- Email: [email protected]
Electrochemical and surface characterization of a nickel–titanium alloy
Biomaterials 19 (1998) 761 — 769
Electrochemical and surface characterization of a nickel—titanium alloy D.J. Wever!, A.G. Veldhuizen!,*, J. de Vries", H.J. Busscher", D.R.A. Uges#, J.R. van Horn! ! Department of Orthopaedics, University Hospital of Groningen, P.O. box 30.001, 9700 RB Groningen, Netherlands " Laboratoria for Materia Technica, University of Groningen, P.O. box 30.001, 9700 RB Groningen, Netherlands # Department of Pharmacy, University Hospital of Groningen, Netherlands Received 25th April 1997; accepted 10th October 1997
Abstract For clinical implantation purposes of shape memory metals the nearly equiatomic nickel—titanium (NiTi) alloy is generally used. In this study, the corrosion properties and surface characteristics of this alloy were investigated and compared with two reference controls, AISI 316 LVM stainless steel and Ti6Al4V. The anodic polarization curves, performed in Hanks’ solution at 37°C, demonstrated a passive behaviour for the NiTi alloy. A more pronounced difference between the corrosion and breakdown potential, i.e. a better resistance to chemical breakdown of passivity was found for the NiTi alloy compared to AISI 316 LVM. X-ray electron spectroscopy (XPS) and scanning electron microscopy (SEM) were undertaken to study the elemental composition and structure of the surface films prior to, and after immersion in Hanks’ solution. The passive film on the NiTi alloy consists of a mainly TiO -based oxide with minimal amounts of nickel in the outermost surface layers. After immersion in Hanks’ solution the 2 growth of a calcium-phosphate layer was observed. The passive diffusion of nickel from the NiTi alloy, measured by atomic absorption spectrophotometry reduced significantly in time from an initial release rate of 14.5]10~7 lg cm~2 s~1 to a nickel release that could not detect anymore after 10 days. It is suggested that the good corrosion properties of the NiTi alloy and the related promising biological response, as reported in literature, may be ascribed to the presence of mainly a TiO -based surface layer and its 2 specific properties, including the formation of a calcium-phosphate layer after exposure to a bioenvironment. ( 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: NiTi; Memory metal; Corrosion; XPS; Nickel release
1. Introduction The almost equiatomic nickel—titanium alloy is unique in that it possesses a so-called shape memory. This shape memory effect involves the recovery of a certain deformation induced at low temperature and as a consequence, a controlled change of shape when the alloy is heated. This phenomenon of the NiTi alloy, first described by Buehler and Wang [1], is the result of a phase transition from the low-temperature phase martensite to the hightemperature phase austenite. In the low-temperature phase the alloy can easily be deformed plastically. When the alloy is exposed to temperatures above the transition temperature, the alloy changes to the high-temperature
* Corresponding author. Tel.: 0031 50 3612802; fax: 0031 503611737.
phase and will revert to its original shape. In this way, deformations even up to 8% strain can be completely recovered. The special properties of NiTi can be put to excellent use for various biomedical applications. For example, at present, the properties of the NiTi alloy are already being used clinically in wires for orthodontic tooth alignment, osteosynthesis staples, vena cava filters and other vascular applications [2—7]. In addition, the NiTi alloy has been suggested for use in the surgical correction of scoliosis [8, 9]. Scoliosis, a three-dimensional deformity of the trunk, is characterized by lateral deviation and axial rotation of the spine, usually accompanied by thoracic deformation. In serious cases, the deformity has to be corrected surgically. The properties of the NiTi alloy make it possible to achieve gradual controlled threedimensional correction during and after the operation. Because of the gradual continuous corrections that are
0142-9612/98/$19.00 ( 1998 Published by Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 7 ) 0 0 2 1 0 - X
762
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
applied in the pre- and post-operative phase, processes like spinal viscoelasticity, growth and bone remodelling can be used. In spite of the intriguing properties of the NiTi alloy and its possible interesting biomedical applications, the implantation of nickel containing materials in human body requires caution. Although nickel is nutritionally essential, it is well known that nickel is capable of eliciting toxic and allergic responses [10—12]. Therefore, in the biocompatibility estimation of NiTi implants, attention has been drawn to the potential nickel release, especially when such NiTi devices are implanted into young patients which results in a potentially longer host exposure. The good anti-corrosive properties and associated good tissue compatibility of titanium and its clinically used alloy Ti6Al4V can be ascribed to a chemically stable oxidized passivation layer. Because titanium has a high affinity for oxygen, these beneficial surface oxide films form spontaneously in an oxygenated environment [13—17]. Consequently, oxide films are present on the metal surface of NiTi alloys as well. Depending on its composition, structure, stability and thickness, this layer may act as a barrier against nickel diffusion. In addition, it is known that calcium-phosphate surface films are naturally formed on titanium alloys in a biological environment [17—19]. It is possible that this layer acts as a further barrier against ion diffusion. The aim of the present study is to determine: (1) The stability of the surface film formed on the NiTi alloy by electrochemical potentiodynamic measurements. (2) The elemental composition and thickness of the surface film formed on the NiTi samples by X-ray electron spectroscopy, studied prior to and after immersion in a physiological solution. (3) The passive diffusion of nickel from the NiTi samples immersed in a physiological solution measured by atomic absorption spectrophotometry. (4) The surface structure by scanning electron microscopy of the surface films. The results are compared with two clinical reference controls, i.e. the implantable grade of stainless steel, AISI 316 LVM, with a nickel content of 13—15% and the well-known Ti6Al4V alloy with superior biocompatibility. It is assumed that the knowledge about the surface characteristics and related corrosion behaviour of the NiTi alloy will contribute to the biocompatibility estimation of this particular alloy, and therefore will finally determine the clinical success of implanted NiTi devices.
&50 at.% nickel and &50 at.% titanium with minimal amounts of trace elements in the final alloy (total of carbon, oxygen and other trace elements (0.5 wt%). Memory properties of the alloy were activated by heating the alloy up to 450°C for 30 min. For the electrochemical potentiodynamic measurements and surface characterization studies, small disks with a diameter of 8 mm and a length of 5 mm were prepared. Cylinders with a diameter 8 mm and a length of 40 mm were prepared for the nickel release study. All samples were mechanically polished to a mean stylus surface roughness Ra between 0.1 and 0.2 lm. Thereafter, the samples were ultrasonically cleaned in an alkaline soap solution and deionized water and then passivated according to the guidelines of the supplier, followed by rinsing in deionized water. The samples of the reference materials, Ti6Al4V (ASTM F136-92 [20]) and AISI 316 LVM, stainless steel (ASTM F138-92 [21]) were obtained from Tisto (Dusseldorf, Germany) and Carpenter (Brussels, Belgium), respectively. The reference samples were prepared to similar dimensions as the NiTi samples. After the final polishing, the Ti6Al4V and AISI 316 LVM samples were ultrasonically cleaned and passivated in nitric acid (20—40 vol%) for 30 min, followed by rinsing in deionized water. Finally, all samples, including the reference samples were air dried, sterilized by conventional steam sterilization at 134°C for 30 min and packaged in aluminium foil to prevent surface contamination. 2.2. Potentiodynamic measurements Potentiodynamic tests were performed on NiTi, Ti6Al4V and AISI 316 LVM in a Hanks’ physiological solution on a Princeton Applied Research potentiostat (Model 173) according to ASTM G5-94 procedures [22]. The electrolyte composition of Hanks’ solution is summarized in Table 1. During the measuring period the solution was deaerated by bubbling with high purity nitrogen (1.7 l min~1) and held at a constant temperature of 37°C. A saturated calomel electrode (SCE) was used as a reference. The potentiodynamic measurements were made after 60 min immersion in the physiological solution at an open-circuit potential. The potential scan was started from a potential of 100 mV more cathodic than the measured corrosion potential. The scan rate was 0.17 mV min~1, toward the anodic direction until 1600 mV (SCE). 2.3. XPS measurements
2. Materials and methods 2.1. Test materials The NiTi alloy was obtained from AMT, Swiss metal (Herk-de-Stad, Belgium) and consisted of about
X-ray photoelectron spectroscopy (XPS) was performed on the NiTi and Ti6Al4V samples prior to and after 1, 3, and 17 days immersion in Hanks’ solution. A S-probe spectrometer (Surface Science Instruments, Mountain View, CA, USA) equipped with an aluminium
D.J. Wever et al. / Biomaterials 19 (1998) 761—769 Table 1 Composition of Hanks’ solution Component
Concentration (g l~1)
CaCl H O 2 2 KCl KH PO 2 4 MgCl 6H O 2 2 MgSO 7H O 2 2 NaCl NaHCO 3 Na HPO 2 4 D-glucose
0.18 0.40 0.06 0.10 0.10 8.00 0.35 0.48 1.00
763
release was determined for five solution samples of both alloys in which after 10 days Hanks’ solution was refreshed. In these volume units the nickel concentration was measured 1 day after the refreshment. Blank solutions were subjected to the same procedures as used for the experimental samples. 2.5. Scanning electron microscopy The surface of the NiTi samples prior to and after 17 days immersion in a Hanks’ solution was examined in a Jeol Field Emission Scanning Electron Microscope, type 6301F operating at 5 kV.
anode (10 kV, 22 mA) and a quartz monochromator was used. The direction of the photoelectron detection angle was 60° with normal to the sample. Broad scan spectra were made with a 250]1000 lm spot and a pass energy of 150 eV while narrow scans were performed of C , O , 14 14 Ti and Ni using a pass energy of 50 eV. The 21 21(3@2) binding energy scale was calibrated to the C peak at 14 284.8 eV. The experimental peaks were integrated after linear background subtraction and the peaks were decomposed assuming a Gaussian/Lorentzian ratio 85/15 by using the SSI PC software package. Elemental surface compositions were calculated from the integrated peak areas employing instrumental sensitivity factors as supplied by the manufacturer and expressed in atomic %. Elemental depth profiling of the surface films of both alloys, prior to and after immersion in the physiological solution, was determined by sputtering off the films using a VG ion gun AG 2.1. The argon ion acceleration was 10 keV at a partial argon pressure of 5]10~7 mbar, resulting in a sputtering rate of 0.053 nm s~1 for titanium oxide films.
3. Results
2.4. Atomic absorption spectrophotometry
3.2. XPS measurements
A Varian atomic absorption spectrophotometer (model SpectrAA 300 plus) with graphite furnace GTA (Zeeman 96 plus) was used for the determination of the passive release rate of nickel out of the NiTi and AISI 316 LVM samples in Hanks’ solution. One millilitre Hanks’ solution with, respectively, 0, 5, 10, 20, 30, or 40 ll of a 0.5 mg l~1 nickel solution added was used for calibration. The minimal detection limit of nickel in the solution was 2.5 lg l~1. The NiTi and AISI 316 LVM samples were stored in Hanks’ solution, one cylinder per volume unit of 100 ml. The solution samples were placed in a 37°C water bath and were moved gently. To determine the nickel release rate, the average nickel concentration of five solution samples in each case was determined after 1, 2, 3, 7, 10, 17 and 31 days. To exclude saturation effects of Hanks’ solution the nickel
Tabel 2 summarizes the atomic percentages and electron binding energies of the surface elements found on the NiTi and Ti6Al4V samples studied prior to immersion in a Hanks’ solution. The surfaces of both alloys showed high atomic percentages of titanium, oxygen and carbon, and only small percentages of nickel were detected on the surface of the NiTi samples. The main component of the O peak at 530.1 eV was indicative for 14 oxygen bound in TiO . Other small components were 2 due to oxygen in C"O and O—C"O bonds. The Ti 21 peak had its main component at 458.8 eV, which also corresponds to TiO . No components were found indica2 tive of metallic titanium, usually found around 454.9 eV. Two peaks of Ni could be distinguished on the NiTi 21(3@2) samples. The main component at 856.2 eV corresponds with Ni O , while the other component at 852.6 eV can 2 3
3.1. Potentiodynamic measurements The potentiodynamic polarization curves of the NiTi, Ti6Al4V and AISI 316 LVM samples are shown in Fig 1. In Hanks’ solution at 37°C all samples exhibited passive behaviour. Ti6Al4V showed this passive behaviour over the entire potential range; the passive current density only increased slightly at higher potentials. The curves of the NiTi and AISI 316 LVM samples demonstrated an obvious breakdown potential, indicating a chemical breakdown of the passive film at this potential. Inspection of the samples after the polarization showed pitting of the surfaces of the AISI 316 LVM and NiTi samples, while the surface of the Ti6Al4V samples appeared to be intact. The difference between corrosion and breakdown potential, was more pronounced for the NiTi samples compared to the AISI 316 LVM samples, i.e. that the NiTi samples were more resistant to chemical breakdown of passivity than the AISI 316 LVM samples.
764
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
Fig. 1. Anodic polarization curves in a deaerated Hanks’ solution at 37°C of the NiTi, Ti6Al4V and AISI 316 LVM samples.
Table 2 Atomic percentages and bindings energies of the elements found in the Ti6Al4V samples and NiTi samples prior immersion in Hank’s solution (t"0) when analysed by XPS Sample
Element
Atomic percentage (%)
Bindings energy (eV) main components
Ti6Al4V
C 14 O 14 Ti 21 Ni 21(3@2) C 14 O 14 Ti 21 Ni 21(3@2)
60.3$15.02! 31.3$14.17! 7.3$2.84! —
284.8 530.1 458.8 —
51.3$5.83! 38.2$6.49! 8.2$1.92! 0.1$0.13!
284.8 530.1 458.8 856.2
NiTi
! Mean$standard deviation (five samples).
be assigned to metallic nickel. It was demonstrated that most carbon was involved in C—C bonds (284.8 eV), i.e. originating mainly from atmospheric contamination. No significant differences between the C , O , and Ti 14 14 21 peaks could be found between the NiTi and Ti6Al4V samples, indicating that the outermost surface layers of the NiTi alloy consist of TiO -based passivation layer 2 with minimal amounts of nickel. Figure 2a—f shows the XPS element depth profiles for Ti, Ni, C, O, Ca and P of the NiTi samples prior to (t"0)
and after 1, 3, and 17 days immersion in a Hanks’ solution obtained by sputtering the surface with Ar` ions. The titanium and oxygen depth profiles (Fig. 2a and c) of the samples prior to immersion in Hanks’ solution were typical for a surface oxide layer; The oxygen signal decreased, while the titanium signal increased towards the bulk metal. The depth at which the oxygen signal decreased with 50% of its maximum value was reached after 6 min. Assuming a constant sputtering rate of 0.053 nm s~1, an oxide thickness of approximately 19 nm was calculated. The nickel signal (Fig. 2b), hardly detected in the outermost surface, also increased during the Ar` ion sputter but reached the maximum values not as fast as titanium. XPS depth profiles of the surfaces after immersion in Hanks’ solution showed the formation of a calcium-phosphate layer on the NiTi samples (Fig. 2e and f ). An increase in Ca and P atomic percentages was found with longer immersion in Hanks’ solution. During the formation of this calcium-phosphate layer a significant change in the titanium depth profile was found. The change in slope of the curve was probably due to the oxide formation, while the change in its location was because of the growing calcium-phosphate layer on the top of the oxide layer (Fig. 2a). In contrast to titanium, the depth profiles of oxygen and nickel did not change significantly during immersion in Hanks’ solution (Fig. 2b and c). The carbon surface contamination layer, found in smaller percentages after immersion in Hanks’
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
765
Fig. 2. XPS depth profiles of the NiTi samples prior to, and after immersion in a Hank’s solution; showing the depth distribution of titanium (Fig. 2a), nickel (Fig. 2b), oxygen (Fig. 2c), carbon (Fig. 2d), calcium (Fig. 2e) and phosphorus (Fig. 2f ) versus sputtering rate (—L—) t"0, (—n—) t"1 day (—e—) t"3 days, (—K—) t"17 days.
solution disappeared almost immediately after Ar` ion sputtering (Fig. 2d). 3.3. Atomic absorption spectrophotometry The nickel release with the NiTi samples obtained by atomic absorption spectrophotometry reduced in time during the measuring period from an initial release rate of 14.5]10~7 lg cm~2 s~1 to a nickel release that could not be detected anymore after 10 days (Fig. 3). Nickel release was not demonstrated in the solution samples with the NiTi alloy in which the Hanks solution after 10 days was refreshed. Therefore, the decrease in nickel release cannot be due to saturation effects of Hanks’ solution. In contrast to the NiTi samples the nickel release of the AISI 316 LVM alloy remained under the detection limit (2.5 lg l~1) during the measuring period. Pitting corrosion was not observed after immersion of the both samples in Hanks’ solution.
3.4. Scanning electron microscopy Typical SEM micrographs of a NiTi sample prior to, and after 17 days immersion in a Hanks’ solution were shown in Fig 4a and b, respectively. In the micrograph of the NiTi surface before immersion, several dried in water droplets from the sterilization procedure can be seen. The micrograph of the NiTi surface after immersion showed an irregular appearance of the calcium-phosphate layer with varying thickness.
4. Discussion Assessing corrosion and surface characteristics is essential in determining the biological response of metallic implant materials. In our study, the NiTi alloy exhibited passive electrochemical behaviour comparable to other clinically used alloys. A more pronounced difference
766
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
Fig. 2. (continued).
between the corrosion and breakdown potential, i.e. a better resistance to chemical breakdown of passivity, was found for the NiTi alloy, compared to AISI 316 LVM stainless steel. As expected, the Ti6Al4V alloy demonstrated a superior corrosion resistance without pitting. XPS analyses of the NiTi alloy before immersion in physiological solution confirm the presence of a passive surface film on the NiTi alloy. This passive film consists of a mainly TiO -based oxide, comparable with Ti6Al4V. 2 Minimal amounts of nickel were found in the outermost surface layers. After immersion of the NiTi alloy in a physiological solution a calcium-phosphate layer was formed on the surface oxide. During the formation of this layer the titanium depth profiles changed significantly. This change could be attributed to both the thickening of the titanium oxide and the formation of a calcium-phosphate layer. Both in vitro studies and implantation stud-
ies have revealed that titanium-based alloys naturally form a calcium-phosphate layer on their passive oxide film after exposure to a bioenvironment [17—19, 23]. It is suggested that the formation and growth of this layer is governed by the existence of titanium oxide. As with any corrosion resistant alloy, despite the presence of stable oxide film, a small release of ions will still be present [13]. The rate of this passive release and the toxic activity of the released ions play an important role in the potential toxicity of an implant material. Because of the known toxic activity of nickel, the release of nickel out of the NiTi alloy should be within acceptable limits to prevent toxic responses [10—12]. The passive release rate of nickel from the NiTi alloy in a physiological solution was determined by atomic absorption spectrophotometry. Contrary to AISI 316 LVM stainless steel, a measurable nickel release from the NiTi alloy was detected, which significantly reduced within 10 days.
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
767
Fig. 3. Mean nickel release rate (lg/cm2 s) versus time (days) of the NiTi samples in Hanks’ solution, determined by atomic absorption spectrophotometry (error bar"standard deviation over five samples).
A similar time-related decrease was observed for the in vitro passive titanium and aluminium release of the Ti6Al4V alloy by others [24—28]. It has been shown in literature that the ion release of titanium and the Ti6AL4V alloy is influenced by the thickness and structure of the surface oxide layer. Increasing the thickness of the surface oxide layer of titanium alloys by thermal oxidation or electrochemical preparation improved the corrosion properties [24, 27, 29, 30]. Moreover, Lowenberg et al. observed an increase of trace element release after ASTM-F86 passivation of Ti6Al4V compared to non-passivated Ti6Al4V [25, 26]. Surface analysis revealed that this passivation method, which was originally developed for stainless steel and the CoCr alloy, reduced the thickness of the TiO film of 2 Ti6Al4V while the concentration of the metallic components in the outer most surface was increased. Therefore, to extend the comparison to the NiTi alloy, it is assumed that the initial nickel release from the NiTi alloy is controlled by the concentration of nickel in the outermost surface layers, as well as the thickness and structure of the TiO -based oxide film on the alloy. It was reported 2 that the nickel surface concentration of the NiTi alloy can be varied in the range of 0—30 at.% depending on the preparation method [31]. This means that passivation
methods and other follow-up treatments are crucial in the initial nickel release and potential biological response to NiTi implants. Various models have been described in literature dealing with surface modification of titanium alloys along with the dissolution process during implantation [23, 27, 28, 32]. It was postulated that the specific reaction of TiO with elements of the bioenvironment as calcium 2 and phosphorus may serve as a further barrier against ion diffusion. The simultaneous occurrence of a reduced time-related nickel release and the formation of the calcium-phosphate layer on the NiTi alloy in the physiological solution, observed in our study, supports this hypothesis. It would be an interesting topic for future research to find out whether this decrease in nickel release during exposure to the bioenvironment also occurs for in vivo situations and whether the process may be influenced by different passivation or other surface modifying techniques. The possible toxic effects of the observed initial passive nickel release of the NiTi samples have already been studied by performing short-term biocompatibility tests according to ISO regulations [33]. In the extract tests, the NiTi samples, treated as described in this study, provoked no cytotoxic, allergic and genotoxic responses.
768
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
Fig. 4. Typical SEM micrographs showing the surface topography of a NiTi sample prior to (Fig. 4a; bar equals 10 lm), and after 17 days immersion in a Hanks’ solution (Fig. 4b; bar equals 2 lm).
This promising biological response was also observed in other cytotoxicity tests as well as in several implantation studies: tissue reactions towards the NiTi alloy were indistinguishable from other implantable grade metals [10, 11, 34, 35]. These results confirm that the NiTi alloy can be regarded as a biological safe implant material. However, it should be emphasized that only the passive electrochemical nickel release of the NiTi alloy and the biological response to this release was studied. In clinical applica-
tions, especially in orthopaedic implant conditions, localized mechanical stress or fretting may damage the surface layer and can cause an increase in ion release or even a release of wear particles. Just as with other implantable grade metals, the extent of this wear related corrosion, as well as the implant location and other environmental conditions will ultimately determine the biological reaction of an implanted NiTi device [36, 37]. In conclusion, the results presented in this study revealed good corrosion resistance of the NiTi alloy
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
compared to other implantable grade metals. It is suggested that these good corrosion properties and the related promising biological response, as reported in literature, may be ascribed to the presence of a mainly TiO -based 2 surface layer and its specific properties, including the formation of a calcium-phosphate layer after exposure to a bioenvironment.
Acknowledgements This work was supported by STW (Netherlands Technology Foundation). We gratefully acknowledge M.M. Sanders, from ArcoMed BV, Rotterdam, for providing the samples and for useful discussion, J. Ijmker, from the Pharmacy Department of the University Hospital Groningen for assistance of the atomic absorption measurements and TEG Daenen and P.H. Joosten from Philips PMF, Eindhoven, for assistance of the electrochemical measurements.
References [1] Buehler WJ, Wang FE. A summary of recent research on nitinol alloys and their potential application. Ocean Eng 1968;1:105. [2] Andreasen G. A clinical trial of alignment of teeth using 0.019 inch thermal nitinol wire with transitional temperature range between 31°C and 45°C. Am J Orthod 1980;78:528—37. [3] Bensman G, Baumgart F, Haaster J. Osteosyntheseklammern aus nickel titan: herstellung, versuche und klinischer einsatz. Tech Mitt Forsch Ber 1982;40:123—34. [4] Cragg AH, De Jong SC, Barnhart WH, Landas SK, Smith TP. Nitinol intra vascular stent: results of preclinical evaluation. Radiology 1993;189:775—8. [5] Kambic H, Sutton C, Oku T et al. Biological performance of TiNi shape memory alloy vascular ring protheses: a two year study. Int J Artif Organs 1988;11:487—92. [6] Prince MR, Salzman EW, Schoen FJ, Palestrant AM, Simon M. Local intravascular effects of the nitinol blood clot filter. Invest Radiol 1988;23:249—300. [7] Simon M, Athanasoulis CA, Kim D et al. Simon nitinol inferior vena cava filter: initial clinical experience: work in progress. Radiology 1989;172:99—103. [8] Sanders JO, Sanders AE, More R, Ashman RB. A preliminary investigation of shape memory alloys in the surgical correction of scoliosis. Spine 1993;18:1640—46. [9] Veldhuizen AG, Sanders MM, Cool JC. A scoliosis correction device based on memory metal. Med Eng Phys 1997;19:171—9. [10] Castleman LS, Motzkin SM. The biocompatibility of nitinol. In: Williams DF, editor. Biocompatibility of clinical implant materials. Boca Raton, USA: CRC press, 1981:1129—54. [11] Putters JL, Kaulesar Sukul DM, Zeeuw de GR, Bijma A, Besseling PA. Comparative cell culture effects of shape memory metal (Nitinol), nickel and titanium: a biocompatibility estimation. Eur Surg Res 1992;24:378—82. [12] McKay GC, Macnair R, MacDonald C, MH Grant. Interactions of orthopaedic metals with an immortalized rat osteoblast cell line. Biomaterials 1996;17:1339—44. [13] Williams DF. Corrosion of orthopaedic implants. In: Williams DF, editor. Biocompatibility of orthopaedic implants. Boca Raton, USA: CRC Press, 1982:197—229.
769
[14] Ameen AP, Short RD, Johns R, Schwach G. The surface analysis of implant materials. 1. The surface composition of a titanium dental implant material. Clin Oral Implants Res 1993;4:144—50. [15] Lausmaa J, Kasemo B. Surface spectroscopic characterization of titanium implant materials. Appl Surf Sci 1990;44:133—46. [16] Ask M, Lausmaa J, Kasemo B. Preparation and surface spectroscopic characterization of oxide films on Ti6Al4V. Appl Surf Sci 1989;35:283—301. [17] Demri B, Hage-Ali M, Moritz M, Muster D. Surface characterization of C/Ti-6Al-4V coating treated with ion beam. Biomaterials 1997;18:305—10. [18] Hanawa T, Ota M. Characterization of surface film formed on titanium in electrolyte using XPS. Appl Surf Sci 1992;55:269—76. [19] Hanawa T, Ota M. Calcium phosphate naturally formed on titanium in electrolyte solution. Biomaterials 1991;12:767—74. [20] Standard specification for wrought titanium 6Al-4V ELI alloy for surgical implant applications. F136—92. Annual Book of ASTM Standards, vol. 04.12:1993;19—21. [21] Standard specification for stainless steel bar and wire for surgical implants. F138-92. Annual Book of ASTM Standards, vol. 04. 12:1993;22—24. [22] Standard reference test method for making potentiostatic and potentiodynamic anodic polarization measurements. G5-94. Annual Book of ASTM Standards, vol. 01.11:1994;63—73. [23] Sundgren JE, Bodo¨ P, Lundstro¨m I. Auger electron spectroscopic studies of the interface between human tissue and implants of titanium and stainless steel. J Colloid Interface Sci 1986;110:9—20. [24] Wisbey A, Gregson PJ, Peter LM, Tuke M. Effect of surface treatment on the dissolution of titanium-based implant materials. Biomaterials 1991;12:470—3. [25] Lowenberg BF, Lugowski S, Chipman M, Davies JE. ASTM-F86 passivation increases trace element release from Ti6AL4V into culture medium. J Mater Sci Mater Med 1994;5:467—72. [26] Callen BW, Lowenberg BF, Lugowski S, Sodhi RN, Davies JE. Nitric acid passivation of Ti6Al4V reduces thickness of surface oxide layer and increases trace element release. J Biomed Mater Res 1995;29:279—90. [27] Browne M, Gregson PJ. Surface modification of titanium alloy implants. Biomaterials 1994;15:894—8. [28] Healy KE, Ducheyne P. The mechanisms of passive dissolution of titanium in a model physiological environment. J Biomed Mater Res 1992;26:319—38. [29] Cigada A, Cabrini M, Pedeferri P. Increasing of the corrosion resistance of the Ti6Al4V alloy by high thickness anodic oxidation. J Mater Sci Mater Med 1992;3:408—12. [30] Shirkhanzadeh M. Electrochemical preparation of protective oxide coatings on titanium surgical alloys. J Mater Sci Mater Med 1992;3:322—5. [31] Shabalovskaya SA, Anderegg JW. Surface spectroscopic characterization of NiTi nearly equiatomic shape memory alloys for implants. J Vac Sci Technol 1995;13:2624—32. [32] Ellingsen JE. A study on the mechanism of protein adsorption to TiO2. Biomaterials 1991;12:593—6. [33] Wever DJ, Veldhuizen AG, Sanders MM, Schakenraad JM, Horn v JR. The cytotoxic, allergic and genotoxic activity of a nickel—titanium alloy. Biomaterials 1997;18:1115—20. [34] Castleman LS, Motzkin SM, Alicandri FP, Bonawit VL. Biocompatibility of Nitinol alloy as an implant material. Biomed Res 1976;10:695—731. [35] Cutwright DE, Bhaskar SN, Perez B, Johnson RM, Cowan GSM Tissue reaktion to Nitinol wire alloy. Oral Surg 1973;35: 578—84. [36] Galante JO, Lemons J, Spector M, Wilson PD, Wright TM. The biologic effects of implant materials. J Orthop Res 1991;9: 760—75. [37] Williams DF. Titanium: epitome of biocompatibility or cause for concern. J Bone Joint Surg (Br) 1994;76:348—9.
Electrochemical and surface characterization of a nickel—titanium alloy D.J. Wever!, A.G. Veldhuizen!,*, J. de Vries", H.J. Busscher", D.R.A. Uges#, J.R. van Horn! ! Department of Orthopaedics, University Hospital of Groningen, P.O. box 30.001, 9700 RB Groningen, Netherlands " Laboratoria for Materia Technica, University of Groningen, P.O. box 30.001, 9700 RB Groningen, Netherlands # Department of Pharmacy, University Hospital of Groningen, Netherlands Received 25th April 1997; accepted 10th October 1997
Abstract For clinical implantation purposes of shape memory metals the nearly equiatomic nickel—titanium (NiTi) alloy is generally used. In this study, the corrosion properties and surface characteristics of this alloy were investigated and compared with two reference controls, AISI 316 LVM stainless steel and Ti6Al4V. The anodic polarization curves, performed in Hanks’ solution at 37°C, demonstrated a passive behaviour for the NiTi alloy. A more pronounced difference between the corrosion and breakdown potential, i.e. a better resistance to chemical breakdown of passivity was found for the NiTi alloy compared to AISI 316 LVM. X-ray electron spectroscopy (XPS) and scanning electron microscopy (SEM) were undertaken to study the elemental composition and structure of the surface films prior to, and after immersion in Hanks’ solution. The passive film on the NiTi alloy consists of a mainly TiO -based oxide with minimal amounts of nickel in the outermost surface layers. After immersion in Hanks’ solution the 2 growth of a calcium-phosphate layer was observed. The passive diffusion of nickel from the NiTi alloy, measured by atomic absorption spectrophotometry reduced significantly in time from an initial release rate of 14.5]10~7 lg cm~2 s~1 to a nickel release that could not detect anymore after 10 days. It is suggested that the good corrosion properties of the NiTi alloy and the related promising biological response, as reported in literature, may be ascribed to the presence of mainly a TiO -based surface layer and its 2 specific properties, including the formation of a calcium-phosphate layer after exposure to a bioenvironment. ( 1998 Published by Elsevier Science Ltd. All rights reserved Keywords: NiTi; Memory metal; Corrosion; XPS; Nickel release
1. Introduction The almost equiatomic nickel—titanium alloy is unique in that it possesses a so-called shape memory. This shape memory effect involves the recovery of a certain deformation induced at low temperature and as a consequence, a controlled change of shape when the alloy is heated. This phenomenon of the NiTi alloy, first described by Buehler and Wang [1], is the result of a phase transition from the low-temperature phase martensite to the hightemperature phase austenite. In the low-temperature phase the alloy can easily be deformed plastically. When the alloy is exposed to temperatures above the transition temperature, the alloy changes to the high-temperature
* Corresponding author. Tel.: 0031 50 3612802; fax: 0031 503611737.
phase and will revert to its original shape. In this way, deformations even up to 8% strain can be completely recovered. The special properties of NiTi can be put to excellent use for various biomedical applications. For example, at present, the properties of the NiTi alloy are already being used clinically in wires for orthodontic tooth alignment, osteosynthesis staples, vena cava filters and other vascular applications [2—7]. In addition, the NiTi alloy has been suggested for use in the surgical correction of scoliosis [8, 9]. Scoliosis, a three-dimensional deformity of the trunk, is characterized by lateral deviation and axial rotation of the spine, usually accompanied by thoracic deformation. In serious cases, the deformity has to be corrected surgically. The properties of the NiTi alloy make it possible to achieve gradual controlled threedimensional correction during and after the operation. Because of the gradual continuous corrections that are
0142-9612/98/$19.00 ( 1998 Published by Elsevier Science Ltd. All rights reserved. PII S 0 1 4 2 - 9 6 1 2 ( 9 7 ) 0 0 2 1 0 - X
762
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
applied in the pre- and post-operative phase, processes like spinal viscoelasticity, growth and bone remodelling can be used. In spite of the intriguing properties of the NiTi alloy and its possible interesting biomedical applications, the implantation of nickel containing materials in human body requires caution. Although nickel is nutritionally essential, it is well known that nickel is capable of eliciting toxic and allergic responses [10—12]. Therefore, in the biocompatibility estimation of NiTi implants, attention has been drawn to the potential nickel release, especially when such NiTi devices are implanted into young patients which results in a potentially longer host exposure. The good anti-corrosive properties and associated good tissue compatibility of titanium and its clinically used alloy Ti6Al4V can be ascribed to a chemically stable oxidized passivation layer. Because titanium has a high affinity for oxygen, these beneficial surface oxide films form spontaneously in an oxygenated environment [13—17]. Consequently, oxide films are present on the metal surface of NiTi alloys as well. Depending on its composition, structure, stability and thickness, this layer may act as a barrier against nickel diffusion. In addition, it is known that calcium-phosphate surface films are naturally formed on titanium alloys in a biological environment [17—19]. It is possible that this layer acts as a further barrier against ion diffusion. The aim of the present study is to determine: (1) The stability of the surface film formed on the NiTi alloy by electrochemical potentiodynamic measurements. (2) The elemental composition and thickness of the surface film formed on the NiTi samples by X-ray electron spectroscopy, studied prior to and after immersion in a physiological solution. (3) The passive diffusion of nickel from the NiTi samples immersed in a physiological solution measured by atomic absorption spectrophotometry. (4) The surface structure by scanning electron microscopy of the surface films. The results are compared with two clinical reference controls, i.e. the implantable grade of stainless steel, AISI 316 LVM, with a nickel content of 13—15% and the well-known Ti6Al4V alloy with superior biocompatibility. It is assumed that the knowledge about the surface characteristics and related corrosion behaviour of the NiTi alloy will contribute to the biocompatibility estimation of this particular alloy, and therefore will finally determine the clinical success of implanted NiTi devices.
&50 at.% nickel and &50 at.% titanium with minimal amounts of trace elements in the final alloy (total of carbon, oxygen and other trace elements (0.5 wt%). Memory properties of the alloy were activated by heating the alloy up to 450°C for 30 min. For the electrochemical potentiodynamic measurements and surface characterization studies, small disks with a diameter of 8 mm and a length of 5 mm were prepared. Cylinders with a diameter 8 mm and a length of 40 mm were prepared for the nickel release study. All samples were mechanically polished to a mean stylus surface roughness Ra between 0.1 and 0.2 lm. Thereafter, the samples were ultrasonically cleaned in an alkaline soap solution and deionized water and then passivated according to the guidelines of the supplier, followed by rinsing in deionized water. The samples of the reference materials, Ti6Al4V (ASTM F136-92 [20]) and AISI 316 LVM, stainless steel (ASTM F138-92 [21]) were obtained from Tisto (Dusseldorf, Germany) and Carpenter (Brussels, Belgium), respectively. The reference samples were prepared to similar dimensions as the NiTi samples. After the final polishing, the Ti6Al4V and AISI 316 LVM samples were ultrasonically cleaned and passivated in nitric acid (20—40 vol%) for 30 min, followed by rinsing in deionized water. Finally, all samples, including the reference samples were air dried, sterilized by conventional steam sterilization at 134°C for 30 min and packaged in aluminium foil to prevent surface contamination. 2.2. Potentiodynamic measurements Potentiodynamic tests were performed on NiTi, Ti6Al4V and AISI 316 LVM in a Hanks’ physiological solution on a Princeton Applied Research potentiostat (Model 173) according to ASTM G5-94 procedures [22]. The electrolyte composition of Hanks’ solution is summarized in Table 1. During the measuring period the solution was deaerated by bubbling with high purity nitrogen (1.7 l min~1) and held at a constant temperature of 37°C. A saturated calomel electrode (SCE) was used as a reference. The potentiodynamic measurements were made after 60 min immersion in the physiological solution at an open-circuit potential. The potential scan was started from a potential of 100 mV more cathodic than the measured corrosion potential. The scan rate was 0.17 mV min~1, toward the anodic direction until 1600 mV (SCE). 2.3. XPS measurements
2. Materials and methods 2.1. Test materials The NiTi alloy was obtained from AMT, Swiss metal (Herk-de-Stad, Belgium) and consisted of about
X-ray photoelectron spectroscopy (XPS) was performed on the NiTi and Ti6Al4V samples prior to and after 1, 3, and 17 days immersion in Hanks’ solution. A S-probe spectrometer (Surface Science Instruments, Mountain View, CA, USA) equipped with an aluminium
D.J. Wever et al. / Biomaterials 19 (1998) 761—769 Table 1 Composition of Hanks’ solution Component
Concentration (g l~1)
CaCl H O 2 2 KCl KH PO 2 4 MgCl 6H O 2 2 MgSO 7H O 2 2 NaCl NaHCO 3 Na HPO 2 4 D-glucose
0.18 0.40 0.06 0.10 0.10 8.00 0.35 0.48 1.00
763
release was determined for five solution samples of both alloys in which after 10 days Hanks’ solution was refreshed. In these volume units the nickel concentration was measured 1 day after the refreshment. Blank solutions were subjected to the same procedures as used for the experimental samples. 2.5. Scanning electron microscopy The surface of the NiTi samples prior to and after 17 days immersion in a Hanks’ solution was examined in a Jeol Field Emission Scanning Electron Microscope, type 6301F operating at 5 kV.
anode (10 kV, 22 mA) and a quartz monochromator was used. The direction of the photoelectron detection angle was 60° with normal to the sample. Broad scan spectra were made with a 250]1000 lm spot and a pass energy of 150 eV while narrow scans were performed of C , O , 14 14 Ti and Ni using a pass energy of 50 eV. The 21 21(3@2) binding energy scale was calibrated to the C peak at 14 284.8 eV. The experimental peaks were integrated after linear background subtraction and the peaks were decomposed assuming a Gaussian/Lorentzian ratio 85/15 by using the SSI PC software package. Elemental surface compositions were calculated from the integrated peak areas employing instrumental sensitivity factors as supplied by the manufacturer and expressed in atomic %. Elemental depth profiling of the surface films of both alloys, prior to and after immersion in the physiological solution, was determined by sputtering off the films using a VG ion gun AG 2.1. The argon ion acceleration was 10 keV at a partial argon pressure of 5]10~7 mbar, resulting in a sputtering rate of 0.053 nm s~1 for titanium oxide films.
3. Results
2.4. Atomic absorption spectrophotometry
3.2. XPS measurements
A Varian atomic absorption spectrophotometer (model SpectrAA 300 plus) with graphite furnace GTA (Zeeman 96 plus) was used for the determination of the passive release rate of nickel out of the NiTi and AISI 316 LVM samples in Hanks’ solution. One millilitre Hanks’ solution with, respectively, 0, 5, 10, 20, 30, or 40 ll of a 0.5 mg l~1 nickel solution added was used for calibration. The minimal detection limit of nickel in the solution was 2.5 lg l~1. The NiTi and AISI 316 LVM samples were stored in Hanks’ solution, one cylinder per volume unit of 100 ml. The solution samples were placed in a 37°C water bath and were moved gently. To determine the nickel release rate, the average nickel concentration of five solution samples in each case was determined after 1, 2, 3, 7, 10, 17 and 31 days. To exclude saturation effects of Hanks’ solution the nickel
Tabel 2 summarizes the atomic percentages and electron binding energies of the surface elements found on the NiTi and Ti6Al4V samples studied prior to immersion in a Hanks’ solution. The surfaces of both alloys showed high atomic percentages of titanium, oxygen and carbon, and only small percentages of nickel were detected on the surface of the NiTi samples. The main component of the O peak at 530.1 eV was indicative for 14 oxygen bound in TiO . Other small components were 2 due to oxygen in C"O and O—C"O bonds. The Ti 21 peak had its main component at 458.8 eV, which also corresponds to TiO . No components were found indica2 tive of metallic titanium, usually found around 454.9 eV. Two peaks of Ni could be distinguished on the NiTi 21(3@2) samples. The main component at 856.2 eV corresponds with Ni O , while the other component at 852.6 eV can 2 3
3.1. Potentiodynamic measurements The potentiodynamic polarization curves of the NiTi, Ti6Al4V and AISI 316 LVM samples are shown in Fig 1. In Hanks’ solution at 37°C all samples exhibited passive behaviour. Ti6Al4V showed this passive behaviour over the entire potential range; the passive current density only increased slightly at higher potentials. The curves of the NiTi and AISI 316 LVM samples demonstrated an obvious breakdown potential, indicating a chemical breakdown of the passive film at this potential. Inspection of the samples after the polarization showed pitting of the surfaces of the AISI 316 LVM and NiTi samples, while the surface of the Ti6Al4V samples appeared to be intact. The difference between corrosion and breakdown potential, was more pronounced for the NiTi samples compared to the AISI 316 LVM samples, i.e. that the NiTi samples were more resistant to chemical breakdown of passivity than the AISI 316 LVM samples.
764
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
Fig. 1. Anodic polarization curves in a deaerated Hanks’ solution at 37°C of the NiTi, Ti6Al4V and AISI 316 LVM samples.
Table 2 Atomic percentages and bindings energies of the elements found in the Ti6Al4V samples and NiTi samples prior immersion in Hank’s solution (t"0) when analysed by XPS Sample
Element
Atomic percentage (%)
Bindings energy (eV) main components
Ti6Al4V
C 14 O 14 Ti 21 Ni 21(3@2) C 14 O 14 Ti 21 Ni 21(3@2)
60.3$15.02! 31.3$14.17! 7.3$2.84! —
284.8 530.1 458.8 —
51.3$5.83! 38.2$6.49! 8.2$1.92! 0.1$0.13!
284.8 530.1 458.8 856.2
NiTi
! Mean$standard deviation (five samples).
be assigned to metallic nickel. It was demonstrated that most carbon was involved in C—C bonds (284.8 eV), i.e. originating mainly from atmospheric contamination. No significant differences between the C , O , and Ti 14 14 21 peaks could be found between the NiTi and Ti6Al4V samples, indicating that the outermost surface layers of the NiTi alloy consist of TiO -based passivation layer 2 with minimal amounts of nickel. Figure 2a—f shows the XPS element depth profiles for Ti, Ni, C, O, Ca and P of the NiTi samples prior to (t"0)
and after 1, 3, and 17 days immersion in a Hanks’ solution obtained by sputtering the surface with Ar` ions. The titanium and oxygen depth profiles (Fig. 2a and c) of the samples prior to immersion in Hanks’ solution were typical for a surface oxide layer; The oxygen signal decreased, while the titanium signal increased towards the bulk metal. The depth at which the oxygen signal decreased with 50% of its maximum value was reached after 6 min. Assuming a constant sputtering rate of 0.053 nm s~1, an oxide thickness of approximately 19 nm was calculated. The nickel signal (Fig. 2b), hardly detected in the outermost surface, also increased during the Ar` ion sputter but reached the maximum values not as fast as titanium. XPS depth profiles of the surfaces after immersion in Hanks’ solution showed the formation of a calcium-phosphate layer on the NiTi samples (Fig. 2e and f ). An increase in Ca and P atomic percentages was found with longer immersion in Hanks’ solution. During the formation of this calcium-phosphate layer a significant change in the titanium depth profile was found. The change in slope of the curve was probably due to the oxide formation, while the change in its location was because of the growing calcium-phosphate layer on the top of the oxide layer (Fig. 2a). In contrast to titanium, the depth profiles of oxygen and nickel did not change significantly during immersion in Hanks’ solution (Fig. 2b and c). The carbon surface contamination layer, found in smaller percentages after immersion in Hanks’
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
765
Fig. 2. XPS depth profiles of the NiTi samples prior to, and after immersion in a Hank’s solution; showing the depth distribution of titanium (Fig. 2a), nickel (Fig. 2b), oxygen (Fig. 2c), carbon (Fig. 2d), calcium (Fig. 2e) and phosphorus (Fig. 2f ) versus sputtering rate (—L—) t"0, (—n—) t"1 day (—e—) t"3 days, (—K—) t"17 days.
solution disappeared almost immediately after Ar` ion sputtering (Fig. 2d). 3.3. Atomic absorption spectrophotometry The nickel release with the NiTi samples obtained by atomic absorption spectrophotometry reduced in time during the measuring period from an initial release rate of 14.5]10~7 lg cm~2 s~1 to a nickel release that could not be detected anymore after 10 days (Fig. 3). Nickel release was not demonstrated in the solution samples with the NiTi alloy in which the Hanks solution after 10 days was refreshed. Therefore, the decrease in nickel release cannot be due to saturation effects of Hanks’ solution. In contrast to the NiTi samples the nickel release of the AISI 316 LVM alloy remained under the detection limit (2.5 lg l~1) during the measuring period. Pitting corrosion was not observed after immersion of the both samples in Hanks’ solution.
3.4. Scanning electron microscopy Typical SEM micrographs of a NiTi sample prior to, and after 17 days immersion in a Hanks’ solution were shown in Fig 4a and b, respectively. In the micrograph of the NiTi surface before immersion, several dried in water droplets from the sterilization procedure can be seen. The micrograph of the NiTi surface after immersion showed an irregular appearance of the calcium-phosphate layer with varying thickness.
4. Discussion Assessing corrosion and surface characteristics is essential in determining the biological response of metallic implant materials. In our study, the NiTi alloy exhibited passive electrochemical behaviour comparable to other clinically used alloys. A more pronounced difference
766
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
Fig. 2. (continued).
between the corrosion and breakdown potential, i.e. a better resistance to chemical breakdown of passivity, was found for the NiTi alloy, compared to AISI 316 LVM stainless steel. As expected, the Ti6Al4V alloy demonstrated a superior corrosion resistance without pitting. XPS analyses of the NiTi alloy before immersion in physiological solution confirm the presence of a passive surface film on the NiTi alloy. This passive film consists of a mainly TiO -based oxide, comparable with Ti6Al4V. 2 Minimal amounts of nickel were found in the outermost surface layers. After immersion of the NiTi alloy in a physiological solution a calcium-phosphate layer was formed on the surface oxide. During the formation of this layer the titanium depth profiles changed significantly. This change could be attributed to both the thickening of the titanium oxide and the formation of a calcium-phosphate layer. Both in vitro studies and implantation stud-
ies have revealed that titanium-based alloys naturally form a calcium-phosphate layer on their passive oxide film after exposure to a bioenvironment [17—19, 23]. It is suggested that the formation and growth of this layer is governed by the existence of titanium oxide. As with any corrosion resistant alloy, despite the presence of stable oxide film, a small release of ions will still be present [13]. The rate of this passive release and the toxic activity of the released ions play an important role in the potential toxicity of an implant material. Because of the known toxic activity of nickel, the release of nickel out of the NiTi alloy should be within acceptable limits to prevent toxic responses [10—12]. The passive release rate of nickel from the NiTi alloy in a physiological solution was determined by atomic absorption spectrophotometry. Contrary to AISI 316 LVM stainless steel, a measurable nickel release from the NiTi alloy was detected, which significantly reduced within 10 days.
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
767
Fig. 3. Mean nickel release rate (lg/cm2 s) versus time (days) of the NiTi samples in Hanks’ solution, determined by atomic absorption spectrophotometry (error bar"standard deviation over five samples).
A similar time-related decrease was observed for the in vitro passive titanium and aluminium release of the Ti6Al4V alloy by others [24—28]. It has been shown in literature that the ion release of titanium and the Ti6AL4V alloy is influenced by the thickness and structure of the surface oxide layer. Increasing the thickness of the surface oxide layer of titanium alloys by thermal oxidation or electrochemical preparation improved the corrosion properties [24, 27, 29, 30]. Moreover, Lowenberg et al. observed an increase of trace element release after ASTM-F86 passivation of Ti6Al4V compared to non-passivated Ti6Al4V [25, 26]. Surface analysis revealed that this passivation method, which was originally developed for stainless steel and the CoCr alloy, reduced the thickness of the TiO film of 2 Ti6Al4V while the concentration of the metallic components in the outer most surface was increased. Therefore, to extend the comparison to the NiTi alloy, it is assumed that the initial nickel release from the NiTi alloy is controlled by the concentration of nickel in the outermost surface layers, as well as the thickness and structure of the TiO -based oxide film on the alloy. It was reported 2 that the nickel surface concentration of the NiTi alloy can be varied in the range of 0—30 at.% depending on the preparation method [31]. This means that passivation
methods and other follow-up treatments are crucial in the initial nickel release and potential biological response to NiTi implants. Various models have been described in literature dealing with surface modification of titanium alloys along with the dissolution process during implantation [23, 27, 28, 32]. It was postulated that the specific reaction of TiO with elements of the bioenvironment as calcium 2 and phosphorus may serve as a further barrier against ion diffusion. The simultaneous occurrence of a reduced time-related nickel release and the formation of the calcium-phosphate layer on the NiTi alloy in the physiological solution, observed in our study, supports this hypothesis. It would be an interesting topic for future research to find out whether this decrease in nickel release during exposure to the bioenvironment also occurs for in vivo situations and whether the process may be influenced by different passivation or other surface modifying techniques. The possible toxic effects of the observed initial passive nickel release of the NiTi samples have already been studied by performing short-term biocompatibility tests according to ISO regulations [33]. In the extract tests, the NiTi samples, treated as described in this study, provoked no cytotoxic, allergic and genotoxic responses.
768
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
Fig. 4. Typical SEM micrographs showing the surface topography of a NiTi sample prior to (Fig. 4a; bar equals 10 lm), and after 17 days immersion in a Hanks’ solution (Fig. 4b; bar equals 2 lm).
This promising biological response was also observed in other cytotoxicity tests as well as in several implantation studies: tissue reactions towards the NiTi alloy were indistinguishable from other implantable grade metals [10, 11, 34, 35]. These results confirm that the NiTi alloy can be regarded as a biological safe implant material. However, it should be emphasized that only the passive electrochemical nickel release of the NiTi alloy and the biological response to this release was studied. In clinical applica-
tions, especially in orthopaedic implant conditions, localized mechanical stress or fretting may damage the surface layer and can cause an increase in ion release or even a release of wear particles. Just as with other implantable grade metals, the extent of this wear related corrosion, as well as the implant location and other environmental conditions will ultimately determine the biological reaction of an implanted NiTi device [36, 37]. In conclusion, the results presented in this study revealed good corrosion resistance of the NiTi alloy
D.J. Wever et al. / Biomaterials 19 (1998) 761—769
compared to other implantable grade metals. It is suggested that these good corrosion properties and the related promising biological response, as reported in literature, may be ascribed to the presence of a mainly TiO -based 2 surface layer and its specific properties, including the formation of a calcium-phosphate layer after exposure to a bioenvironment.
Acknowledgements This work was supported by STW (Netherlands Technology Foundation). We gratefully acknowledge M.M. Sanders, from ArcoMed BV, Rotterdam, for providing the samples and for useful discussion, J. Ijmker, from the Pharmacy Department of the University Hospital Groningen for assistance of the atomic absorption measurements and TEG Daenen and P.H. Joosten from Philips PMF, Eindhoven, for assistance of the electrochemical measurements.
References [1] Buehler WJ, Wang FE. A summary of recent research on nitinol alloys and their potential application. Ocean Eng 1968;1:105. [2] Andreasen G. A clinical trial of alignment of teeth using 0.019 inch thermal nitinol wire with transitional temperature range between 31°C and 45°C. Am J Orthod 1980;78:528—37. [3] Bensman G, Baumgart F, Haaster J. Osteosyntheseklammern aus nickel titan: herstellung, versuche und klinischer einsatz. Tech Mitt Forsch Ber 1982;40:123—34. [4] Cragg AH, De Jong SC, Barnhart WH, Landas SK, Smith TP. Nitinol intra vascular stent: results of preclinical evaluation. Radiology 1993;189:775—8. [5] Kambic H, Sutton C, Oku T et al. Biological performance of TiNi shape memory alloy vascular ring protheses: a two year study. Int J Artif Organs 1988;11:487—92. [6] Prince MR, Salzman EW, Schoen FJ, Palestrant AM, Simon M. Local intravascular effects of the nitinol blood clot filter. Invest Radiol 1988;23:249—300. [7] Simon M, Athanasoulis CA, Kim D et al. Simon nitinol inferior vena cava filter: initial clinical experience: work in progress. Radiology 1989;172:99—103. [8] Sanders JO, Sanders AE, More R, Ashman RB. A preliminary investigation of shape memory alloys in the surgical correction of scoliosis. Spine 1993;18:1640—46. [9] Veldhuizen AG, Sanders MM, Cool JC. A scoliosis correction device based on memory metal. Med Eng Phys 1997;19:171—9. [10] Castleman LS, Motzkin SM. The biocompatibility of nitinol. In: Williams DF, editor. Biocompatibility of clinical implant materials. Boca Raton, USA: CRC press, 1981:1129—54. [11] Putters JL, Kaulesar Sukul DM, Zeeuw de GR, Bijma A, Besseling PA. Comparative cell culture effects of shape memory metal (Nitinol), nickel and titanium: a biocompatibility estimation. Eur Surg Res 1992;24:378—82. [12] McKay GC, Macnair R, MacDonald C, MH Grant. Interactions of orthopaedic metals with an immortalized rat osteoblast cell line. Biomaterials 1996;17:1339—44. [13] Williams DF. Corrosion of orthopaedic implants. In: Williams DF, editor. Biocompatibility of orthopaedic implants. Boca Raton, USA: CRC Press, 1982:197—229.
769
[14] Ameen AP, Short RD, Johns R, Schwach G. The surface analysis of implant materials. 1. The surface composition of a titanium dental implant material. Clin Oral Implants Res 1993;4:144—50. [15] Lausmaa J, Kasemo B. Surface spectroscopic characterization of titanium implant materials. Appl Surf Sci 1990;44:133—46. [16] Ask M, Lausmaa J, Kasemo B. Preparation and surface spectroscopic characterization of oxide films on Ti6Al4V. Appl Surf Sci 1989;35:283—301. [17] Demri B, Hage-Ali M, Moritz M, Muster D. Surface characterization of C/Ti-6Al-4V coating treated with ion beam. Biomaterials 1997;18:305—10. [18] Hanawa T, Ota M. Characterization of surface film formed on titanium in electrolyte using XPS. Appl Surf Sci 1992;55:269—76. [19] Hanawa T, Ota M. Calcium phosphate naturally formed on titanium in electrolyte solution. Biomaterials 1991;12:767—74. [20] Standard specification for wrought titanium 6Al-4V ELI alloy for surgical implant applications. F136—92. Annual Book of ASTM Standards, vol. 04.12:1993;19—21. [21] Standard specification for stainless steel bar and wire for surgical implants. F138-92. Annual Book of ASTM Standards, vol. 04. 12:1993;22—24. [22] Standard reference test method for making potentiostatic and potentiodynamic anodic polarization measurements. G5-94. Annual Book of ASTM Standards, vol. 01.11:1994;63—73. [23] Sundgren JE, Bodo¨ P, Lundstro¨m I. Auger electron spectroscopic studies of the interface between human tissue and implants of titanium and stainless steel. J Colloid Interface Sci 1986;110:9—20. [24] Wisbey A, Gregson PJ, Peter LM, Tuke M. Effect of surface treatment on the dissolution of titanium-based implant materials. Biomaterials 1991;12:470—3. [25] Lowenberg BF, Lugowski S, Chipman M, Davies JE. ASTM-F86 passivation increases trace element release from Ti6AL4V into culture medium. J Mater Sci Mater Med 1994;5:467—72. [26] Callen BW, Lowenberg BF, Lugowski S, Sodhi RN, Davies JE. Nitric acid passivation of Ti6Al4V reduces thickness of surface oxide layer and increases trace element release. J Biomed Mater Res 1995;29:279—90. [27] Browne M, Gregson PJ. Surface modification of titanium alloy implants. Biomaterials 1994;15:894—8. [28] Healy KE, Ducheyne P. The mechanisms of passive dissolution of titanium in a model physiological environment. J Biomed Mater Res 1992;26:319—38. [29] Cigada A, Cabrini M, Pedeferri P. Increasing of the corrosion resistance of the Ti6Al4V alloy by high thickness anodic oxidation. J Mater Sci Mater Med 1992;3:408—12. [30] Shirkhanzadeh M. Electrochemical preparation of protective oxide coatings on titanium surgical alloys. J Mater Sci Mater Med 1992;3:322—5. [31] Shabalovskaya SA, Anderegg JW. Surface spectroscopic characterization of NiTi nearly equiatomic shape memory alloys for implants. J Vac Sci Technol 1995;13:2624—32. [32] Ellingsen JE. A study on the mechanism of protein adsorption to TiO2. Biomaterials 1991;12:593—6. [33] Wever DJ, Veldhuizen AG, Sanders MM, Schakenraad JM, Horn v JR. The cytotoxic, allergic and genotoxic activity of a nickel—titanium alloy. Biomaterials 1997;18:1115—20. [34] Castleman LS, Motzkin SM, Alicandri FP, Bonawit VL. Biocompatibility of Nitinol alloy as an implant material. Biomed Res 1976;10:695—731. [35] Cutwright DE, Bhaskar SN, Perez B, Johnson RM, Cowan GSM Tissue reaktion to Nitinol wire alloy. Oral Surg 1973;35: 578—84. [36] Galante JO, Lemons J, Spector M, Wilson PD, Wright TM. The biologic effects of implant materials. J Orthop Res 1991;9: 760—75. [37] Williams DF. Titanium: epitome of biocompatibility or cause for concern. J Bone Joint Surg (Br) 1994;76:348—9.