- Email: [email protected]
Effect of thermal oxidation on corrosion and corrosion–wear behaviour of a Ti–6Al–4V alloy
ARTICLE IN PRESS
Biomaterials 25 (2004) 3325–3333
Effect of thermal oxidation on corrosion and corrosion–wear behaviour of a Ti–6Al–4V alloy . . Huseyin . Hasan Gulery uz, C - imeno&glu* Department of Metallurgy and Materials Engineering, Istanbul Technical University, ITU Ayazaga Kampusu Maslak, 34469 Maslak, Istanbul, Turkey Received 20 June 2003; accepted 30 September 2003
Abstract In this study, comparative investigation of thermal oxidation treatment for Ti–6Al–4V was carried out to determine the optimum oxidation conditions for further evaluation of corrosion–wear performance. Characterization of modified surface layers was made by means of microscopic examinations, hardness measurements and X-ray diffraction analysis. Optimum oxidation condition was determined according to the results of accelerated corrosion tests made in 5 m HCl solution The examined Ti–6Al–4V alloy exhibited excellent resistance to corrosion after oxidation at 600 C for 60 h. This oxidation condition achieved 25 times higher wear resistance than the untreated alloy during reciprocating wear test conducted in a 0.9% NaCl solution. r 2003 Elsevier Ltd. All rights reserved. Keywords: Corrosion; Corrosion–wear; Titanium; Ti–6Al–4V alloy; Thermal oxidation; Wear
1. Introduction Biomedical devices are subjected to action of sliding and rubbing contact of articulating surfaces during their service in the body [1]. This situation leads to localized stresses at the contact regions and may cause heavy damage on their surfaces. Additionally, wear may progress on the implant surface very rapidly due to the combined effect of corrosion in the human body and corrosion assisted wear resulting in severe surface damage. Titanium and its alloys are known as the most appropriate materials for biomedical applications, due to their well-established corrosion resistance and biocompatibility. They owe their excellent corrosion resistance to passive oxide film formation at room temperature. This film, which mainly consists of TiO2, provides chemical inertness in many aqueous media and assures their biocompatibility as a biomaterial [1–3]. However, in their native form TiO2 films have poor mechanical properties and they are easily fractured under fretting and sliding wear conditions. Sustained dissolution of underlying metal after the disruption of *Corresponding author. Tel.: +90-212-285-68-34; fax: +90-212285-34-27. E-mail address: [email protected] (H. C - imeno&glu). 0142-9612/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2003.10.009
oxide film and the reformation of passive oxide layer result in gradual consumption of the material [4–7]. Furthermore, formation of wear debris and release of metal ions cause adverse tissue reactions, implant loosening and eventual revision surgery [8,9]. Thermal oxidation treatment tends to improve surface characteristics of titanium and its alloys. Oxidation, particularly at temperatures above 200 C, promotes the development of a crystalline oxide film. Increasing temperature induces the formation of a thicker oxide layer, which is accompanied with dissolution of oxygen beneath it [10]. Since the formation of mechanically stable and chemically resistant oxide layers affects corrosion and wear behavior of titanium and its alloys [7,11–13], in this study, we aimed to determine the optimum thermal oxidation condition of a Ti–6Al–4V alloy on the basis of corrosion and corrosion–wear response [14].
2. Materials and methods Ti–6Al–4V alloy utilized in the present investigation was received as 8 mm diameter cold drawn rod. The cylindrical samples cut from the rod were prepared with a surface-finishing route to achieve a mirror like appearance. Later they were cleaned in acetone and
ARTICLE IN PRESS 3326
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
dried in hot air. Thermal oxidation treatments were conducted at 600 C and 650 C for between 12 and 60 h at normal atmospheric condition. After characterization studies, corrosion and corrosion–wear tests were performed on untreated and oxidized samples. 2.1. Characterization tests Characterization of the surfaces was made by surface roughness and hardness measurements, X-ray diffraction (XRD) analysis and optical microscopic examinations. Three samples were averaged at each characterization test to accurately determine relevant properties. 2.1.1. Surface roughness measurements A Mahr Perthen Perthometer S8P optical profilometer was used to examine the surface roughness of the samples. The effect of oxidation conditions on surface roughness was determined according to average roughness (Ra) values, which define the arithmetic mean of departure of a surface profile from a mean line. 2.1.2. Hardness measurements Hardness tests were carried out with a Vickers pyramid indenter, using a Fischer HP 100 XY-PROG ultramicrohardness tester. Hardness measurements were performed on the surfaces of the samples under four different indentation loads, ranging from 250 to 1000 mN. At each load level, at least 10 successive measurements were made. 2.1.3. XRD analysis A Philips RV 3710 X-ray diffractometer was used for the XRD analysis. The glancing incidence X-ray diffraction technique was used for surface phase identification of oxidized samples. CuKa radiation source was used and the incidence beam angle was 2 . Diffraction angle range was between 10 and 90 , with a step increment of 0.02 and a count time of 1 s.
containing 5 m HCl for 60 h. The minimum amount of the solution reacting in the corrosion tests was determined, taking into account the surface area of the samples as 0.3 ml/mm2 [15]. During the corrosion tests the temperature of the solution was 2474 C. The results of the corrosion tests were evaluated by measuring the weight loss of the samples at certain intervals, with an accuracy of 0.1 mg. After the test period of 60 h the surface appearances of the samples were examined with a microscope at 10 magnification. 2.3. Corrosion–wear tests Corrosion–wear tests were conducted on a reciprocating wear tester described in ASTM G133 standard. The length and the width of the wear test samples, which were sliced from the transverse section of the 8 mm diameter rod, were 20 and 8 mm, respectively. Two samples were tested to examine the wear performance of each surface condition. During wear testing 150 g normal load was applied on the surfaces of the samples with a 10 mm diameter Al2O3 ball. A commercially available isotonic serum containing 0.9% NaCl was used as corrosive environment to simulate the corrosive effect of the human body. Sliding speed of the ball was 20 mm/s and the stroke of the reciprocating motion was 9 mm. Wear tests were interrupted at certain intervals to determine the progress of wear. At each interval samples were cleaned with alcohol and wear tracks were examined by a surface profilometer. After examining the wear tracks developed on the surfaces, the samples were returned to the wear tester in their original arrangement. Wear tests were performed for 300 min. After the wear tests, wear tracks that developed on the untreated and oxidized surfaces were investigated by a JEOL JSM 6335 F scanning electron microscope (SEM). The wear of the Al2O3 balls was also examined with an optical microscope at the end of the wear tests.
3. Results 2.1.4. Microscopic examinations Microstructural examinations were conducted on the cross-sections of the oxidized samples with a Zeiss light optical microscope. Cross-sections of the oxidized samples were prepared by standard metallographic technique. After grinding and polishing metallographic samples were etched with 2% HF solution. 2.2. Corrosion tests Accelerated corrosion tests were conducted to evaluate the relative corrosion resistance of untreated and oxidized samples. Two samples were used for each surface condition. Samples having 8 mm diameter and 8 mm height were suspended in an acidic solution
3.1. Characterization tests After thermal oxidation the surface of the Ti–6Al–4V alloy was covered with a dark colored oxide layer. Roughness of oxidized surfaces increased drastically with increasing oxidation time and temperature as presented in Fig. 1. Average roughness of untreated sample was 0.17 mm before oxidation. XRD patterns of the untreated and oxidized samples are given in Fig. 2. a-Ti and b-Ti peaks were obtained at different diffraction angles corresponding to different crystallographic planes of untreated alloy (Fig. 2a). They were also observed after oxidation at 600 C (Fig. 2b). This was basically, due to penetration of
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
3327
Fig. 1. The effect of oxidation temperature and time on the average roughness of oxidized surfaces.
X-ray beyond the thin oxide layer, through the substrate. Further analysis revealed that diffraction angles of Ti peaks shifted slightly left from their original positions; probably caused by the dissolution of oxygen in the subsurface zone. Oxidized surfaces principally consisted of rutile modification of TiO2 (Fig. 2b and c). The anatase form of TiO2 was also detected at limited number of diffraction angles, especially after oxidation at 600 C (Fig. 2b). However, at 650 C rutile totally dominated the oxide structure (Fig. 2c). Surface hardness of untreated and oxidized samples under varying indentation loads are presented in Fig. 3. The hardness values for the oxidized samples show a large amount of scatter as indicated by the standard deviation bars. Higher oxidation temperature and/or lower indentation load yielded larger scatter in hardness values. This observation can be attributed to contribution of surface roughness to the indentation response of oxidized surfaces. In untreated samples the surface hardness is almost constant at about 3500 MPa with very low scatter for indentation loads of 250–1000 mN. However, a dramatic increment of up to 9000 MPa in surface hardness was achieved upon oxidation. On the oxidized surfaces the decrease of hardness with increasing indentation load can be attributed to the involvement of softer regions at high penetration depths of the indenter. It should be noted that, higher hardness values were maintained for a wide indentation load range, with increasing oxidation time and temperature. This indicates that, deeper hardened layers were achieved at high oxidation temperatures and/or times. Cross-sectional optical micrographs of the oxidized samples are given in Fig. 4. Beneath the oxide layer an oxygen diffusion zone appeared as a white colored region after etching. At low temperature relatively thin
Fig. 2. XRD patterns of untreated and oxidized surfaces (a: hcp titanium, b: bcc titanium, A: anatase, R: rutile): (a) Before oxidation (untreated condition); (b) after oxidation at 600 C for 48 h; (c) after oxidation at 650 C for 48 h.
oxide films were observed, which remained intact with the subsurface oxygen diffusion zone (Fig. 4a). Increasing temperature and oxidation time promoted formation of thick oxide layers and deeper penetration of oxygen into the metal (Fig. 4b). The variation of oxide layer thickness and oxygen diffusion zone with respect to oxidation conditions are presented in Fig. 5. 3.2. Corrosion tests Fig. 6 presents the averaged results of two individual corrosion tests as weight loss of the samples with respect to duration time in 5 m HCl solution. Continuous dissolution of untreated alloy, which exhibited the lowest corrosion resistance among the investigated
ARTICLE IN PRESS 3328
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
Fig. 3. The effect of oxidation temperature on the Vickers hardness (HV) values of oxidized surfaces for oxidation times of: (a) 12, (b) 24, (c) 36, (d) 48 and (e) 60 h.
Fig. 4. Cross-sectional optical micrographs of the samples oxidized (a) at 600 C for 60 h and (b) at 650 C for 60 h (OL: Oxide layer, ODZ: Oxygen diffusion zone).
samples, was sustained throughout the testing period. No measurable weight loss was obtained on the samples that were oxidized at 650 C for duration times of 12 and 36 h. Prolonged corrosion tests resulted in abrupt increase in weight loss due to flaking and removal of the oxide scales. For the samples oxidized at 600 C, weight loss gradually increased with increasing duration
time. Longer oxidation times at 600 C formed a thicker protective surface layer, which did not show any weight loss in 36 h. Among the investigated materials, the sample oxidized at 600 C for 60 h exhibited the maximum corrosion resistance. Fig. 7 shows the optical micrographs of the cylindrical surfaces of the samples after corrosion testing time of
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
3329
Fig. 5. The effect of oxidation time and temperature on thickness of OL and ODZ.
Fig. 6. Weight loss data obtained through the accelerated corrosion tests.
60 h. The surfaces of the samples can be characterized by dark and light colored regions. Light colored regions are the substrate, while the dark colored regions are the oxide layer. During corrosion testing, light colored regions appeared due to local removal of dark colored oxide layers. Fragments of oxide lost contact with the substrate and separated from the surface without dissolving in the acidic solution. The surface of the sample, which exhibited the maximum corrosion resistance, was still covered with oxide layer. This indicates that, very stable oxide formed on the surface of the examined Ti–6Al–4V alloy after oxidation at 600 C for 60 h. Thick oxide layers (formed on surfaces of the samples oxidized at 650 C) resisted the attack of corrosive media for 36 h without any weight loss (Fig. 6). However, at longer testing times fragments of oxide layers detached from the surfaces allowing further corrosion progress at a rate similar to that for untreated alloy. Therefore, after a testing period of 60 h very few
Fig. 7. Photographs of surface appearances of oxidized samples after 60 h of accelerated corrosion testing.
oxide pieces (dark regions) were observed on sample surfaces (Fig. 7). Since the oxide layer is very effective in protecting the titanium and its alloys from the detrimental effects of corrosive media, the higher the stability of the oxide layer through out the testing period, the better is the corrosion resistance. 3.3. Corrosion–wear tests Corrosion–wear tests were carried out to compare the wear resistance of untreated alloy with that of the alloy
ARTICLE IN PRESS 3330
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
Fig. 8. 3-D surface profiles of untreated and oxidized surfaces at certain intervals of corrosion wear tests.
oxidized at 600 C for 60 h in a mild corrosive media. Fig. 8 displays the 3-D profiles of the wear tracks developed on the untreated and oxidized surfaces at certain intervals of corrosion–wear tests. It is clearly seen that shallower and narrower wear track was formed on oxidized surfaces than untreated ones. The results of corrosion–wear tests were quantified by measuring the cross-sectional areas of the wear tracks from their 2-D profiles. Fig. 9 graphically shows the progress of wear track area with respect to testing time. Steady wear of untreated surface was restrained by thermal oxidation. Accordingly, oxidized surface exhibited approximately 25 times higher corrosion–wear resistance than the untreated alloy, after testing time of 300 min. In Fig. 10 SEM micrographs of the wear tracks developed on the untreated and oxidized surfaces after a
testing time of 300 min are presented. The wear tracks produced on these samples exhibited different topography. A typical feature of rough wear track produced on untreated surface was extensive shear deformation due to ploughing action of the Al2O3 ball (Fig. 10a). The wear tracks on oxidized surface were smooth (Fig. 10b) and contained microcracks with traces of local material removal (Fig. 10c). Thus, on oxidized surface wear was initiated by disruption of oxide layer and followed by the wear of oxygen diffusion zone, which was locally removed by cracking at extended testing times. The Al2O3 balls were also subjected to wear during testing. The wear scars formed on the balls are depicted in Fig. 11. Optical microscopic examinations revealed that, heavy wear of the ball on untreated surface progressed by grain pull out mechanism. The topogra-
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
phy of wear scar developed on the ball that slid on the oxidized surface was almost similar with that of untreated surface but smaller by a factor of 16. Thus, the size of wear scar was proportional to the size of the wear track formed on the surface of the material being tested.
Fig. 9. Progress of wear track areas during corrosion–wear tests.
3331
4. Discussion Thermal oxidation at 600 C and 650 C modified the surface of the examined Ti–6Al–4V alloy by introducing an oxide layer (anatase and/or rutile) with an oxygen diffusion zone beneath it. Increase of surface hardness after thermal oxidation can be attributed to the hard oxide layer covering the surface and contribution of strains in the matrix due to dissolution of oxygen [16,17]. The thickness of the oxide layer and oxygen diffusion zone increased with increasing oxidation time. However, growth of the oxide layer during thermal oxidation was accompanied by surface roughening. The thickness and the roughness of the oxide layer increased in an accelerated manner, when oxidation was conducted at 650 C. This situation affirms the role of temperature on oxidation rate by activating diffusing species. On the other hand, accelerated oxidation rate due to increasing temperature may put the whole oxide layer in a more stressed condition as a result of intrinsic and thermal effects [18,19]. The formation of defective oxide structure provides easy diffusion paths for oxygen and metal ions allowing oxidation progress in an uncontrollable way. Since the corrosion resistance of thermally oxidized surface layers relies on their chemical stability and defect free structure, this observation may also be
Fig. 10. SEM micrographs of the wear tracks developed on (a) untreated and (b) oxidized surfaces. High magnification of the framed region is shown in the in micrograph (c).
ARTICLE IN PRESS 3332
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
Fig. 11. The wear scars that formed on Al2O3 balls during corrosion–wear testing of: (a) untreated, and (b) oxidized surfaces.
valid for explaining the corrosion behavior of oxidized samples because strong reducing acids would not dissolve heavy oxide layer on titanium [20]. Small structural defects (microcracks and micropores) may be responsible for the failure of oxidized surfaces [21]. During incubation period, aggressive ions penetrate towards the oxide–metal interface and steadily extend, undermining the oxide because of the galvanic coupling of the oxide and substrate metal. Oxide layers gradually detached from the surface as fragments. This observation points out the fact that corrosive solution has entered the interface through individual defects. As soon as the oxide fragments lost their contact with the substrate, they detached. The corrosion behavior of the samples that were oxidized for short times at 600 C differed to that of the other samples. The relatively thin oxide layer that covered these samples was a mixture of anatase and rutile. Anatase was readily attacked and dissolved in reducing acids whereas rutile was inert and resistant to attack [2,21]. Dissolution of anatase at the beginning of the corrosion test left some sites of oxidized samples unprotected and further corrosion continued with total removal of the oxide layer. Increasing oxidation time yielded higher amount of rutile in the oxidized surface, which increased protective nature of oxide layer against the corrosive environment. Less defective structures of oxide containing more rutile formed at 600 C after longer oxidation times. And this combination presented the most corrosion resistant oxide layer that formed on Ti–6Al–4V alloy. Corrosion–wear test results confirmed the superiority of the thermally oxidized surface over untreated alloy. Surface degradation of untreated alloy commenced with the onset of sliding action and large amount of wear debris was produced through the entire test duration. The oxide layer prevented the wear of Ti–6Al–4V alloy over a certain period of testing time. The removal of oxide layer and diffusion hardened zone occurred due to mechanical and chemical effects. Even though the oxidized (at 600 C for 60 h) surface exhibited the best corrosion resistance in HCl solution with its less
defective structure, it was subjected to mechanical stresses during corrosion–wear test. This situation caused the local breakdown of the oxide layer and allowed the corrosive solution to enter the oxide–oxygen diffusion zone interface. The oxide layer was gradually removed by the combined action of corrosion and wear together. Eventually, corrosion–wear continued on the oxygen diffusion zone by producing a smaller amount of wear debris. After extended testing times, cracking of oxygen diffusion zone caused local material removal, and thereafter the depth of the wear track reached to the substrate. When the Al2O3 ball contacted with the substrate, it was subjected to wear with the mechanism (grain pull out) which was observed during testing of untreated surface Apparently, the wear mechanism of Al2O3 ball on untreated surface was maintained on oxidized surface when the depth of the wear track reached to substrate. Grain pull out wear mechanism is inevitable when the adhesion stresses between the sliding surfaces tend to be greater than the cohesive strength of the materials [22]. It should be emphasized that, the extent of wear developed on the Al2O3 ball is controlled by the size of the contact area between the ball and the surface being tested.
5. Conclusion The Ti–6Al–4V alloy was thermally oxidized at 600 C and 650 C to produce corrosion and wear resistant surface layers. Oxide thickness and oxygen diffusion zone depth steadily increased with increasing oxidation temperature and time. Oxide layers were composed of anatase and rutile structures of TiO2. Roughening of oxidized surfaces was more severe after oxidation at 650 C. A significant increment in surface hardness (3500 to 9000 HV) was achieved due to formation of a hard oxide layer and an oxygen diffusion zone beneath it. According to results of corrosion tests carried out in an aggressive acidic solution (5 m HCl), oxidation at 600 C for 60 h produced the most corrosion resistant surface on Ti–6Al–4V alloy. Wear tests that were
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
conducted in a mild corrosive solution (0.9% NaCl), indicated that thermal oxidation treatment performed at 600 C for 60 h prevented extensive corrosion–wear of Ti–6Al–4V alloy and increased wear resistance by a factor of 25.
[9]
[10]
Acknowledgements The authors gratefully acknowledge the support of the NATO TU PVD Coatings project for the equipment supplied to the Metallurgy and Materials Engineering Department of Istanbul Technical University, which was utilized in characterization work carried out in this study. One of the authors (H.G.) would like to thank The State Planning Organization of Turkey for the support of his MSc. study.
[11]
[12]
[13]
[14]
References [1] Brunette DM, Tengvall P, Textor M, Thomsen P. Titanium in medicine. Heidelberg: Springer; 2001. [2] Boyer R, Welsch G, Collings EW. Materials properties handbook: titanium alloys. Materials Park, OH: ASM International; 1994. [3] Donachie MJ. Titanium: a technical guide. Materials Park, OH: ASM International; 1989. [4] Gilbert JL, Buckley CA, Lautenschlager EP. Titanium oxide film fracture and repassivation: The effect of potential, pH and aeration. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its alloys: the material and biological issues, ASTM STP 1272. Philadelphia: ASTM; 1996. p. 199–214. [5] Komotori J, Lee BJ, Dong H, Dearnley PA. Corrosion response of surface engineered titanium alloys damaged by prior abrasion. Wear 2001;88–98:1–11. [6] Galliano F, Galvanetto E, Mischler S, Landolt D. Tribocorrosion behavior of plasma nitrided Ti–6Al–4V alloy in neutral NaCl solution. Surf Coat Technol 2001;145:121–31. [7] Long M, Rack HJ. Titanium alloys in total joint replacement—a materials science perspective. Biomaterials 1998;19:1621–39. [8] Kovacs P, Davidson JA. Chemical and electrochemical aspects of biocompatibility of titanium and its alloys. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its
[15] [16]
[17]
[18]
[19]
[20] [21]
[22]
3333
alloys: the material and biological issues, ASTM STP 1272. Philadelphia: ASTM; 1996. p. 163–77. Dearnley PA. A review of metallic, ceramic and surface-treated metals used for bearing surfaces in human joint replacements. Institution of Mechanical Engineers proceedings, Part H. J Eng Med 1999;213:107–35. Kofstad P. High temperature corrosion. Essex: Elsevier Applied Science; 1988. Mishra AK, Davidson JA, Poggie RA, Kovacs P, FitzGerald TJ. Mechanical and tribological properties and biocompatibility of diffusion hardened Ti–13Nb–13Zr—a new titanium alloy for surgical implants. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its alloys: the material and biological issues, ASTM STP 1272. Philadelphia: ASTM; 1996. p. 96–112. Lopez MF, Jimenez JA, Gutierrez A. Corrosion study of surfacemodified vanadium-free titanium alloys. Electrochim Acta 2003; 48:1395–401. Garcia-Alonso MC, Saldana L, Valles G, Gonzalez-Garrasco JL, Gonzalez-Cabrero J, Martinez ME, Gil-Garay E, Munuera L. In vitro corrosion behavior and osteoblast response of thermally oxidised Ti6Al4V alloy. Biomaterials 2003;24:19–26. . . H. Surface treatment of Ti–6Al–4V alloy by thermal Gulery uz oxidation. M.Sc. thesis, 2003. Standard practice for laboratory immersion corrosion testing of metals. ASTM G31–72. Philadelphia: ASTM; 1995. Weissman S, Shrier A. Strain distribution in oxidized alpha titanium crystals. In: Jaffee RI, Promisel NE, editors. The science, technology and application of titanium. London: Pergamon Press; 1968. p. 441–58. Borgioli F, Galvanetto E, Galliano FP, Bacci T. Air treatment of pure titanium by furnace and glow-discharge processes. Surf Coat Technol 2001;141:103–7. Coddet C, Chaze AM, Beranger G. Measurement of the adhesion of thermal oxide films: application to the oxidation of titanium. J Mater Sci 1987;22:2969–74. Schutze . M. Failure of oxide scales on advanced materials due to . presence of stresses. In: Saito Y, Onay B, Maruyama T, editors. High temperature corrosion of advanced materials and protective coatings. B.V. Amsterdam: Elsevier Science Publishers; 1992. p. 39–49. Baboian R. Corrosion tests and standards: application and interpretation. Philadelphia: ASTM Manual Series; 1995. Bloyce A, Qi P-Y, Dong H, Bell T. Surface modification of titanium alloys for combined improvements in corrosion and wear resistance. Surf Coat Technol 1998;107:125–32. Dearnley PA, Dahm KL, Cimenoglu H. The corrosion–wear behaviour of thermally oxidised Cp–Ti and Ti–6Al–4V. Wear, in press.
Biomaterials 25 (2004) 3325–3333
Effect of thermal oxidation on corrosion and corrosion–wear behaviour of a Ti–6Al–4V alloy . . Huseyin . Hasan Gulery uz, C - imeno&glu* Department of Metallurgy and Materials Engineering, Istanbul Technical University, ITU Ayazaga Kampusu Maslak, 34469 Maslak, Istanbul, Turkey Received 20 June 2003; accepted 30 September 2003
Abstract In this study, comparative investigation of thermal oxidation treatment for Ti–6Al–4V was carried out to determine the optimum oxidation conditions for further evaluation of corrosion–wear performance. Characterization of modified surface layers was made by means of microscopic examinations, hardness measurements and X-ray diffraction analysis. Optimum oxidation condition was determined according to the results of accelerated corrosion tests made in 5 m HCl solution The examined Ti–6Al–4V alloy exhibited excellent resistance to corrosion after oxidation at 600 C for 60 h. This oxidation condition achieved 25 times higher wear resistance than the untreated alloy during reciprocating wear test conducted in a 0.9% NaCl solution. r 2003 Elsevier Ltd. All rights reserved. Keywords: Corrosion; Corrosion–wear; Titanium; Ti–6Al–4V alloy; Thermal oxidation; Wear
1. Introduction Biomedical devices are subjected to action of sliding and rubbing contact of articulating surfaces during their service in the body [1]. This situation leads to localized stresses at the contact regions and may cause heavy damage on their surfaces. Additionally, wear may progress on the implant surface very rapidly due to the combined effect of corrosion in the human body and corrosion assisted wear resulting in severe surface damage. Titanium and its alloys are known as the most appropriate materials for biomedical applications, due to their well-established corrosion resistance and biocompatibility. They owe their excellent corrosion resistance to passive oxide film formation at room temperature. This film, which mainly consists of TiO2, provides chemical inertness in many aqueous media and assures their biocompatibility as a biomaterial [1–3]. However, in their native form TiO2 films have poor mechanical properties and they are easily fractured under fretting and sliding wear conditions. Sustained dissolution of underlying metal after the disruption of *Corresponding author. Tel.: +90-212-285-68-34; fax: +90-212285-34-27. E-mail address: [email protected] (H. C - imeno&glu). 0142-9612/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2003.10.009
oxide film and the reformation of passive oxide layer result in gradual consumption of the material [4–7]. Furthermore, formation of wear debris and release of metal ions cause adverse tissue reactions, implant loosening and eventual revision surgery [8,9]. Thermal oxidation treatment tends to improve surface characteristics of titanium and its alloys. Oxidation, particularly at temperatures above 200 C, promotes the development of a crystalline oxide film. Increasing temperature induces the formation of a thicker oxide layer, which is accompanied with dissolution of oxygen beneath it [10]. Since the formation of mechanically stable and chemically resistant oxide layers affects corrosion and wear behavior of titanium and its alloys [7,11–13], in this study, we aimed to determine the optimum thermal oxidation condition of a Ti–6Al–4V alloy on the basis of corrosion and corrosion–wear response [14].
2. Materials and methods Ti–6Al–4V alloy utilized in the present investigation was received as 8 mm diameter cold drawn rod. The cylindrical samples cut from the rod were prepared with a surface-finishing route to achieve a mirror like appearance. Later they were cleaned in acetone and
ARTICLE IN PRESS 3326
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
dried in hot air. Thermal oxidation treatments were conducted at 600 C and 650 C for between 12 and 60 h at normal atmospheric condition. After characterization studies, corrosion and corrosion–wear tests were performed on untreated and oxidized samples. 2.1. Characterization tests Characterization of the surfaces was made by surface roughness and hardness measurements, X-ray diffraction (XRD) analysis and optical microscopic examinations. Three samples were averaged at each characterization test to accurately determine relevant properties. 2.1.1. Surface roughness measurements A Mahr Perthen Perthometer S8P optical profilometer was used to examine the surface roughness of the samples. The effect of oxidation conditions on surface roughness was determined according to average roughness (Ra) values, which define the arithmetic mean of departure of a surface profile from a mean line. 2.1.2. Hardness measurements Hardness tests were carried out with a Vickers pyramid indenter, using a Fischer HP 100 XY-PROG ultramicrohardness tester. Hardness measurements were performed on the surfaces of the samples under four different indentation loads, ranging from 250 to 1000 mN. At each load level, at least 10 successive measurements were made. 2.1.3. XRD analysis A Philips RV 3710 X-ray diffractometer was used for the XRD analysis. The glancing incidence X-ray diffraction technique was used for surface phase identification of oxidized samples. CuKa radiation source was used and the incidence beam angle was 2 . Diffraction angle range was between 10 and 90 , with a step increment of 0.02 and a count time of 1 s.
containing 5 m HCl for 60 h. The minimum amount of the solution reacting in the corrosion tests was determined, taking into account the surface area of the samples as 0.3 ml/mm2 [15]. During the corrosion tests the temperature of the solution was 2474 C. The results of the corrosion tests were evaluated by measuring the weight loss of the samples at certain intervals, with an accuracy of 0.1 mg. After the test period of 60 h the surface appearances of the samples were examined with a microscope at 10 magnification. 2.3. Corrosion–wear tests Corrosion–wear tests were conducted on a reciprocating wear tester described in ASTM G133 standard. The length and the width of the wear test samples, which were sliced from the transverse section of the 8 mm diameter rod, were 20 and 8 mm, respectively. Two samples were tested to examine the wear performance of each surface condition. During wear testing 150 g normal load was applied on the surfaces of the samples with a 10 mm diameter Al2O3 ball. A commercially available isotonic serum containing 0.9% NaCl was used as corrosive environment to simulate the corrosive effect of the human body. Sliding speed of the ball was 20 mm/s and the stroke of the reciprocating motion was 9 mm. Wear tests were interrupted at certain intervals to determine the progress of wear. At each interval samples were cleaned with alcohol and wear tracks were examined by a surface profilometer. After examining the wear tracks developed on the surfaces, the samples were returned to the wear tester in their original arrangement. Wear tests were performed for 300 min. After the wear tests, wear tracks that developed on the untreated and oxidized surfaces were investigated by a JEOL JSM 6335 F scanning electron microscope (SEM). The wear of the Al2O3 balls was also examined with an optical microscope at the end of the wear tests.
3. Results 2.1.4. Microscopic examinations Microstructural examinations were conducted on the cross-sections of the oxidized samples with a Zeiss light optical microscope. Cross-sections of the oxidized samples were prepared by standard metallographic technique. After grinding and polishing metallographic samples were etched with 2% HF solution. 2.2. Corrosion tests Accelerated corrosion tests were conducted to evaluate the relative corrosion resistance of untreated and oxidized samples. Two samples were used for each surface condition. Samples having 8 mm diameter and 8 mm height were suspended in an acidic solution
3.1. Characterization tests After thermal oxidation the surface of the Ti–6Al–4V alloy was covered with a dark colored oxide layer. Roughness of oxidized surfaces increased drastically with increasing oxidation time and temperature as presented in Fig. 1. Average roughness of untreated sample was 0.17 mm before oxidation. XRD patterns of the untreated and oxidized samples are given in Fig. 2. a-Ti and b-Ti peaks were obtained at different diffraction angles corresponding to different crystallographic planes of untreated alloy (Fig. 2a). They were also observed after oxidation at 600 C (Fig. 2b). This was basically, due to penetration of
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
3327
Fig. 1. The effect of oxidation temperature and time on the average roughness of oxidized surfaces.
X-ray beyond the thin oxide layer, through the substrate. Further analysis revealed that diffraction angles of Ti peaks shifted slightly left from their original positions; probably caused by the dissolution of oxygen in the subsurface zone. Oxidized surfaces principally consisted of rutile modification of TiO2 (Fig. 2b and c). The anatase form of TiO2 was also detected at limited number of diffraction angles, especially after oxidation at 600 C (Fig. 2b). However, at 650 C rutile totally dominated the oxide structure (Fig. 2c). Surface hardness of untreated and oxidized samples under varying indentation loads are presented in Fig. 3. The hardness values for the oxidized samples show a large amount of scatter as indicated by the standard deviation bars. Higher oxidation temperature and/or lower indentation load yielded larger scatter in hardness values. This observation can be attributed to contribution of surface roughness to the indentation response of oxidized surfaces. In untreated samples the surface hardness is almost constant at about 3500 MPa with very low scatter for indentation loads of 250–1000 mN. However, a dramatic increment of up to 9000 MPa in surface hardness was achieved upon oxidation. On the oxidized surfaces the decrease of hardness with increasing indentation load can be attributed to the involvement of softer regions at high penetration depths of the indenter. It should be noted that, higher hardness values were maintained for a wide indentation load range, with increasing oxidation time and temperature. This indicates that, deeper hardened layers were achieved at high oxidation temperatures and/or times. Cross-sectional optical micrographs of the oxidized samples are given in Fig. 4. Beneath the oxide layer an oxygen diffusion zone appeared as a white colored region after etching. At low temperature relatively thin
Fig. 2. XRD patterns of untreated and oxidized surfaces (a: hcp titanium, b: bcc titanium, A: anatase, R: rutile): (a) Before oxidation (untreated condition); (b) after oxidation at 600 C for 48 h; (c) after oxidation at 650 C for 48 h.
oxide films were observed, which remained intact with the subsurface oxygen diffusion zone (Fig. 4a). Increasing temperature and oxidation time promoted formation of thick oxide layers and deeper penetration of oxygen into the metal (Fig. 4b). The variation of oxide layer thickness and oxygen diffusion zone with respect to oxidation conditions are presented in Fig. 5. 3.2. Corrosion tests Fig. 6 presents the averaged results of two individual corrosion tests as weight loss of the samples with respect to duration time in 5 m HCl solution. Continuous dissolution of untreated alloy, which exhibited the lowest corrosion resistance among the investigated
ARTICLE IN PRESS 3328
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
Fig. 3. The effect of oxidation temperature on the Vickers hardness (HV) values of oxidized surfaces for oxidation times of: (a) 12, (b) 24, (c) 36, (d) 48 and (e) 60 h.
Fig. 4. Cross-sectional optical micrographs of the samples oxidized (a) at 600 C for 60 h and (b) at 650 C for 60 h (OL: Oxide layer, ODZ: Oxygen diffusion zone).
samples, was sustained throughout the testing period. No measurable weight loss was obtained on the samples that were oxidized at 650 C for duration times of 12 and 36 h. Prolonged corrosion tests resulted in abrupt increase in weight loss due to flaking and removal of the oxide scales. For the samples oxidized at 600 C, weight loss gradually increased with increasing duration
time. Longer oxidation times at 600 C formed a thicker protective surface layer, which did not show any weight loss in 36 h. Among the investigated materials, the sample oxidized at 600 C for 60 h exhibited the maximum corrosion resistance. Fig. 7 shows the optical micrographs of the cylindrical surfaces of the samples after corrosion testing time of
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
3329
Fig. 5. The effect of oxidation time and temperature on thickness of OL and ODZ.
Fig. 6. Weight loss data obtained through the accelerated corrosion tests.
60 h. The surfaces of the samples can be characterized by dark and light colored regions. Light colored regions are the substrate, while the dark colored regions are the oxide layer. During corrosion testing, light colored regions appeared due to local removal of dark colored oxide layers. Fragments of oxide lost contact with the substrate and separated from the surface without dissolving in the acidic solution. The surface of the sample, which exhibited the maximum corrosion resistance, was still covered with oxide layer. This indicates that, very stable oxide formed on the surface of the examined Ti–6Al–4V alloy after oxidation at 600 C for 60 h. Thick oxide layers (formed on surfaces of the samples oxidized at 650 C) resisted the attack of corrosive media for 36 h without any weight loss (Fig. 6). However, at longer testing times fragments of oxide layers detached from the surfaces allowing further corrosion progress at a rate similar to that for untreated alloy. Therefore, after a testing period of 60 h very few
Fig. 7. Photographs of surface appearances of oxidized samples after 60 h of accelerated corrosion testing.
oxide pieces (dark regions) were observed on sample surfaces (Fig. 7). Since the oxide layer is very effective in protecting the titanium and its alloys from the detrimental effects of corrosive media, the higher the stability of the oxide layer through out the testing period, the better is the corrosion resistance. 3.3. Corrosion–wear tests Corrosion–wear tests were carried out to compare the wear resistance of untreated alloy with that of the alloy
ARTICLE IN PRESS 3330
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
Fig. 8. 3-D surface profiles of untreated and oxidized surfaces at certain intervals of corrosion wear tests.
oxidized at 600 C for 60 h in a mild corrosive media. Fig. 8 displays the 3-D profiles of the wear tracks developed on the untreated and oxidized surfaces at certain intervals of corrosion–wear tests. It is clearly seen that shallower and narrower wear track was formed on oxidized surfaces than untreated ones. The results of corrosion–wear tests were quantified by measuring the cross-sectional areas of the wear tracks from their 2-D profiles. Fig. 9 graphically shows the progress of wear track area with respect to testing time. Steady wear of untreated surface was restrained by thermal oxidation. Accordingly, oxidized surface exhibited approximately 25 times higher corrosion–wear resistance than the untreated alloy, after testing time of 300 min. In Fig. 10 SEM micrographs of the wear tracks developed on the untreated and oxidized surfaces after a
testing time of 300 min are presented. The wear tracks produced on these samples exhibited different topography. A typical feature of rough wear track produced on untreated surface was extensive shear deformation due to ploughing action of the Al2O3 ball (Fig. 10a). The wear tracks on oxidized surface were smooth (Fig. 10b) and contained microcracks with traces of local material removal (Fig. 10c). Thus, on oxidized surface wear was initiated by disruption of oxide layer and followed by the wear of oxygen diffusion zone, which was locally removed by cracking at extended testing times. The Al2O3 balls were also subjected to wear during testing. The wear scars formed on the balls are depicted in Fig. 11. Optical microscopic examinations revealed that, heavy wear of the ball on untreated surface progressed by grain pull out mechanism. The topogra-
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
phy of wear scar developed on the ball that slid on the oxidized surface was almost similar with that of untreated surface but smaller by a factor of 16. Thus, the size of wear scar was proportional to the size of the wear track formed on the surface of the material being tested.
Fig. 9. Progress of wear track areas during corrosion–wear tests.
3331
4. Discussion Thermal oxidation at 600 C and 650 C modified the surface of the examined Ti–6Al–4V alloy by introducing an oxide layer (anatase and/or rutile) with an oxygen diffusion zone beneath it. Increase of surface hardness after thermal oxidation can be attributed to the hard oxide layer covering the surface and contribution of strains in the matrix due to dissolution of oxygen [16,17]. The thickness of the oxide layer and oxygen diffusion zone increased with increasing oxidation time. However, growth of the oxide layer during thermal oxidation was accompanied by surface roughening. The thickness and the roughness of the oxide layer increased in an accelerated manner, when oxidation was conducted at 650 C. This situation affirms the role of temperature on oxidation rate by activating diffusing species. On the other hand, accelerated oxidation rate due to increasing temperature may put the whole oxide layer in a more stressed condition as a result of intrinsic and thermal effects [18,19]. The formation of defective oxide structure provides easy diffusion paths for oxygen and metal ions allowing oxidation progress in an uncontrollable way. Since the corrosion resistance of thermally oxidized surface layers relies on their chemical stability and defect free structure, this observation may also be
Fig. 10. SEM micrographs of the wear tracks developed on (a) untreated and (b) oxidized surfaces. High magnification of the framed region is shown in the in micrograph (c).
ARTICLE IN PRESS 3332
& / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
Fig. 11. The wear scars that formed on Al2O3 balls during corrosion–wear testing of: (a) untreated, and (b) oxidized surfaces.
valid for explaining the corrosion behavior of oxidized samples because strong reducing acids would not dissolve heavy oxide layer on titanium [20]. Small structural defects (microcracks and micropores) may be responsible for the failure of oxidized surfaces [21]. During incubation period, aggressive ions penetrate towards the oxide–metal interface and steadily extend, undermining the oxide because of the galvanic coupling of the oxide and substrate metal. Oxide layers gradually detached from the surface as fragments. This observation points out the fact that corrosive solution has entered the interface through individual defects. As soon as the oxide fragments lost their contact with the substrate, they detached. The corrosion behavior of the samples that were oxidized for short times at 600 C differed to that of the other samples. The relatively thin oxide layer that covered these samples was a mixture of anatase and rutile. Anatase was readily attacked and dissolved in reducing acids whereas rutile was inert and resistant to attack [2,21]. Dissolution of anatase at the beginning of the corrosion test left some sites of oxidized samples unprotected and further corrosion continued with total removal of the oxide layer. Increasing oxidation time yielded higher amount of rutile in the oxidized surface, which increased protective nature of oxide layer against the corrosive environment. Less defective structures of oxide containing more rutile formed at 600 C after longer oxidation times. And this combination presented the most corrosion resistant oxide layer that formed on Ti–6Al–4V alloy. Corrosion–wear test results confirmed the superiority of the thermally oxidized surface over untreated alloy. Surface degradation of untreated alloy commenced with the onset of sliding action and large amount of wear debris was produced through the entire test duration. The oxide layer prevented the wear of Ti–6Al–4V alloy over a certain period of testing time. The removal of oxide layer and diffusion hardened zone occurred due to mechanical and chemical effects. Even though the oxidized (at 600 C for 60 h) surface exhibited the best corrosion resistance in HCl solution with its less
defective structure, it was subjected to mechanical stresses during corrosion–wear test. This situation caused the local breakdown of the oxide layer and allowed the corrosive solution to enter the oxide–oxygen diffusion zone interface. The oxide layer was gradually removed by the combined action of corrosion and wear together. Eventually, corrosion–wear continued on the oxygen diffusion zone by producing a smaller amount of wear debris. After extended testing times, cracking of oxygen diffusion zone caused local material removal, and thereafter the depth of the wear track reached to the substrate. When the Al2O3 ball contacted with the substrate, it was subjected to wear with the mechanism (grain pull out) which was observed during testing of untreated surface Apparently, the wear mechanism of Al2O3 ball on untreated surface was maintained on oxidized surface when the depth of the wear track reached to substrate. Grain pull out wear mechanism is inevitable when the adhesion stresses between the sliding surfaces tend to be greater than the cohesive strength of the materials [22]. It should be emphasized that, the extent of wear developed on the Al2O3 ball is controlled by the size of the contact area between the ball and the surface being tested.
5. Conclusion The Ti–6Al–4V alloy was thermally oxidized at 600 C and 650 C to produce corrosion and wear resistant surface layers. Oxide thickness and oxygen diffusion zone depth steadily increased with increasing oxidation temperature and time. Oxide layers were composed of anatase and rutile structures of TiO2. Roughening of oxidized surfaces was more severe after oxidation at 650 C. A significant increment in surface hardness (3500 to 9000 HV) was achieved due to formation of a hard oxide layer and an oxygen diffusion zone beneath it. According to results of corrosion tests carried out in an aggressive acidic solution (5 m HCl), oxidation at 600 C for 60 h produced the most corrosion resistant surface on Ti–6Al–4V alloy. Wear tests that were
ARTICLE IN PRESS & / Biomaterials 25 (2004) 3325–3333 H. Gulery uz, . . H. C - imenoglu
conducted in a mild corrosive solution (0.9% NaCl), indicated that thermal oxidation treatment performed at 600 C for 60 h prevented extensive corrosion–wear of Ti–6Al–4V alloy and increased wear resistance by a factor of 25.
[9]
[10]
Acknowledgements The authors gratefully acknowledge the support of the NATO TU PVD Coatings project for the equipment supplied to the Metallurgy and Materials Engineering Department of Istanbul Technical University, which was utilized in characterization work carried out in this study. One of the authors (H.G.) would like to thank The State Planning Organization of Turkey for the support of his MSc. study.
[11]
[12]
[13]
[14]
References [1] Brunette DM, Tengvall P, Textor M, Thomsen P. Titanium in medicine. Heidelberg: Springer; 2001. [2] Boyer R, Welsch G, Collings EW. Materials properties handbook: titanium alloys. Materials Park, OH: ASM International; 1994. [3] Donachie MJ. Titanium: a technical guide. Materials Park, OH: ASM International; 1989. [4] Gilbert JL, Buckley CA, Lautenschlager EP. Titanium oxide film fracture and repassivation: The effect of potential, pH and aeration. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its alloys: the material and biological issues, ASTM STP 1272. Philadelphia: ASTM; 1996. p. 199–214. [5] Komotori J, Lee BJ, Dong H, Dearnley PA. Corrosion response of surface engineered titanium alloys damaged by prior abrasion. Wear 2001;88–98:1–11. [6] Galliano F, Galvanetto E, Mischler S, Landolt D. Tribocorrosion behavior of plasma nitrided Ti–6Al–4V alloy in neutral NaCl solution. Surf Coat Technol 2001;145:121–31. [7] Long M, Rack HJ. Titanium alloys in total joint replacement—a materials science perspective. Biomaterials 1998;19:1621–39. [8] Kovacs P, Davidson JA. Chemical and electrochemical aspects of biocompatibility of titanium and its alloys. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its
[15] [16]
[17]
[18]
[19]
[20] [21]
[22]
3333
alloys: the material and biological issues, ASTM STP 1272. Philadelphia: ASTM; 1996. p. 163–77. Dearnley PA. A review of metallic, ceramic and surface-treated metals used for bearing surfaces in human joint replacements. Institution of Mechanical Engineers proceedings, Part H. J Eng Med 1999;213:107–35. Kofstad P. High temperature corrosion. Essex: Elsevier Applied Science; 1988. Mishra AK, Davidson JA, Poggie RA, Kovacs P, FitzGerald TJ. Mechanical and tribological properties and biocompatibility of diffusion hardened Ti–13Nb–13Zr—a new titanium alloy for surgical implants. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its alloys: the material and biological issues, ASTM STP 1272. Philadelphia: ASTM; 1996. p. 96–112. Lopez MF, Jimenez JA, Gutierrez A. Corrosion study of surfacemodified vanadium-free titanium alloys. Electrochim Acta 2003; 48:1395–401. Garcia-Alonso MC, Saldana L, Valles G, Gonzalez-Garrasco JL, Gonzalez-Cabrero J, Martinez ME, Gil-Garay E, Munuera L. In vitro corrosion behavior and osteoblast response of thermally oxidised Ti6Al4V alloy. Biomaterials 2003;24:19–26. . . H. Surface treatment of Ti–6Al–4V alloy by thermal Gulery uz oxidation. M.Sc. thesis, 2003. Standard practice for laboratory immersion corrosion testing of metals. ASTM G31–72. Philadelphia: ASTM; 1995. Weissman S, Shrier A. Strain distribution in oxidized alpha titanium crystals. In: Jaffee RI, Promisel NE, editors. The science, technology and application of titanium. London: Pergamon Press; 1968. p. 441–58. Borgioli F, Galvanetto E, Galliano FP, Bacci T. Air treatment of pure titanium by furnace and glow-discharge processes. Surf Coat Technol 2001;141:103–7. Coddet C, Chaze AM, Beranger G. Measurement of the adhesion of thermal oxide films: application to the oxidation of titanium. J Mater Sci 1987;22:2969–74. Schutze . M. Failure of oxide scales on advanced materials due to . presence of stresses. In: Saito Y, Onay B, Maruyama T, editors. High temperature corrosion of advanced materials and protective coatings. B.V. Amsterdam: Elsevier Science Publishers; 1992. p. 39–49. Baboian R. Corrosion tests and standards: application and interpretation. Philadelphia: ASTM Manual Series; 1995. Bloyce A, Qi P-Y, Dong H, Bell T. Surface modification of titanium alloys for combined improvements in corrosion and wear resistance. Surf Coat Technol 1998;107:125–32. Dearnley PA, Dahm KL, Cimenoglu H. The corrosion–wear behaviour of thermally oxidised Cp–Ti and Ti–6Al–4V. Wear, in press.