variable ratio of sp2/sp3 structures, leading to an increase in surface hardness. So in this research Ti6Al4V samples were coated with DLC using sputtering process to evaluate the debris release. The Ti6Al4V and Ti6Al4V plus DLC coating surfaces were analyzed using Raman spectroscopy and instrumented indentation (hardness). The wear behavior was tested using a reciprocating linear tribometer. The wear rate was smaller in the coated samples, producing less debris than the untreated Ti6Al4V alloy. Debris morphology was also evaluated, using scanning electronic microscopy, and it was observed that debris size from the coated samples were bigger than those observed from the uncoated Ti6Al4V alloy, above the size that generally triggers biological response from the host.
1 Introduction Human body can move due to a complex system of muscles and articulations. Some degeneration, like osteoarthritis, can reduce the function and quality of life of the patient, being the most common disease of the adult hip [1]. In some cases, the total hip arthroplasty procedure can improve the pain and
recover hip function, giving the patient a better quality of life [2]. Prosthesis is an artificial component implanted to substitute some body function. The lifetime of prostheses vary according to many factors: what kind of body function is being replaced, lifestyle of the patient, physiological characteristics, prostheses model and materials, fixation method etc. When prosthesis reaches its end-of-life a replacement surgery is required and the patient is subject to all its inherent risks. Lifetime is increased when excessive wear and corrosion are reduced [3]. Many researches are made to increase biomechanical properties of prostheses for articular replacement. Among all possible mechanical treatments, one possibility is the surface modification by deposition of hard coatings on titanium alloys. Diamond-like carbon (DLC) is a coating that has low coefficient of friction, high hardness, low wear rate and chemical stability [4, 5]. These characteristics allow an increase in prostheses lifetime avoiding body rejection from the adjacent host tissue [6]. The increase of wear resistance leads to reduction of the amount of debris, an important characteristic in articular prostheses [7, 8]. In the present research we investigated the debris released using tribological tests in titanium alloy (Ti6Al4V) coated with DLC deposited by sputtering method. Morphological characterization of debris was performed in order to evaluate its medium size and roundness.
2 Materials and methods L. Y. S. Oliveira (&) N. K. Kuromoto C. J. M. Siqueira Programa de Pos-Graduac¸a˜o em Engenharia Mecanica (PG-MEC), Universidade Federal do Parana´ (UFPR), Rua Francisco H. dos Santos, P.O. Box 19011, Curitiba, PR, Brazil e-mail: [email protected]
2.1 Materials Samples were prepared with biomedical Ti6Al4V alloy (ASTM F136). The chemical composition was assured by
123
J Mater Sci: Mater Med
X-ray spectrometry and chemical gas composition, presenting the following elements (wt%): 0.006 C; 0.0037 H; 0.1151 O; 0.0066 N; 6.13 Al; 0.18 Fe; 3.97 V and balance Ti. Ti6Al4V rods, with 15.88 mm diameter, supplied by Vulcanium Technology, were cut into discs with 5 mm thickness. The discs were grounded and polished to obtain mirror-like smoothness. The grinding was done in 5 stages. First step grinding was done with 120 grit silicon carbide abrasive paper, followed by 220, 320, 500, 1,200 and 2,400 grit abrasive papers. Subsequently, the grounded surfaces were polished with coarse and fine colloidal silica solution to get mirror-like smoothness. The polished samples were cleaned with aqueous acetone solution in an ultrasonic bath tank and air dried. The average surface roughness (Ra) values of the sample were maintained at (0.04 ± 0.013) lm, measured with a profilometer, in order to achieve a roughness comparable to that of orthopedic devices surface. After the metallographic preparation the Ti6Al4V samples were submitted to diamond-like carbon deposition. 2.2 DLC deposition The DLC deposition was realized in Bodycote Brasimet using the sputtering method. Before the coating deposition, Ti6Al4V samples were cleaned ultrasonically in acetone. For the deposition of the DLC a direct current discharge was used in a mixed argon/acetylene/nitrogen atmosphere. The following parameters were used: • • • • • • •
Sputtering process; Temperature: 180–200 C; Coating structure: Multilayer; Bias voltage (interlayers): -400 V; Solid sources: carbon and Ti cathodes; Sputtering power: 5 kW; Total deposition time: 10–15 h.
With these parameters the expected layer thickness of 2–3 lm and a hardness of 4–4.5 GPa were obtained. 2.3 Raman spectroscopy Raman spectroscopy was used to evaluate chemical and structural modifications generated during the DLC deposition. Witec Focus Innovations equipment was used with the following parameters: • • • •
Standard: silicon; Integration: 5 s; Accumulations: 109; Laser: Nd:YAG laser, 532 nm;
123
Raman spectra were measured in a range of 100–2,000 cm-1 at an operating power of 25 mW. 2.4 Hardness The nanohardness (H) values were determined using a MTS Nano Indenter XP equipment with a diamond Berkovich geometry tip. The maximum applied load was 400 mN and 9 indentations were made on each sample with a distance of 100 lm. The nanohardness was calculated from 8 loading–unloading cycles using Oliver and Pharr method [9]. 2.5 Reciprocating linear tribometer This test was performed to evaluate the wear rate of the samples, with and without the DLC coating. A reciprocating linear tribometer from CSM Instruments was used. The reciprocating frequency was 2 Hz, the same used by ¨ rstele et al [10]. O For the Ti6Al4V samples, a load of 5 N was applied. For the coated samples the load was increased to 10 N, since 5 N was insufficient to produce debris in sufficient quantity to microscopical analysis. Other parameters used in this test were: • • • •
Number of cycles: 5,000; Tip: round Tungsten Carbide (WC), 6 mm diameter; Frequency: 2 Hz; Amplitude: 4 mm.
The wear track section was obtained using a profilometer (Veeco Dektak 150). The transversal section area and the wear rate were calculated using the Tribox Software (v 2.0), from CSM Instruments. 2.6 Collecting debris After the tribological test, samples were deposited in an aluminum recipient (30 mm diameter and 20 mm height). Inside this recipient, samples were washed out with acetone. This set was submitted to ultrasonic cleaning for 50 min. The sample was removed from the recipient and the recipient with acetone containing the suspended particles was taken to an oven to evaporate. With the complete evaporation of the acetone, debris was left in the bottom of the recipient, allowing further tests. 2.7 Scanning electronic microscopy The collected debris was analyzed using scanning electronic microscopy (SEM). The SEM analyses were performed in Philips XL30 equipment. Semi-quantitative
J Mater Sci: Mater Med
elementary analyses were also done, using EDAX X-ray dispersive spectroscope. The parameters were: • •
Magnification: 1009, 2009, 5009 and 10009 Acceleration voltage: 15 and 20 kV
Those parameters were adjusted according to the size of each fragment.
The DLC diamond peak begins at 1,340 cm-1, and it’s known that the natural, unstressed diamond has a single sharp peak at 1,332 cm-1 [6]. There’s a second, broader peak at 1,580 cm-1, corresponding to the G band (graphite-like film). The intensity of the sp2 peak is slightly bigger than the sp3 peak (approximately 2 % bigger), indicating that the deposited film has almost the same amount of diamond and graphite [6].
2.8 Debris analysis through images 3.2 Hardness From SEM, debris was analyzed regarding their area, length and Feret ratio. Measurements were performed using the software from the Carl Zeiss Axionvision optical microscope. The area measure is obtained from the visible area of the debris, captured from the image. The Feret ratio is the result of dividing the largest Feret diameter by the smallest Feret diameter. The largest Feret diameter is the largest distance among two points in the debris boundary. The smallest Feret diameter is the smallest distance among two points in the debris boundary that is perpendicular to the largest Feret diameter. Feret ratio will always be a value in the range 0–1. Values closer to 1 indicate that the debris is rounder, while Feret ratios closer to 0 will indicate that its shape is oblong.
The hardness values obtained with instrumented indentation are shown in Fig. 3. It’s possible to observe that the hardness is increased just for a small displacement, indicating that the film was thin (2–3 lm, as expected). Given that the graphite corresponded to almost 50 % of the film content, the hardness of
3 Results 3.1 Raman analysis Raman spectroscopy was performed in Ti6Al4V and Ti6Al4V ? DLC samples. In Fig. 1 it is possible to observe the peaks that correspond to the Ti6Al4V alloy. In Fig. 2 it is possible to observe the Raman spectroscopy for the Ti6Al4V ? DLC sample, specifically the shift region of the DLC coating.
Fig. 1 Raman spectroscopy of the Ti6Al4V sample
Fig. 2 Raman spectroscopy of the Ti6Al4V ? DLC sample
Fig. 3 Hardness results of Ti6Al4V and Ti6Al4V ? DLC samples surfaces
123
J Mater Sci: Mater Med
average wear rate was of 12 9 10-6 mm3/N/m. The DLC deposition increased the wear resistance in about 35 times. The high sp2 content lead to a lubricating property with very small wear being observed in the WC counterpart. 3.4 Coefficient of friction The coefficient of friction (COF) values are shown in Fig. 5. The DLC coated sample had a smaller COF than the Ti6Al4V sample. This corroborates the lubricating property of the high graphitic film, with the COF being steady as the number of cycles increased. On the uncoated sample, the COF increases with the number of cycles, as debris from the wear accumulate. 3.5 Scanning electron microscope of the cross-section SEM analysis were performed in the cross-section of the samples, as shown if Fig. 6. The DLC coating has a width of 2.5 lm, within the manufacturer specification tolerance. Small differences in the width and some dark spots are a consequence of deposition parameters (such as small temperature variation) and sample preparation to the SEM imaging. 3.6 Scanning electron microscope and energydispersive X-ray spectroscopy of debris
Fig. 4 Wear track of a Ti6Al4V and b Ti6Al4V? DLC samples
4.3 GPa was equivalent of a soft a-C:H amorphous carbon composition. 3.3 Wear track and wear rate The wear track for the Ti6Al4V and Ti6Al4V ? DLC samples are shown in Fig. 4. It is possible to observe that the wear track for the coated sample is smaller than the Ti6Al4V alloy, with approximately 33 % of its width (approximately 250 lm against 750 lm for the uncoated sample). As mentioned earlier, tribological test had a different load for each sample, with a load of 5 N on the Ti6Al4V uncoated sample and a load of 10 N on the DLC coated sample. Using a smaller load on the Ti6Al4V ? DLC sample barely produced any wear, preventing the observation of debris. And a higher load (10 N) on the Ti6Al4V sample lead to an overrun of the equipment limits, causing excessive tangential force. The wear rate for the Ti6Al4V sample was on average 437 9 10-6 mm3/N/m. For the DLC coated samples, the
123
SEM analysis was performed on debris collected from the tribological test. Energy-dispersive X-ray spectroscopy (EDS) was also performed in some debris, evaluating if the chemical composition was correspondent to the expected from the sample substrate, and coating or even from WC spheres. In Fig. 7 it is possible to observe some debris from the Ti6Al4V sample, with a magnification of 5009. The EDS analysis was also performed, showing the expected composition of Ti, Al and V. It’s possible to observe that the debris size distribution is very irregular, with some particles having a length of 1 lm or less. The shape of the larger particles is also very irregular, including agglomerations of particles with many sizes in multiple layers (for example, in the lower right of Fig. 7b). In Fig. 8 debris from the Ti6Al4V ? DLC samples are shown, with a magnification of 5009. EDS results showed the presence of C, O, Si and W, being Carbon and Silicon elements from the DLC coating: C is the main element of the film deposition and Si is used in a pre-deposition step, to increase the adherence of the DLC film to the substrate. Tungsten comes from the WC sphere and just small traces were found on the EDS test.
J Mater Sci: Mater Med Fig. 5 Coefficient of friction of Ti6Al4V and Ti6Al4V ? DLC samples
4 Discussion In the Raman spectroscopy of the Ti6Al4V ? DLC samples it’s possible to observe that the DLC coating structure is responsible for two characteristic peaks at the 1,332 and 1,580 cm-1 region, being respectively the D (diamond— sp3) and G (graphite—sp2) band [11]. The D and G bands can be fitted in the Raman spectra using two Gaussian curves, and their ratio can be used to estimate the sp2/sp3 ratio. Our DLC coating had high presence of sp2 structure, corroborating the graphitic properties found in the tests [12, 13]. It was observed that the DLC deposition increases the hardness in approximately 15 %. After 1,000 lm the values of both surfaces are the same. There is influence of the substrate in the coating hardness. According to Tsui and Pharr [14] when the depth of the nanoindentation is bigger than 10 % of the coating thickness, results are influenced by the substrate. Since the DLC coating had a thickness of 2,500 nm, all measures with a displacement over 250 nm have influence from the Ti6Al4V hardness. The observed DLC hardness is smaller from the theoretical hardness values due to the film thickness, substrate hardness, sp2 to sp3 ratio, and other factors [15]. The deposition of the DLC coating was sufficient to increase the resistance against the sliding wear caused by the linear reciprocal movement. The wear track edge of the TI6Al4V ? DLC sample is also smoother, without plucked chips, indicating that the wear mechanism probably wasn’t related to crack propagation.
The DLC deposition parameters lead to a higher content of sp2 hybridizations. According to Wen et al. [16] DLC coatings with higher sp2 content have more evident graphitic properties, lowering the COF. It’s possible to observe that as the number of cycles increases, the Ti6Al4V COF also increases, probably due to the accumulation of debris on the sample surface. The self-lubrication characteristic of the graphitic DLC coating helps the steady COF, independently of the number of cycles. Debris from the Ti6Al4V sample had a size, on average, smaller than 20 lm, and in a great amount, sometimes forming small clusters with an elongated irregular shape. According to Sargeant and Goswami, debris with that size promotes an increased response in macrophages, leading to an intense biological host response [17]. Hallab and Jacobs assert that elongated debris, smaller than 10 lm, stimulate more inflammatory host response, activating phagocytosis from macrophages. Wright and Goodman [18] state that debris with size smaller than 20 lm stimulates the release of the tumor necrosis factor (TNF) cytotoxin. Much debris from the Ti6Al4V ? DLC samples had a length bigger than 60 lm. Very few debris was produced from those samples, and they were isolated on the recipient, allowing individual images of 46 particles. Unlike the debris from the uncoated Ti6Al4V sample, this debris does not resemble a cluster. The majority seems to be made of one single fragment, and some seem to be made of two parts. This debris geometry indicates elongated plates, probably generated by a delamination from the
123
J Mater Sci: Mater Med
Fig. 6 Cross-section of the a Ti6Al4V and b Ti6Al4V ? DLC samples
Fig. 7 Debris from the tribological test on the Ti6Al4V sample, 9500 magnification
DLC coating, facilitated by the sp2 structure. It is expected that these debris produce less intense immunological host response, which is desirable in long-life prostheses [19, 20]. It was possible to calculate the area and Feret ratio and measure the length of the particles. These results are shown in Table 1. The particles individualized by the SEM software had some parameters calculated. The average particles area was 105.9 lm2 for the Ti6Al4V samples and 4,477.8 lm2 for the DLC coated samples. It’s worth mentioning that some debris from the Ti6Al4V samples were clustered and could not be analyzed individually, and so the average area of each particle was probably much smaller. The length (maximum distance among two points on the edge of the particle) was measured, and the average particle length was 12.5 lm for the Ti6Al4V samples and 107.40 lm for the Ti6Al4V ? DLC samples. The standard
deviation of the measure was proportionally bigger for the titanium-alloy samples, since some clusters were measured as a single particle. The Feret ratio was also calculated. A Feret ratio closer to 1 would indicate a round shape. Values closer to 0.5 indicate that the width of the debris is, roughly, half of the length. A small difference was found with Ti6Al4V debris being slightly bigger. In Fig. 9 the debris length distribution is shown for the Ti6Al4V samples. It’s possible to observe that almost half of the particles have a size of 5 lm or less. The 90th percentile was calculated as 25.3 lm. In Fig. 10 the debris length distribution is shown for the Ti6Al4V ? DLC samples. It’s possible to observe that\10 % of the particles had a length of 40 lm or less. The 90th percentile was calculated as 200 lm.
123
J Mater Sci: Mater Med Table 1 Area, diameter and Feret Ti6Al4V ? DLC samples debris
These results can be compared with other studies, where debris from hip prosthesis made from different materials were analyzed. Commercially, hip prosthesis
of
Ti6Al4V
and
Ti6Al4V
Ti6Al4V ? DLC
Average area (lm2)
105.9
4,477.8
Average length (lm)
12.5
107.4
Average Feret ratio
Fig. 8 Debris from the tribological test on the Ti6Al4V ? DLC sample, 9500 magnification
ratio
0.63
0.58
are commonly made from a metallic (Cobalt Chrome or Stainless Steel) or ceramic (alumina–zirconia) femoral head which articulates on a polymeric acetabular cup (UHMWPE or PUC) [21–23]. Tipper et al. [21] characterized debris from UHMWPE against zirconia ceramic and found that approximately 85 % of the debris had 0.1–0.5 lm. They also tested a metal-on-metal pair (low carbon wrought cobalt chrome alloy against high carbon wrought cobalt chromium alloy) and found debris with an average length of 30 nm. For a ceramic-on-ceramic pair (alumina) the debris was even smaller, with an average length of 9 nm. Affatato et al. [22] tested UHMWPE against ceramics heads and found debris that had, mainly, a length of 200–300 lm for a 20 % alumina—80 % zirconia composition; 10–15 lm for a 40 % alumina–60 % zirconia composition and up to 100 lm for a pure alumina ceramic. Wu and Peng [24] studied debris from UHMWPE against high carbon cobalt chromium alloy and found that approximately 22.4 % of them were smaller than 0.1 lm, 71.6 % were in a size range of 0.1–10 lm and 6 % were bigger than 10 lm. Wang et al. [25] tested UHMWPE against cobalt chromium alloy and found two types of UHMWPE debris: a rounded type, which was smaller than 0.3 lm, and an elongated type, with most debris having a length of 1–2 lm. Debris size can be affected by testing conditions, like the use of a hip simulator, applied load and lubricating conditions. Nevertheless, debris generated from the DLCcoated Ti6Al4V has a size comparable to those obtained with UHMWPE against ceramic pairs (alumina or alumina–zirconia). Comparing particles sizes from DLC-coated and uncoated samples, it’s possible to observe that the DLC coating increased the average particle size and (based on the wear rate and difficulties in finding those debris using SEM) reduced its amounts. As mentioned earlier, the literature [16–20] considers that the immunological response from the host is more intense for particles with smaller sizes, including a greater release of the tumor necrosis factor cytotoxin.
123
J Mater Sci: Mater Med Fig. 9 Debris length distribution and cumulative curves for the Ti6Al4V samples
Fig. 10 Debris length distribution and cumulative curves for the Ti6Al4V ? DLC samples
5 Conclusion Increasing the lifetime of orthopedic prostheses would benefit patients, as less replacement surgeries would be required. Hard coatings represent one possibility, since an increase in tribological properties of prostheses materials could minimize its wear due to sliding forces that are applied during everyday movements of the patient. DLC coating have desirable properties: the sp2/sp3 ratio determines the hardness and self-lubricating characteristic, and the chemical composition and structure are biological inert, avoiding the host biological response. In this research Ti alloys used as implant were coated with DLC layer with the goal to obtain surface that can
123
increase the lifetime of orthopedic prostheses. The results indicate that the wear rate decreases when the Ti6Al4V sample is coated with DLC, producing less debris. Debris morphology also changes, with Ti6Al4V ? DLC leading to bigger particles, a factor that reduces phagocytosis and the deployment of tumor necrosis factor. DLC coating might be a viable superficial treatment for Ti6Al4V prostheses that have some associated movement, like hip prostheses. Acknowledgments The authors would like to thanks Dr. Ronaldo Ruas, from Brasimet Bodycote, for the DLC deposition; Irineu Vitor Leite and Dr. Geninho Thome´, from Neoortho Produtos Ortope´dicos S/A, for the Ti6Al4V samples; LabNano and CME from UFPR for hardness tests and Raman spectroscopy.
J Mater Sci: Mater Med
References 1. Hartofilakidis G, Babis GC, Lampropoulou-Adamidou K. Congenital hip disease in adults. Berlin: Springer; 2014. p. 3–10. 2. Singh J, Lewallen DG. Patients with osteoarthritis and avascular necrosis have better functional outcomes and those with avascular necrosis worse pain outcomes compared to rheumatoid arthritis after primary hip arthroplasty: a cohort study. BMC Med. 2013;11:210. 3. Huang L, et al. Nanomechanical properties of nanostructured titanium prepared by SMAT. Surf Coat Technol. 2006;201:208–13. 4. Marciano FR, et al. Thermodynamic aspects of fibroblastic spreading on diamond-like carbon films containing titanium dioxide nanoparticles. Theor Chem Acc. 2011;10:24. 5. Wu Y, et al. Preparation and properties of Ag/DLC nanocomposite films fabricated by unbalanced magnetron sputtering. Appl Surf Sci. 2013;284:165–70. 6. Nibennanoune Z, et al. Improving diamond coating on Ti6Al4V substrate using a diamond like carbon interlayer: raman residual stress evaluation and AFM analyses. Diam Relat Mater. 2012;22:105–12. 7. Reuter S, et al. Correlation of structural properties of commercial DLC coatings to their tribological performance in biomedical applications. Wear. 2006;261:419–25. 8. Brossa F, et al. Tribological behavior of Ti6Ai4V modified by surface treatments. J Mater Sci Mater Med. 1996;7:471–4. 9. Oliver WC, Pharr GM. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J Mater Res. 2011;7:1564–83. ¨ sterle W, et al. Potential of wear resistant coatings on Ti–6Al–4V 10. O for artificial hip joint bearing surfaces. Wear. 2008;264:505–17. 11. Schwan J, et al. Raman spectroscopy on amorphous carbon films. J Appl Phys. 1996;80:440. 12. Yadav VS, et al. Study of Raman spectra of nano-crystalline diamond like carbon (DLC) films composition (sp2:sp3) with
13.
14.
15. 16. 17. 18.
19. 20. 21.
22.
23.
24. 25.
substrate temperature. Proc World Congr Eng Comput Sci. 2009;1:4. Tai FC, et al. Correlation between ID/IG Ratio from visible Raman spectra and sp2/sp3 ratio from XPS spectra of annealed hydrogenated DLC film. Mater Trans. 2006;47(7):1847–52. Tsui TY, Pharr GM. substrate effects on nanoindentation mechanical property measurements of soft films on hard substrates. J Mater Res. 1999;14:292–301. Fischer-Cripps AC. Nanoindentaion. New York: Springer; 2004. Wen, et al. Surface evolution of a gradient structured Ti in hydrogen peroxide solution. Appl Surf Sci. 2008;254:2905–10. Sargeant A, Goswami T. Hip implants: paper V. Physiological effects. Mater Des. 2006;27:287–307. Wright TM, Goodman SB. Implant wear in total joint replacement: clinical and biological issues, material and design considerations. American Academy of Orthopaedic Surgeons; 2002. Hallab NJ, Jacobs JJ. Biologic effects of implant debris. Bull NYU Hosp Jt Dis. 2009;67:182–8. Hunt JA, Willians DF. The effect of titanium debris on soft tissue response. J Mater Sci Mater Med. 1994;5:381–3. Tipper JL, et al. Characterization of wear debris from UHMWPE on zirconia ceramic, metal-on-metal and alumina ceramic-onceramic hip prostheses generated in a physiological anatomical hip joint simulator. Wear. 2001;250:120–8. Affatato S, et al. Isolation and morphological characterization of UHMWPE wear debris generated in vitro. Biomaterials. 2001;22:2325–31. Elsner JJ, et al. Wear rate evaluation of a novel polycarbonateurethane cushion form bearing for artificial hip joints. Acta Biomater. 2010;6:4698–707. Wu J, Peng Z. Investigation of the geometries and surface topographies of UHMWPE wear particles. Tribol Int. 2013;66:208–18. Wang A, et al. Comparison of the size and morphology of UHMWPE wear debris produced by a hip joint simulator under serum and water lubricated conditions. Biomaterials. 1996;17:865–71.
123
Treating orthopedic prosthesis with diamond-like carbon: minimizing debris in Ti6Al4V Luciane Y. S. Oliveira • Neide K. Kuromoto Carlos J. M. Siqueira
•
Received: 13 December 2013 / Accepted: 3 June 2014 Springer Science+Business Media New York 2014
Abstract Prostheses are subject to various forms of failing mechanisms, including wear from ordinary patient motion. Superficial treatments can improve tribological properties of the contact pair, minimizing wear and increasing prostheses lifetime. One possibility is the diamond-like carbon (DLC) coating, where Carbon is deposited with variable ratio of sp2/sp3 structures, leading to an increase in surface hardness. So in this research Ti6Al4V samples were coated with DLC using sputtering process to evaluate the debris release. The Ti6Al4V and Ti6Al4V plus DLC coating surfaces were analyzed using Raman spectroscopy and instrumented indentation (hardness). The wear behavior was tested using a reciprocating linear tribometer. The wear rate was smaller in the coated samples, producing less debris than the untreated Ti6Al4V alloy. Debris morphology was also evaluated, using scanning electronic microscopy, and it was observed that debris size from the coated samples were bigger than those observed from the uncoated Ti6Al4V alloy, above the size that generally triggers biological response from the host.
1 Introduction Human body can move due to a complex system of muscles and articulations. Some degeneration, like osteoarthritis, can reduce the function and quality of life of the patient, being the most common disease of the adult hip [1]. In some cases, the total hip arthroplasty procedure can improve the pain and
recover hip function, giving the patient a better quality of life [2]. Prosthesis is an artificial component implanted to substitute some body function. The lifetime of prostheses vary according to many factors: what kind of body function is being replaced, lifestyle of the patient, physiological characteristics, prostheses model and materials, fixation method etc. When prosthesis reaches its end-of-life a replacement surgery is required and the patient is subject to all its inherent risks. Lifetime is increased when excessive wear and corrosion are reduced [3]. Many researches are made to increase biomechanical properties of prostheses for articular replacement. Among all possible mechanical treatments, one possibility is the surface modification by deposition of hard coatings on titanium alloys. Diamond-like carbon (DLC) is a coating that has low coefficient of friction, high hardness, low wear rate and chemical stability [4, 5]. These characteristics allow an increase in prostheses lifetime avoiding body rejection from the adjacent host tissue [6]. The increase of wear resistance leads to reduction of the amount of debris, an important characteristic in articular prostheses [7, 8]. In the present research we investigated the debris released using tribological tests in titanium alloy (Ti6Al4V) coated with DLC deposited by sputtering method. Morphological characterization of debris was performed in order to evaluate its medium size and roundness.
2 Materials and methods L. Y. S. Oliveira (&) N. K. Kuromoto C. J. M. Siqueira Programa de Pos-Graduac¸a˜o em Engenharia Mecanica (PG-MEC), Universidade Federal do Parana´ (UFPR), Rua Francisco H. dos Santos, P.O. Box 19011, Curitiba, PR, Brazil e-mail: [email protected]
2.1 Materials Samples were prepared with biomedical Ti6Al4V alloy (ASTM F136). The chemical composition was assured by
123
J Mater Sci: Mater Med
X-ray spectrometry and chemical gas composition, presenting the following elements (wt%): 0.006 C; 0.0037 H; 0.1151 O; 0.0066 N; 6.13 Al; 0.18 Fe; 3.97 V and balance Ti. Ti6Al4V rods, with 15.88 mm diameter, supplied by Vulcanium Technology, were cut into discs with 5 mm thickness. The discs were grounded and polished to obtain mirror-like smoothness. The grinding was done in 5 stages. First step grinding was done with 120 grit silicon carbide abrasive paper, followed by 220, 320, 500, 1,200 and 2,400 grit abrasive papers. Subsequently, the grounded surfaces were polished with coarse and fine colloidal silica solution to get mirror-like smoothness. The polished samples were cleaned with aqueous acetone solution in an ultrasonic bath tank and air dried. The average surface roughness (Ra) values of the sample were maintained at (0.04 ± 0.013) lm, measured with a profilometer, in order to achieve a roughness comparable to that of orthopedic devices surface. After the metallographic preparation the Ti6Al4V samples were submitted to diamond-like carbon deposition. 2.2 DLC deposition The DLC deposition was realized in Bodycote Brasimet using the sputtering method. Before the coating deposition, Ti6Al4V samples were cleaned ultrasonically in acetone. For the deposition of the DLC a direct current discharge was used in a mixed argon/acetylene/nitrogen atmosphere. The following parameters were used: • • • • • • •
Sputtering process; Temperature: 180–200 C; Coating structure: Multilayer; Bias voltage (interlayers): -400 V; Solid sources: carbon and Ti cathodes; Sputtering power: 5 kW; Total deposition time: 10–15 h.
With these parameters the expected layer thickness of 2–3 lm and a hardness of 4–4.5 GPa were obtained. 2.3 Raman spectroscopy Raman spectroscopy was used to evaluate chemical and structural modifications generated during the DLC deposition. Witec Focus Innovations equipment was used with the following parameters: • • • •
Standard: silicon; Integration: 5 s; Accumulations: 109; Laser: Nd:YAG laser, 532 nm;
123
Raman spectra were measured in a range of 100–2,000 cm-1 at an operating power of 25 mW. 2.4 Hardness The nanohardness (H) values were determined using a MTS Nano Indenter XP equipment with a diamond Berkovich geometry tip. The maximum applied load was 400 mN and 9 indentations were made on each sample with a distance of 100 lm. The nanohardness was calculated from 8 loading–unloading cycles using Oliver and Pharr method [9]. 2.5 Reciprocating linear tribometer This test was performed to evaluate the wear rate of the samples, with and without the DLC coating. A reciprocating linear tribometer from CSM Instruments was used. The reciprocating frequency was 2 Hz, the same used by ¨ rstele et al [10]. O For the Ti6Al4V samples, a load of 5 N was applied. For the coated samples the load was increased to 10 N, since 5 N was insufficient to produce debris in sufficient quantity to microscopical analysis. Other parameters used in this test were: • • • •
Number of cycles: 5,000; Tip: round Tungsten Carbide (WC), 6 mm diameter; Frequency: 2 Hz; Amplitude: 4 mm.
The wear track section was obtained using a profilometer (Veeco Dektak 150). The transversal section area and the wear rate were calculated using the Tribox Software (v 2.0), from CSM Instruments. 2.6 Collecting debris After the tribological test, samples were deposited in an aluminum recipient (30 mm diameter and 20 mm height). Inside this recipient, samples were washed out with acetone. This set was submitted to ultrasonic cleaning for 50 min. The sample was removed from the recipient and the recipient with acetone containing the suspended particles was taken to an oven to evaporate. With the complete evaporation of the acetone, debris was left in the bottom of the recipient, allowing further tests. 2.7 Scanning electronic microscopy The collected debris was analyzed using scanning electronic microscopy (SEM). The SEM analyses were performed in Philips XL30 equipment. Semi-quantitative
J Mater Sci: Mater Med
elementary analyses were also done, using EDAX X-ray dispersive spectroscope. The parameters were: • •
Magnification: 1009, 2009, 5009 and 10009 Acceleration voltage: 15 and 20 kV
Those parameters were adjusted according to the size of each fragment.
The DLC diamond peak begins at 1,340 cm-1, and it’s known that the natural, unstressed diamond has a single sharp peak at 1,332 cm-1 [6]. There’s a second, broader peak at 1,580 cm-1, corresponding to the G band (graphite-like film). The intensity of the sp2 peak is slightly bigger than the sp3 peak (approximately 2 % bigger), indicating that the deposited film has almost the same amount of diamond and graphite [6].
2.8 Debris analysis through images 3.2 Hardness From SEM, debris was analyzed regarding their area, length and Feret ratio. Measurements were performed using the software from the Carl Zeiss Axionvision optical microscope. The area measure is obtained from the visible area of the debris, captured from the image. The Feret ratio is the result of dividing the largest Feret diameter by the smallest Feret diameter. The largest Feret diameter is the largest distance among two points in the debris boundary. The smallest Feret diameter is the smallest distance among two points in the debris boundary that is perpendicular to the largest Feret diameter. Feret ratio will always be a value in the range 0–1. Values closer to 1 indicate that the debris is rounder, while Feret ratios closer to 0 will indicate that its shape is oblong.
The hardness values obtained with instrumented indentation are shown in Fig. 3. It’s possible to observe that the hardness is increased just for a small displacement, indicating that the film was thin (2–3 lm, as expected). Given that the graphite corresponded to almost 50 % of the film content, the hardness of
3 Results 3.1 Raman analysis Raman spectroscopy was performed in Ti6Al4V and Ti6Al4V ? DLC samples. In Fig. 1 it is possible to observe the peaks that correspond to the Ti6Al4V alloy. In Fig. 2 it is possible to observe the Raman spectroscopy for the Ti6Al4V ? DLC sample, specifically the shift region of the DLC coating.
Fig. 1 Raman spectroscopy of the Ti6Al4V sample
Fig. 2 Raman spectroscopy of the Ti6Al4V ? DLC sample
Fig. 3 Hardness results of Ti6Al4V and Ti6Al4V ? DLC samples surfaces
123
J Mater Sci: Mater Med
average wear rate was of 12 9 10-6 mm3/N/m. The DLC deposition increased the wear resistance in about 35 times. The high sp2 content lead to a lubricating property with very small wear being observed in the WC counterpart. 3.4 Coefficient of friction The coefficient of friction (COF) values are shown in Fig. 5. The DLC coated sample had a smaller COF than the Ti6Al4V sample. This corroborates the lubricating property of the high graphitic film, with the COF being steady as the number of cycles increased. On the uncoated sample, the COF increases with the number of cycles, as debris from the wear accumulate. 3.5 Scanning electron microscope of the cross-section SEM analysis were performed in the cross-section of the samples, as shown if Fig. 6. The DLC coating has a width of 2.5 lm, within the manufacturer specification tolerance. Small differences in the width and some dark spots are a consequence of deposition parameters (such as small temperature variation) and sample preparation to the SEM imaging. 3.6 Scanning electron microscope and energydispersive X-ray spectroscopy of debris
Fig. 4 Wear track of a Ti6Al4V and b Ti6Al4V? DLC samples
4.3 GPa was equivalent of a soft a-C:H amorphous carbon composition. 3.3 Wear track and wear rate The wear track for the Ti6Al4V and Ti6Al4V ? DLC samples are shown in Fig. 4. It is possible to observe that the wear track for the coated sample is smaller than the Ti6Al4V alloy, with approximately 33 % of its width (approximately 250 lm against 750 lm for the uncoated sample). As mentioned earlier, tribological test had a different load for each sample, with a load of 5 N on the Ti6Al4V uncoated sample and a load of 10 N on the DLC coated sample. Using a smaller load on the Ti6Al4V ? DLC sample barely produced any wear, preventing the observation of debris. And a higher load (10 N) on the Ti6Al4V sample lead to an overrun of the equipment limits, causing excessive tangential force. The wear rate for the Ti6Al4V sample was on average 437 9 10-6 mm3/N/m. For the DLC coated samples, the
123
SEM analysis was performed on debris collected from the tribological test. Energy-dispersive X-ray spectroscopy (EDS) was also performed in some debris, evaluating if the chemical composition was correspondent to the expected from the sample substrate, and coating or even from WC spheres. In Fig. 7 it is possible to observe some debris from the Ti6Al4V sample, with a magnification of 5009. The EDS analysis was also performed, showing the expected composition of Ti, Al and V. It’s possible to observe that the debris size distribution is very irregular, with some particles having a length of 1 lm or less. The shape of the larger particles is also very irregular, including agglomerations of particles with many sizes in multiple layers (for example, in the lower right of Fig. 7b). In Fig. 8 debris from the Ti6Al4V ? DLC samples are shown, with a magnification of 5009. EDS results showed the presence of C, O, Si and W, being Carbon and Silicon elements from the DLC coating: C is the main element of the film deposition and Si is used in a pre-deposition step, to increase the adherence of the DLC film to the substrate. Tungsten comes from the WC sphere and just small traces were found on the EDS test.
J Mater Sci: Mater Med Fig. 5 Coefficient of friction of Ti6Al4V and Ti6Al4V ? DLC samples
4 Discussion In the Raman spectroscopy of the Ti6Al4V ? DLC samples it’s possible to observe that the DLC coating structure is responsible for two characteristic peaks at the 1,332 and 1,580 cm-1 region, being respectively the D (diamond— sp3) and G (graphite—sp2) band [11]. The D and G bands can be fitted in the Raman spectra using two Gaussian curves, and their ratio can be used to estimate the sp2/sp3 ratio. Our DLC coating had high presence of sp2 structure, corroborating the graphitic properties found in the tests [12, 13]. It was observed that the DLC deposition increases the hardness in approximately 15 %. After 1,000 lm the values of both surfaces are the same. There is influence of the substrate in the coating hardness. According to Tsui and Pharr [14] when the depth of the nanoindentation is bigger than 10 % of the coating thickness, results are influenced by the substrate. Since the DLC coating had a thickness of 2,500 nm, all measures with a displacement over 250 nm have influence from the Ti6Al4V hardness. The observed DLC hardness is smaller from the theoretical hardness values due to the film thickness, substrate hardness, sp2 to sp3 ratio, and other factors [15]. The deposition of the DLC coating was sufficient to increase the resistance against the sliding wear caused by the linear reciprocal movement. The wear track edge of the TI6Al4V ? DLC sample is also smoother, without plucked chips, indicating that the wear mechanism probably wasn’t related to crack propagation.
The DLC deposition parameters lead to a higher content of sp2 hybridizations. According to Wen et al. [16] DLC coatings with higher sp2 content have more evident graphitic properties, lowering the COF. It’s possible to observe that as the number of cycles increases, the Ti6Al4V COF also increases, probably due to the accumulation of debris on the sample surface. The self-lubrication characteristic of the graphitic DLC coating helps the steady COF, independently of the number of cycles. Debris from the Ti6Al4V sample had a size, on average, smaller than 20 lm, and in a great amount, sometimes forming small clusters with an elongated irregular shape. According to Sargeant and Goswami, debris with that size promotes an increased response in macrophages, leading to an intense biological host response [17]. Hallab and Jacobs assert that elongated debris, smaller than 10 lm, stimulate more inflammatory host response, activating phagocytosis from macrophages. Wright and Goodman [18] state that debris with size smaller than 20 lm stimulates the release of the tumor necrosis factor (TNF) cytotoxin. Much debris from the Ti6Al4V ? DLC samples had a length bigger than 60 lm. Very few debris was produced from those samples, and they were isolated on the recipient, allowing individual images of 46 particles. Unlike the debris from the uncoated Ti6Al4V sample, this debris does not resemble a cluster. The majority seems to be made of one single fragment, and some seem to be made of two parts. This debris geometry indicates elongated plates, probably generated by a delamination from the
123
J Mater Sci: Mater Med
Fig. 6 Cross-section of the a Ti6Al4V and b Ti6Al4V ? DLC samples
Fig. 7 Debris from the tribological test on the Ti6Al4V sample, 9500 magnification
DLC coating, facilitated by the sp2 structure. It is expected that these debris produce less intense immunological host response, which is desirable in long-life prostheses [19, 20]. It was possible to calculate the area and Feret ratio and measure the length of the particles. These results are shown in Table 1. The particles individualized by the SEM software had some parameters calculated. The average particles area was 105.9 lm2 for the Ti6Al4V samples and 4,477.8 lm2 for the DLC coated samples. It’s worth mentioning that some debris from the Ti6Al4V samples were clustered and could not be analyzed individually, and so the average area of each particle was probably much smaller. The length (maximum distance among two points on the edge of the particle) was measured, and the average particle length was 12.5 lm for the Ti6Al4V samples and 107.40 lm for the Ti6Al4V ? DLC samples. The standard
deviation of the measure was proportionally bigger for the titanium-alloy samples, since some clusters were measured as a single particle. The Feret ratio was also calculated. A Feret ratio closer to 1 would indicate a round shape. Values closer to 0.5 indicate that the width of the debris is, roughly, half of the length. A small difference was found with Ti6Al4V debris being slightly bigger. In Fig. 9 the debris length distribution is shown for the Ti6Al4V samples. It’s possible to observe that almost half of the particles have a size of 5 lm or less. The 90th percentile was calculated as 25.3 lm. In Fig. 10 the debris length distribution is shown for the Ti6Al4V ? DLC samples. It’s possible to observe that\10 % of the particles had a length of 40 lm or less. The 90th percentile was calculated as 200 lm.
123
J Mater Sci: Mater Med Table 1 Area, diameter and Feret Ti6Al4V ? DLC samples debris
These results can be compared with other studies, where debris from hip prosthesis made from different materials were analyzed. Commercially, hip prosthesis
of
Ti6Al4V
and
Ti6Al4V
Ti6Al4V ? DLC
Average area (lm2)
105.9
4,477.8
Average length (lm)
12.5
107.4
Average Feret ratio
Fig. 8 Debris from the tribological test on the Ti6Al4V ? DLC sample, 9500 magnification
ratio
0.63
0.58
are commonly made from a metallic (Cobalt Chrome or Stainless Steel) or ceramic (alumina–zirconia) femoral head which articulates on a polymeric acetabular cup (UHMWPE or PUC) [21–23]. Tipper et al. [21] characterized debris from UHMWPE against zirconia ceramic and found that approximately 85 % of the debris had 0.1–0.5 lm. They also tested a metal-on-metal pair (low carbon wrought cobalt chrome alloy against high carbon wrought cobalt chromium alloy) and found debris with an average length of 30 nm. For a ceramic-on-ceramic pair (alumina) the debris was even smaller, with an average length of 9 nm. Affatato et al. [22] tested UHMWPE against ceramics heads and found debris that had, mainly, a length of 200–300 lm for a 20 % alumina—80 % zirconia composition; 10–15 lm for a 40 % alumina–60 % zirconia composition and up to 100 lm for a pure alumina ceramic. Wu and Peng [24] studied debris from UHMWPE against high carbon cobalt chromium alloy and found that approximately 22.4 % of them were smaller than 0.1 lm, 71.6 % were in a size range of 0.1–10 lm and 6 % were bigger than 10 lm. Wang et al. [25] tested UHMWPE against cobalt chromium alloy and found two types of UHMWPE debris: a rounded type, which was smaller than 0.3 lm, and an elongated type, with most debris having a length of 1–2 lm. Debris size can be affected by testing conditions, like the use of a hip simulator, applied load and lubricating conditions. Nevertheless, debris generated from the DLCcoated Ti6Al4V has a size comparable to those obtained with UHMWPE against ceramic pairs (alumina or alumina–zirconia). Comparing particles sizes from DLC-coated and uncoated samples, it’s possible to observe that the DLC coating increased the average particle size and (based on the wear rate and difficulties in finding those debris using SEM) reduced its amounts. As mentioned earlier, the literature [16–20] considers that the immunological response from the host is more intense for particles with smaller sizes, including a greater release of the tumor necrosis factor cytotoxin.
123
J Mater Sci: Mater Med Fig. 9 Debris length distribution and cumulative curves for the Ti6Al4V samples
Fig. 10 Debris length distribution and cumulative curves for the Ti6Al4V ? DLC samples
5 Conclusion Increasing the lifetime of orthopedic prostheses would benefit patients, as less replacement surgeries would be required. Hard coatings represent one possibility, since an increase in tribological properties of prostheses materials could minimize its wear due to sliding forces that are applied during everyday movements of the patient. DLC coating have desirable properties: the sp2/sp3 ratio determines the hardness and self-lubricating characteristic, and the chemical composition and structure are biological inert, avoiding the host biological response. In this research Ti alloys used as implant were coated with DLC layer with the goal to obtain surface that can
123
increase the lifetime of orthopedic prostheses. The results indicate that the wear rate decreases when the Ti6Al4V sample is coated with DLC, producing less debris. Debris morphology also changes, with Ti6Al4V ? DLC leading to bigger particles, a factor that reduces phagocytosis and the deployment of tumor necrosis factor. DLC coating might be a viable superficial treatment for Ti6Al4V prostheses that have some associated movement, like hip prostheses. Acknowledgments The authors would like to thanks Dr. Ronaldo Ruas, from Brasimet Bodycote, for the DLC deposition; Irineu Vitor Leite and Dr. Geninho Thome´, from Neoortho Produtos Ortope´dicos S/A, for the Ti6Al4V samples; LabNano and CME from UFPR for hardness tests and Raman spectroscopy.
J Mater Sci: Mater Med
References 1. Hartofilakidis G, Babis GC, Lampropoulou-Adamidou K. Congenital hip disease in adults. Berlin: Springer; 2014. p. 3–10. 2. Singh J, Lewallen DG. Patients with osteoarthritis and avascular necrosis have better functional outcomes and those with avascular necrosis worse pain outcomes compared to rheumatoid arthritis after primary hip arthroplasty: a cohort study. BMC Med. 2013;11:210. 3. Huang L, et al. Nanomechanical properties of nanostructured titanium prepared by SMAT. Surf Coat Technol. 2006;201:208–13. 4. Marciano FR, et al. Thermodynamic aspects of fibroblastic spreading on diamond-like carbon films containing titanium dioxide nanoparticles. Theor Chem Acc. 2011;10:24. 5. Wu Y, et al. Preparation and properties of Ag/DLC nanocomposite films fabricated by unbalanced magnetron sputtering. Appl Surf Sci. 2013;284:165–70. 6. Nibennanoune Z, et al. Improving diamond coating on Ti6Al4V substrate using a diamond like carbon interlayer: raman residual stress evaluation and AFM analyses. Diam Relat Mater. 2012;22:105–12. 7. Reuter S, et al. Correlation of structural properties of commercial DLC coatings to their tribological performance in biomedical applications. Wear. 2006;261:419–25. 8. Brossa F, et al. Tribological behavior of Ti6Ai4V modified by surface treatments. J Mater Sci Mater Med. 1996;7:471–4. 9. Oliver WC, Pharr GM. An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J Mater Res. 2011;7:1564–83. ¨ sterle W, et al. Potential of wear resistant coatings on Ti–6Al–4V 10. O for artificial hip joint bearing surfaces. Wear. 2008;264:505–17. 11. Schwan J, et al. Raman spectroscopy on amorphous carbon films. J Appl Phys. 1996;80:440. 12. Yadav VS, et al. Study of Raman spectra of nano-crystalline diamond like carbon (DLC) films composition (sp2:sp3) with
13.
14.
15. 16. 17. 18.
19. 20. 21.
22.
23.
24. 25.
substrate temperature. Proc World Congr Eng Comput Sci. 2009;1:4. Tai FC, et al. Correlation between ID/IG Ratio from visible Raman spectra and sp2/sp3 ratio from XPS spectra of annealed hydrogenated DLC film. Mater Trans. 2006;47(7):1847–52. Tsui TY, Pharr GM. substrate effects on nanoindentation mechanical property measurements of soft films on hard substrates. J Mater Res. 1999;14:292–301. Fischer-Cripps AC. Nanoindentaion. New York: Springer; 2004. Wen, et al. Surface evolution of a gradient structured Ti in hydrogen peroxide solution. Appl Surf Sci. 2008;254:2905–10. Sargeant A, Goswami T. Hip implants: paper V. Physiological effects. Mater Des. 2006;27:287–307. Wright TM, Goodman SB. Implant wear in total joint replacement: clinical and biological issues, material and design considerations. American Academy of Orthopaedic Surgeons; 2002. Hallab NJ, Jacobs JJ. Biologic effects of implant debris. Bull NYU Hosp Jt Dis. 2009;67:182–8. Hunt JA, Willians DF. The effect of titanium debris on soft tissue response. J Mater Sci Mater Med. 1994;5:381–3. Tipper JL, et al. Characterization of wear debris from UHMWPE on zirconia ceramic, metal-on-metal and alumina ceramic-onceramic hip prostheses generated in a physiological anatomical hip joint simulator. Wear. 2001;250:120–8. Affatato S, et al. Isolation and morphological characterization of UHMWPE wear debris generated in vitro. Biomaterials. 2001;22:2325–31. Elsner JJ, et al. Wear rate evaluation of a novel polycarbonateurethane cushion form bearing for artificial hip joints. Acta Biomater. 2010;6:4698–707. Wu J, Peng Z. Investigation of the geometries and surface topographies of UHMWPE wear particles. Tribol Int. 2013;66:208–18. Wang A, et al. Comparison of the size and morphology of UHMWPE wear debris produced by a hip joint simulator under serum and water lubricated conditions. Biomaterials. 1996;17:865–71.
123
