- Email: [email protected]
Low modulus Ti–Nb–Hf alloy for biomedical applications
Materials Science and Engineering C 42 (2014) 691–695
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Low modulus Ti–Nb–Hf alloy for biomedical applications M. González a,b,⁎, J. Peña a,b, F.J. Gil a,c, J.M. Manero a,c a b c
Department of Materials Science and Metallurgy, Universitat Politècnica de Catalunya (UPC), Avda. Diagonal 647, 08028 Barcelona, Spain Materials Science, Elisava Escola Superior de Disseny i Enginyeria de Barcelona, La Rambla 30-32, 08002 Barcelona, Spain Ciber-BBN, Spain
a r t i c l e
i n f o
Article history: Received 20 November 2013 Received in revised form 13 April 2014 Accepted 9 June 2014 Available online 19 June 2014 Keywords: Corrosion Biocompatibility Cytotoxicity Low elastic modulus Stress shielding effect TiNbHf alloys
a b s t r a c t β-Type titanium alloys with a low elastic modulus are a potential strategy to reduce stress shielding effect and to enhance bone remodeling in implants used to substitute failed hard tissue. For biomaterial application, investigation on the mechanical behavior, the corrosion resistance and the cell response is required. The new Ti25Nb16Hf alloy was studied before and after 95% cold rolling (95% C.R.). The mechanical properties were determined by tensile testing and its corrosion behavior was analyzed by potentiostatic equipment in Hank's solution at 37 °C. The cell response was studied by means of cytotoxicity evaluation, cell adhesion and proliferation measurements. The stress–strain curves showed the lowest elastic modulus (42 GPa) in the cold worked alloy and high tensile strength, similar to that of Ti6Al4V. The new alloy exhibited better corrosion resistance in terms of open circuit potential (EOCP), but was similar in terms of corrosion current density (iCORR) compared to Ti grade II. Cytotoxicity studies revealed that the chemical composition of the alloy does not induce cytotoxic activity. Cell studies in the new alloy showed a lower adhesion and a higher proliferation compared to Ti grade II presenting, therefore, mechanical features similar to those of human cortical bone and, simultaneously, a good cell response. © 2014 Elsevier B.V. All rights reserved.
1. Introduction In the bone substitution metal implant field, one of the main challenges is to achieve materials that exhibit a low elastic modulus, similar to that of human cortical bone (10–30 GPa) [1]. The most commercial alloys used as implant materials present an elastic modulus in the range of 110 GPa of Ti6Al4V up to 220 GPa of cobalt–chrome alloys. This difference gives rise to the stress shielding effect and causes loosening of prostheses [2]. The research of new alloys is specially focusing on β-type Ti alloys that exhibit low elastic modulus and high ductility. They promote a good workability in cold that permits to modify the elastic modulus and its mechanical strength through microstructural changes [3–5]. It is important to highlight that the lowest value of the elastic modulus reported for the β-titanium alloy, Ti35Nb4Sn, and Ti24Nb4Zr7.9Sn, is ~40 GPa [6–8]. In a previous work, using instrumented nanoindentation, we obtained a low elastic modulus of 44 GPa in the cold rolled Ti25Nb16Hf alloy (95% C.R.) and a modulus of 66 GPa in the as cast state (0% C.R.). In the cold-rolled condition, the material presented a nanocrystalline structure, a feature that has gained attention due to its good mechanical properties and corrosion behavior [9,10]. Both results suggested ⁎ Corresponding author at: Department of Materials Science and Metallurgy, Universitat Politècnica de Catalunya (UPC), Avda. Diagonal 647, 08028 Barcelona, Spain. Tel.: +34 93 4010714; fax: +34 93 4016706. E-mail address: [email protected] (M. González).
http://dx.doi.org/10.1016/j.msec.2014.06.010 0928-4931/© 2014 Elsevier B.V. All rights reserved.
the viability of employing this material in bone-implant coupling [11,12]. Recently, Wang B. L. et al. [13] have studied cytotoxicity and hemocompatibility of Ti–22Nb–yZr/Hf (y = 2, 4 and 6 at.%) alloys, in vitro using L-929 fibroblast cell derived from mice. The results show that the Ti–Nb–Hf alloys are highly biocompatible and Hf is a nontoxic element. Both the hemolysis test and the platelet adhesion test prove excellent hemocompatibility. The morphology of the platelets adhered on the surfaces of Ti–Nb–Hf was almost the same as that detected on the surface of commercial pure Ti. However, the cytotoxicity of these Ti–Nb–Hf alloys has never been evaluated by using human osteoblast-like cells. The alloy corrosion resistance and the presence of toxic individual metals are the main factors determining its biocompatibility [14]. Another important issue is the cell–substrate interaction that promotes cell adhesion and proliferation. The main objective of the present paper is the characterization of the mechanical properties obtained by tensile testing, in order to confirm the low elastic modulus resulted from instrumented nanoindentation. The characterization of the corrosion behavior and the human osteoblast-like MG63 cell response has also been analyzed and their results have been compared to those of Ti grade II and Ti6Al4V. The elastic modulus has been confirmed. The alloy is non-cytotoxic and has a good corrosion resistance. As a result, Ti25Nb16Hf can be considered a good biomaterial for hard tissue replacements.
692
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
2. Materials and methods The new Ti alloy Ti25Nb16Hf (wt.%, vacuum arc-melted, treated at 1100 °C for 1.5 h and quenched in a mixture of ethanol/water at 0 °C), was designed using a method based on the molecular orbital theory [11,15]. The alloy has an average valence electron number (e/a) of 4.17. This value is near to the one reported by Hao et al. [6], who recommended it in order to obtain a low elastic modulus. The alloy presents a bond order (Bo) of 2.86 eV, which is near to the 2.87 eV of the Gum Metal alloy studied by Saito et al. [5]. Rectangular samples of the alloy were cold rolled at room temperature to a 95% reduction in thickness in a laboratory rolling mill with tool steel work rolls measuring 92 mm in diameter. Unidirectional multipass rolling was carried out using 5% reduction per pass using a fixed rolling speed of 35 rpm and silicone grease as lubricant. The samples were mechanically polished and finished with colloidal silica to give a mirror-like finish. All the studies used Ti grade II as a control sample due to its excellent biocompatibility [16]. For the tensile tests, Ti6Al4V was also used as a control material due to its excellent mechanical properties and because it is one of the major alloys used in orthopedic implants [17]. 2.1. Microstructural characterization Optical microstructural analysis of the Ti25Nb16Hf alloy in the 0% C.R. condition was carried out after Keller reactive attack (2 mL HF, 3 mL HCl, 5 mL HNO3, and 190 mL distilled H2O). To determine the constituent phases, X-ray diffraction measurement was conducted at room temperature using a D-500 Siemens diffractometer with Cu Kα radiation. 3 mm diameter thin disks of the alloy in the 95% C.R. condition were electropolished for transmission electron microscopy (TEM) in an electrolyte consisting of 400 mL butoxyethanol, 400 mL methanol, and 100 mL perchloric acid. 2.2. Tensile tests The standard surface-smooth tensile samples of the experimental alloy, together with the control materials, were machined in accordance with ASTM-E8-04 [18]. The tensile tests were carried out at a displacement rate of 1 mm min− 1 by a test machine Bionix series 358 (MTS Inc., USA) equipped with extensometer model 632.31F-24. An average of five measurements was taken for each group. 2.3. Corrosion tests Three plates of each material (area 0.8 cm2) were corrosion tested and compared to Ti grade II. The samples were isolated with polymethyl methacrylate (PMMA) in the interface between the air and the testing solution. The corrosion tests were carried out in accordance with ASTM G5 and ASTM G31-92 [19].
The tests were performed in 125 mL Hanks' Balanced Salt Solution, HBSS (Sigma-Aldrich) at 37 °C with a ParStat 2273 Advanced Electrochemical System potentiostat (Princeton Applied Research Company, USA) in a three-electrode cell. A graphite counter electrode was used and a saturated calomel electrode (SCE) with a standard potential of 0.241 V was employed as a reference electrode. In the corrosion tests the following results were determined: the open circuit potential (EOCP), which is the potential of an electrode measured according to the reference electrode without current, the corrosion potential (ECORR), the potential calculated at the intersection where the total oxidation rate is equal to the total reduction rate and the corrosion current density (iCORR), which is the current at the corrosion potential. The open circuit potential (EOCP) was monitored for 3 h in order to allow the leveling-off of the value before the polarization resistance test. For the cyclic polarization measurements, the potential was increased at a rate of 1 mV/s, starting from −0.3 V up to 2 V. 2.4. Cytotoxicity evaluation The possible cytotoxic effect of the alloy was evaluated using three rectangular samples of 9 mm × 9 mm. All specimens were sterilized by autoclave. Extracts of the alloy at concentrations of 1:1, 1:10, 1:100 and 1:1000 were prepared by immersing the samples in Dulbecco's Modified Eagle Medium (DMEM) at 37 °C for 72 h, in accordance with the ISO 10993-5. The MG63 cells were plated on a 96-well tissue culture polystyrene (TCPS) dish and incubated with DMEM media for 24 h. Afterwards, the culture media were replaced by the extract dilutions. Cytotoxicity was evaluated after 24 h by using the water-soluble tetrazolium salt (WST) assay. TCPS wells containing only DMEM media were used as negative control (no reactive response, 100% cell viability) and TCPS wells containing DMEM + SDS 2% were used as positive control (reactive response). Cytotoxicity was evaluated by comparing the number of living cells in each pure extract and dilution with the number of those present in the negative control. 2.5. Cell adhesion In the cell adhesion studies, 3 × 104 MG63 cells were seeded on triplicate specimens and incubated for 4 h in a 48-well culture plate. Ti grade II and the culture dish (TCPS) were used as control materials. After 4 h the cells were fixed with 4% paraformaldehyde/PBS for 15 min at room temperature, permeabilized with 0.05% Triton X100/PBS for 10 min and blocked with 1% BSA/PBS for 30 min for non-specific binding. After that the cells were stained using the double indirect immunofluorescence technique. In order to stain the focal points, primary monoclonal antivinculin V9131 antibodies from mice (Sigma-Aldrich) were used during 30 min. Subsequently the designated antibodies were conjugated with Alexa Fluor 488
Fig. 1. Optical microscopy image of Ti25Nb16Hf (0% C.R.) alloy with an X-ray diffraction pattern (a) and a bright-field TEM image of the 95% C.R. alloy with an electron diffraction pattern of the selected area (b).
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
693
antimouse IgG (Molecular Probes). Then the cells were incubated for 30 min with rhodamine-phalloidin and secondary Hoechst antibodies (Invitrogen), for the actin filaments and cell nuclei, respectively. Between each step, cells were washed with PBS for three times. Coverslips were mounted on glass slides in Mowiol (Calbiochem) mounting medium. The stained cells were photographed with a Nikon E-600 fluorescence microscope and an Olympus DP72 camera. To quantify the number of cells adhered per unit of area, four images (4 mm2 area/image) were studied for each sample using software Image-J. 2.6. Cell proliferation For cell numbers, 12 × 103 MG63 cells were seeded on triplicate specimens in 48-well TCPS culture plates. Cells were cultured for 1, 3, 7 and 14 days on the new alloy and on the control materials Ti grade II and TCPS. Cell numbers were assessed by measuring the enzyme lactate dehydrogenase (LDH) found in the cytoplasm of the tissue cell with Cytotoxity Detection Kit (LDH).
Fig. 2. Stress–strain curves of Ti25Nb16Hf alloy (0 and 95% C.W.), of Ti grade II and of Ti6Al4V.
2.7. Statistical analysis Statistical analyses were performed with significant levels up to 5%. An analysis of variance was carried out with post hoc Fisher's Protected Least Significant Difference (PLSD) to identify statistically significant differences among the experimental groups of results. 3. Results and discussion 3.1. Microstructural characterization The microstructure of the Ti25Nb16Hf (0% C.R.) alloy is shown in Fig. 1a. X-ray diffraction analysis confirms the presence of β-phase, together with a low fraction of α″ phase (inset in Fig. 1a). Fig. 1b shows a bright field image obtained by TEM from the 95% C.R. alloy. The picture shows the formation of a nanocrystalline structure with grain size less than 50 nm. Inset in Fig. 1b shows the corresponding selected area electron diffraction pattern with continual diffraction rings, which confirms the presence of many crystalline grains. The indexation of the rings corresponds to the reticular parameters of βphase. 3.2. Tensile test Tensile test results of the samples conducted before (0% C.R.) and after cold rolling (95% C.R.) are summarized in Table 1. The 0% C.R. alloy presents a low elastic modulus (~66 GPa), which is about 2/3 of that of Ti grade II. It also has a yield strength similar to that of Ti grade II and a ductility similar to Ti6Al4V. Tensile stress–strain curves (Fig. 2) indicate that cold rolling substantially decreases the elastic modulus to ~42 GPa, which is about 1/3 of its original value. These results confirm what has already been observed in the previous nanoindentation study [5]. The low elastic modulus of the 95% C.R. alloy is attributed to the nanocrystalline structure formed during the cold rolling treatment [12]. It has been found that elastic moduli of nanocrystalline materials are usually lower than the corresponding coarse-sized crystalline materials [19]. Elastic modulus decreases, as a
Table 1 Mechanical properties of the alloys studied. Material
E (GPa)
Ti25Nb16Hf (95% C.R.) Ti25Nb16Hf (0% C.R.) Ti grade II Ti6Al4V
42.3 ± 66.2 ± 103 ± 112.7 ±
3.7 5.4 13 3.3
σy (MPa)
σTS (MPa)
εf (%)
709 280 329 905
870 450 450 947
4.0 14.9 31.7 12.7
± ± ± ±
40 60 13 10
± ± ± ±
37 35 17 9
± ± ± ±
consequence of the large fraction of atoms in the grain boundaries having a lower elastic modulus [20,21]. As a result, the elastic deformability of the new alloy after cold rolling reaches 1.8%. This value doubles the result obtained before cold rolling. When the new alloy is cold worked, the ductility decreases, while the yield strength and the tensile strength increase to a value near to that of Ti6Al4V. Probably, this large elastic deformability that occurs in the so-called “superelastic alloys” is known to be due to reversible martensitic transformations resulting from the deformation [5]. The large elastic deformability is an important property as we are interested in designing biomimetic artificial bone and bone is a superelastic material [22]. The obtained results are in accordance with previous works [4,6] which showed that β-Ti alloys offer good cold workability which permits the modification of the elastic modulus and mechanical strength through microstructural changes.
3.3. Corrosion tests The results of the open circuit potentials, corrosion potential and corrosion current densities are summarized in Table 2. The Ti–Nb–Hf alloy presents a good corrosion resistance as it was expected from the high corrosion resistance reported for Hf [23] and Nb [24]. The results are in agreement with the good corrosion resistance of β-type titanium alloys such as Ti–22Nb, Ti–22Nb–6Zr [25] and Ti–29Nb–13Ta–4.6Zr [26]. The Ti25Nb16Hf alloy (95% C.R.) displays the highest corrosion resistance as revealed by its least corrosion current density value of 1.52 μA/cm2. It is worth highlighting that the cold rolled alloy is formed by a β-nanocrystalline structure. This outcome is in agreement with the best corrosion resistance reported in literature for nanocrystalline alloys [11,12]. R. Mishra et al. [12] reported a decrease in the corrosion current density from 83.18 to 24.15 μA/cm2, corresponding to bulk and nanocrystalline nickel, respectively, in experiments conducted in 1 mol/L H2SO4. The enhanced corrosion resistance of the nanocrystalline alloys has been attributed to the high surface/volume ratio due to the presence of a large number of grain boundaries. The higher diffusivity improves the formation of a protective layer [11]. The Ti25Nb16Hf alloy (0% C.R.) displayed a good corrosion resistance, with a corrosion current Table 2 Electrochemical measurement results in Hanks' solution at 37 °C.
1.0 4.8 2.9 1.8
Material
iCORR (μA/cm2)
ECORR (mV)
EOCP (mV)
Ti25Nb16Hf (95% C.R.) Ti25Nb16Hf (0% C.R.) Ti grade II
1.52 ± 0.25 2.18 ± 0.33 2.08 ± 0.26
984 ± 20 1283 ± 25 979 ± 18
−241 ± 3 −341 ± 4 −330 ± 3
694
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
Within the period of measurement, it was observed that the Ti25Nb16Hf (95% C.R.) alloy displayed the highest EOCP value, indicating thus the possibility of thermodynamic stability. The Ti25Nb16Hf (0% C.R.) and Ti grade II had the lowest EOCP value, which indicates that they have a higher thermodynamic tendency to corrode in the studied medium. 3.4. Cytotoxicity evaluation
Fig. 3. Cytotoxicity test of MG63 cells to the substrate of Ti25Nb16Hf alloy. Data represent means and ± standard deviation. The dashed line corresponds to the negative control response.
density value similar to that of Ti grade II. From the corrosion potential results we can observe that the Ti25Nb16Hf alloy (95% C.R.) and Ti grade II displayed an ECORR more cathodic than the Ti25Nb16Hf (0% C.R.) sample.
Fig. 3 shows the results of the cytotoxicity test for the Ti25Nb16Hf alloy samples. The graph presents the percentage of viability of cells cultured in the pure extract and in the dilution of the medium that were in contact with the alloy during 72 h of immersion. A 100% of cell viability can be seen in the extracts of the alloy with no significant differences with the results of the culture TCPS plate used as a negative control. It confirms that the new alloy has an excellent biocompatibility in vitro as the alloying elements present in the material did not induce cytotoxic effect, as it has been reported by literature [14,22]. 3.5. Cell adhesion Ti25Nb16Hf (95% C.R.) alloy was selected for the cell adhesion test due to its low elastic modulus and to its high corrosion resistance. Immunofluorescence images (Fig. 4) show the cells adhering to the new alloy and to the control materials after 4 h of culture. MG63 cells
Fig. 4. Double immunofluorescence staining (blue—nucleus, red—actin filaments and green—focal points) of MG63 adhered cells after 4 h of culture corresponding to the Ti25Nb16Hf alloy, Ti grade II and TCPS substrates. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
695
were stained to visualize focal contacts (green), cytoskeletal structures of actin filaments (red) and nuclei (blue). The observation of the cell morphology shows a good polygon-like adherent growth and good cell spreading with well-defined clusters of focal adhesion points in the different surfaces studied. The figure shows a lower number of adhering cells on the Ti25Nb16Hf specimen, compared to the control materials. The higher cell recruitment corresponds to the TCPS surface. Cell number quantification (Fig. 5) demonstrates that the new alloy induces a lower cell recruitment compared to Ti grade II, with significant difference. The number of attached cells on the new alloy is about half the value determined for TCPS control. Results are in accordance with the fibroblast cell adhesion studies reported by Wang B L et al. [13] who concluded that the number of adhered cells was lower on Ti–Nb–Hf alloy than on pure Ti. 3.6. Cell proliferation Cell number studies were carried out in cell cultures of 1, 3, 7 and 14 days and show that the MG63 cells proliferated properly on the surface of the new alloy (Fig. 6). Although the cell adhesion after 4 h was lower in the new alloy, there were no differences in the proliferation behavior of the MG63 cells after cultures of 1, 3 and 7 days. The differences in the number of cells after 14 days of incubation were statistically lower in the Ti grade II surface than in the Ti–Nb–Hf alloy. The cells proliferated continually in all the substrates, as indicated by the test showing the non-cytotoxic nature of the Ti25Nb16Hf alloy. The results are in agreement with the Arciniegas et al. [27] studies which showed similar proliferation behavior between the Ti41.2Nb6.1Zr and Ti grade II substrates after cultures of 1 and 9 days. 4. Conclusions The low elastic modulus of the Ti25Nb16Hf alloy obtained by tensile testing confirms the results of the previous nanoindentation studies. The elastic modulus of the new alloy decreases after cold rolling. The 95% C.R. alloy shows a satisfactory balance between low elastic modulus (42 GPa), high strength (870 MPa) and high elastic deformability. The developed alloy exhibits a good corrosion resistance, similar to that of Ti grade II. The cell cultures show the excellent cell response in terms of cytotoxicity evaluation, cell adhesion and proliferation. As a whole, these outcomes demonstrate the potentialities of the novel Ti alloy with mechanical properties similar to those of cortical
Fig. 5. Number of cells/cm2 attached to the Ti25Nb16Hf alloy and control substrates after 4 h of culture.
Fig. 6. Proliferation number of MG63 cells cultured on each sample for periods of 1, 3, 7 and 14 days.
bone and with a good osteoblast cell response. This material could be considered a good candidate to be used in implant devices for replacing failed hard tissue. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.msec.2014.06.010. References [1] M. Niinomi, Mater. Sci. Eng. A 243 (1998) 231–236. [2] M.I.Z. Ridzwan, S. Shuib, A.Y. Hassan, A.A. Shokri, M.N. Mohamad Ibrahim, J. Med. Sci. 7 (2007) 460–467. [3] T.W. Duering, K.N. Melton, D. Stoeckel, C.M. Wayman, Engineering aspects of shape memory alloys, Butterworth-Heinemann Ltd, 1990. [4] Y.L. Hao, Titanium alloy with extra-low modulus and superelasticity and its producing method and processing thereof, US 2007/0137742 a1, 2007. [5] T. Saito, T. Furuta, J.H. Hwang, S. Kuramoto, K. Nishino, N. Suzuki, R. Cheng, A. Yamada, K. Ito, Y. Seno, T. Nonaka, H. Ikehata, N. Nagasako, C. Iwamoto, Y. Ikuhara, T. Sakuma, Science 300 (2003) 464–467. [6] Y.L. Hao, S.J. Li, S.Y. Sun, C.Y. Zheng, R. Yang, Acta Biomater. 3 (2008) 277–286. [7] H. Matsumoto, S. Watanabe, S. Hanada, Mater. Trans. 46 (2005) 1070–1078. [8] M. Niinomi, M. Nakai, J. Hieda, Acta Biomater. 8 (2012) 3888–3903. [9] M. González, J. Peña, J.M. Manero, M. Arciniegas, F.J.J. Gil, Mater. Eng. Perform. 18 (2009) 506–510. [10] M. González, F.J. Gil, J.M. Manero, J.J. Peña, Mater. Eng. Perform. 20 (2011) 653–657. [11] D. Zander, U. Köster, Mater. Sci. Eng. A 375–377 (2004) 53–59. [12] R. Mishra, R. Balasubramaniam, Corr. Sci. 46 (2004) 3019–3029. [13] B.L. Wang, L. Li, Y.F. Zheng, Biomed. Mater. 5 (2010) 1–8. [14] A. Balakrishnan, B.C. Lee, T.N. Kim, B.B. Panigrahi, Trends Biomater, Artif. Organs 2 (2008) 58–64. [15] M. Arciniegas, J. Peña, J.M. Manero, J.C. Paniagua, F.J. Gil, Phil. Mag. 88 (2008) 2529–2548. [16] M. Niinomi, Sci. Technol. Adv. Mater. 4 (2003) 445–454. [17] M. Long, H.J. Rack, Biomaterials 19 (1998) 1621–1639. [18] ASTM-E8-04, Standard test methods for tension testing of metallic materials, annual book of ASTM standards, American Society for Testing and Materials, Philadelphia, PA, 2004. [19] American Society for Testing, Materials, Annual Book of ASTM Standard, ASTM, Philadelphia, PA, 1978. 817. [20] J. Schiøtz, F.D. Di Tolla, K.W. Jacobsen, Nature 391 (1998) 561–563. [21] Y.L. Hao, S.J. Li, Y. Sun, C.Y. Zheng, Q.M. Hu, R. Yang, Appl. Phys. Lett. 87 (9) (2005) 1–3. [22] H. Matsuno, A. Yokoyama, F. Watari, M. Uo, T. Kawasaki, Biomaterials 22 (2001) 1253–1262. [23] S. Mohammadi, M. Esposito, M. Cucu, L.E. Ericson, Thomsen P, J. Mater. Sci. Mater. Med. 12 (2001) 603–611. [24] D.L. Moffat, D.C. Larbalestier, Metall. Mater. Trans. A 19 (7) (1988) 1687–1694. [25] B.L. Wang, Y.F. Zheng, L.C. Zhao, Mater. Corros. 60 (2009) 788–794. [26] Y. Tanaka, M. Nakai, T. Akahori, M. Niinomi, Y. Tsutsumi, H. Doi, T. Hanawa, Corros. Sci. 50 (2008) 2111–2116. [27] M. Arciniegas, J. Peña, F.J. Gi, J.M. Manero, J. Biomed. Mater. Res. B 101 (2013) 709–720.
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Low modulus Ti–Nb–Hf alloy for biomedical applications M. González a,b,⁎, J. Peña a,b, F.J. Gil a,c, J.M. Manero a,c a b c
Department of Materials Science and Metallurgy, Universitat Politècnica de Catalunya (UPC), Avda. Diagonal 647, 08028 Barcelona, Spain Materials Science, Elisava Escola Superior de Disseny i Enginyeria de Barcelona, La Rambla 30-32, 08002 Barcelona, Spain Ciber-BBN, Spain
a r t i c l e
i n f o
Article history: Received 20 November 2013 Received in revised form 13 April 2014 Accepted 9 June 2014 Available online 19 June 2014 Keywords: Corrosion Biocompatibility Cytotoxicity Low elastic modulus Stress shielding effect TiNbHf alloys
a b s t r a c t β-Type titanium alloys with a low elastic modulus are a potential strategy to reduce stress shielding effect and to enhance bone remodeling in implants used to substitute failed hard tissue. For biomaterial application, investigation on the mechanical behavior, the corrosion resistance and the cell response is required. The new Ti25Nb16Hf alloy was studied before and after 95% cold rolling (95% C.R.). The mechanical properties were determined by tensile testing and its corrosion behavior was analyzed by potentiostatic equipment in Hank's solution at 37 °C. The cell response was studied by means of cytotoxicity evaluation, cell adhesion and proliferation measurements. The stress–strain curves showed the lowest elastic modulus (42 GPa) in the cold worked alloy and high tensile strength, similar to that of Ti6Al4V. The new alloy exhibited better corrosion resistance in terms of open circuit potential (EOCP), but was similar in terms of corrosion current density (iCORR) compared to Ti grade II. Cytotoxicity studies revealed that the chemical composition of the alloy does not induce cytotoxic activity. Cell studies in the new alloy showed a lower adhesion and a higher proliferation compared to Ti grade II presenting, therefore, mechanical features similar to those of human cortical bone and, simultaneously, a good cell response. © 2014 Elsevier B.V. All rights reserved.
1. Introduction In the bone substitution metal implant field, one of the main challenges is to achieve materials that exhibit a low elastic modulus, similar to that of human cortical bone (10–30 GPa) [1]. The most commercial alloys used as implant materials present an elastic modulus in the range of 110 GPa of Ti6Al4V up to 220 GPa of cobalt–chrome alloys. This difference gives rise to the stress shielding effect and causes loosening of prostheses [2]. The research of new alloys is specially focusing on β-type Ti alloys that exhibit low elastic modulus and high ductility. They promote a good workability in cold that permits to modify the elastic modulus and its mechanical strength through microstructural changes [3–5]. It is important to highlight that the lowest value of the elastic modulus reported for the β-titanium alloy, Ti35Nb4Sn, and Ti24Nb4Zr7.9Sn, is ~40 GPa [6–8]. In a previous work, using instrumented nanoindentation, we obtained a low elastic modulus of 44 GPa in the cold rolled Ti25Nb16Hf alloy (95% C.R.) and a modulus of 66 GPa in the as cast state (0% C.R.). In the cold-rolled condition, the material presented a nanocrystalline structure, a feature that has gained attention due to its good mechanical properties and corrosion behavior [9,10]. Both results suggested ⁎ Corresponding author at: Department of Materials Science and Metallurgy, Universitat Politècnica de Catalunya (UPC), Avda. Diagonal 647, 08028 Barcelona, Spain. Tel.: +34 93 4010714; fax: +34 93 4016706. E-mail address: [email protected] (M. González).
http://dx.doi.org/10.1016/j.msec.2014.06.010 0928-4931/© 2014 Elsevier B.V. All rights reserved.
the viability of employing this material in bone-implant coupling [11,12]. Recently, Wang B. L. et al. [13] have studied cytotoxicity and hemocompatibility of Ti–22Nb–yZr/Hf (y = 2, 4 and 6 at.%) alloys, in vitro using L-929 fibroblast cell derived from mice. The results show that the Ti–Nb–Hf alloys are highly biocompatible and Hf is a nontoxic element. Both the hemolysis test and the platelet adhesion test prove excellent hemocompatibility. The morphology of the platelets adhered on the surfaces of Ti–Nb–Hf was almost the same as that detected on the surface of commercial pure Ti. However, the cytotoxicity of these Ti–Nb–Hf alloys has never been evaluated by using human osteoblast-like cells. The alloy corrosion resistance and the presence of toxic individual metals are the main factors determining its biocompatibility [14]. Another important issue is the cell–substrate interaction that promotes cell adhesion and proliferation. The main objective of the present paper is the characterization of the mechanical properties obtained by tensile testing, in order to confirm the low elastic modulus resulted from instrumented nanoindentation. The characterization of the corrosion behavior and the human osteoblast-like MG63 cell response has also been analyzed and their results have been compared to those of Ti grade II and Ti6Al4V. The elastic modulus has been confirmed. The alloy is non-cytotoxic and has a good corrosion resistance. As a result, Ti25Nb16Hf can be considered a good biomaterial for hard tissue replacements.
692
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
2. Materials and methods The new Ti alloy Ti25Nb16Hf (wt.%, vacuum arc-melted, treated at 1100 °C for 1.5 h and quenched in a mixture of ethanol/water at 0 °C), was designed using a method based on the molecular orbital theory [11,15]. The alloy has an average valence electron number (e/a) of 4.17. This value is near to the one reported by Hao et al. [6], who recommended it in order to obtain a low elastic modulus. The alloy presents a bond order (Bo) of 2.86 eV, which is near to the 2.87 eV of the Gum Metal alloy studied by Saito et al. [5]. Rectangular samples of the alloy were cold rolled at room temperature to a 95% reduction in thickness in a laboratory rolling mill with tool steel work rolls measuring 92 mm in diameter. Unidirectional multipass rolling was carried out using 5% reduction per pass using a fixed rolling speed of 35 rpm and silicone grease as lubricant. The samples were mechanically polished and finished with colloidal silica to give a mirror-like finish. All the studies used Ti grade II as a control sample due to its excellent biocompatibility [16]. For the tensile tests, Ti6Al4V was also used as a control material due to its excellent mechanical properties and because it is one of the major alloys used in orthopedic implants [17]. 2.1. Microstructural characterization Optical microstructural analysis of the Ti25Nb16Hf alloy in the 0% C.R. condition was carried out after Keller reactive attack (2 mL HF, 3 mL HCl, 5 mL HNO3, and 190 mL distilled H2O). To determine the constituent phases, X-ray diffraction measurement was conducted at room temperature using a D-500 Siemens diffractometer with Cu Kα radiation. 3 mm diameter thin disks of the alloy in the 95% C.R. condition were electropolished for transmission electron microscopy (TEM) in an electrolyte consisting of 400 mL butoxyethanol, 400 mL methanol, and 100 mL perchloric acid. 2.2. Tensile tests The standard surface-smooth tensile samples of the experimental alloy, together with the control materials, were machined in accordance with ASTM-E8-04 [18]. The tensile tests were carried out at a displacement rate of 1 mm min− 1 by a test machine Bionix series 358 (MTS Inc., USA) equipped with extensometer model 632.31F-24. An average of five measurements was taken for each group. 2.3. Corrosion tests Three plates of each material (area 0.8 cm2) were corrosion tested and compared to Ti grade II. The samples were isolated with polymethyl methacrylate (PMMA) in the interface between the air and the testing solution. The corrosion tests were carried out in accordance with ASTM G5 and ASTM G31-92 [19].
The tests were performed in 125 mL Hanks' Balanced Salt Solution, HBSS (Sigma-Aldrich) at 37 °C with a ParStat 2273 Advanced Electrochemical System potentiostat (Princeton Applied Research Company, USA) in a three-electrode cell. A graphite counter electrode was used and a saturated calomel electrode (SCE) with a standard potential of 0.241 V was employed as a reference electrode. In the corrosion tests the following results were determined: the open circuit potential (EOCP), which is the potential of an electrode measured according to the reference electrode without current, the corrosion potential (ECORR), the potential calculated at the intersection where the total oxidation rate is equal to the total reduction rate and the corrosion current density (iCORR), which is the current at the corrosion potential. The open circuit potential (EOCP) was monitored for 3 h in order to allow the leveling-off of the value before the polarization resistance test. For the cyclic polarization measurements, the potential was increased at a rate of 1 mV/s, starting from −0.3 V up to 2 V. 2.4. Cytotoxicity evaluation The possible cytotoxic effect of the alloy was evaluated using three rectangular samples of 9 mm × 9 mm. All specimens were sterilized by autoclave. Extracts of the alloy at concentrations of 1:1, 1:10, 1:100 and 1:1000 were prepared by immersing the samples in Dulbecco's Modified Eagle Medium (DMEM) at 37 °C for 72 h, in accordance with the ISO 10993-5. The MG63 cells were plated on a 96-well tissue culture polystyrene (TCPS) dish and incubated with DMEM media for 24 h. Afterwards, the culture media were replaced by the extract dilutions. Cytotoxicity was evaluated after 24 h by using the water-soluble tetrazolium salt (WST) assay. TCPS wells containing only DMEM media were used as negative control (no reactive response, 100% cell viability) and TCPS wells containing DMEM + SDS 2% were used as positive control (reactive response). Cytotoxicity was evaluated by comparing the number of living cells in each pure extract and dilution with the number of those present in the negative control. 2.5. Cell adhesion In the cell adhesion studies, 3 × 104 MG63 cells were seeded on triplicate specimens and incubated for 4 h in a 48-well culture plate. Ti grade II and the culture dish (TCPS) were used as control materials. After 4 h the cells were fixed with 4% paraformaldehyde/PBS for 15 min at room temperature, permeabilized with 0.05% Triton X100/PBS for 10 min and blocked with 1% BSA/PBS for 30 min for non-specific binding. After that the cells were stained using the double indirect immunofluorescence technique. In order to stain the focal points, primary monoclonal antivinculin V9131 antibodies from mice (Sigma-Aldrich) were used during 30 min. Subsequently the designated antibodies were conjugated with Alexa Fluor 488
Fig. 1. Optical microscopy image of Ti25Nb16Hf (0% C.R.) alloy with an X-ray diffraction pattern (a) and a bright-field TEM image of the 95% C.R. alloy with an electron diffraction pattern of the selected area (b).
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
693
antimouse IgG (Molecular Probes). Then the cells were incubated for 30 min with rhodamine-phalloidin and secondary Hoechst antibodies (Invitrogen), for the actin filaments and cell nuclei, respectively. Between each step, cells were washed with PBS for three times. Coverslips were mounted on glass slides in Mowiol (Calbiochem) mounting medium. The stained cells were photographed with a Nikon E-600 fluorescence microscope and an Olympus DP72 camera. To quantify the number of cells adhered per unit of area, four images (4 mm2 area/image) were studied for each sample using software Image-J. 2.6. Cell proliferation For cell numbers, 12 × 103 MG63 cells were seeded on triplicate specimens in 48-well TCPS culture plates. Cells were cultured for 1, 3, 7 and 14 days on the new alloy and on the control materials Ti grade II and TCPS. Cell numbers were assessed by measuring the enzyme lactate dehydrogenase (LDH) found in the cytoplasm of the tissue cell with Cytotoxity Detection Kit (LDH).
Fig. 2. Stress–strain curves of Ti25Nb16Hf alloy (0 and 95% C.W.), of Ti grade II and of Ti6Al4V.
2.7. Statistical analysis Statistical analyses were performed with significant levels up to 5%. An analysis of variance was carried out with post hoc Fisher's Protected Least Significant Difference (PLSD) to identify statistically significant differences among the experimental groups of results. 3. Results and discussion 3.1. Microstructural characterization The microstructure of the Ti25Nb16Hf (0% C.R.) alloy is shown in Fig. 1a. X-ray diffraction analysis confirms the presence of β-phase, together with a low fraction of α″ phase (inset in Fig. 1a). Fig. 1b shows a bright field image obtained by TEM from the 95% C.R. alloy. The picture shows the formation of a nanocrystalline structure with grain size less than 50 nm. Inset in Fig. 1b shows the corresponding selected area electron diffraction pattern with continual diffraction rings, which confirms the presence of many crystalline grains. The indexation of the rings corresponds to the reticular parameters of βphase. 3.2. Tensile test Tensile test results of the samples conducted before (0% C.R.) and after cold rolling (95% C.R.) are summarized in Table 1. The 0% C.R. alloy presents a low elastic modulus (~66 GPa), which is about 2/3 of that of Ti grade II. It also has a yield strength similar to that of Ti grade II and a ductility similar to Ti6Al4V. Tensile stress–strain curves (Fig. 2) indicate that cold rolling substantially decreases the elastic modulus to ~42 GPa, which is about 1/3 of its original value. These results confirm what has already been observed in the previous nanoindentation study [5]. The low elastic modulus of the 95% C.R. alloy is attributed to the nanocrystalline structure formed during the cold rolling treatment [12]. It has been found that elastic moduli of nanocrystalline materials are usually lower than the corresponding coarse-sized crystalline materials [19]. Elastic modulus decreases, as a
Table 1 Mechanical properties of the alloys studied. Material
E (GPa)
Ti25Nb16Hf (95% C.R.) Ti25Nb16Hf (0% C.R.) Ti grade II Ti6Al4V
42.3 ± 66.2 ± 103 ± 112.7 ±
3.7 5.4 13 3.3
σy (MPa)
σTS (MPa)
εf (%)
709 280 329 905
870 450 450 947
4.0 14.9 31.7 12.7
± ± ± ±
40 60 13 10
± ± ± ±
37 35 17 9
± ± ± ±
consequence of the large fraction of atoms in the grain boundaries having a lower elastic modulus [20,21]. As a result, the elastic deformability of the new alloy after cold rolling reaches 1.8%. This value doubles the result obtained before cold rolling. When the new alloy is cold worked, the ductility decreases, while the yield strength and the tensile strength increase to a value near to that of Ti6Al4V. Probably, this large elastic deformability that occurs in the so-called “superelastic alloys” is known to be due to reversible martensitic transformations resulting from the deformation [5]. The large elastic deformability is an important property as we are interested in designing biomimetic artificial bone and bone is a superelastic material [22]. The obtained results are in accordance with previous works [4,6] which showed that β-Ti alloys offer good cold workability which permits the modification of the elastic modulus and mechanical strength through microstructural changes.
3.3. Corrosion tests The results of the open circuit potentials, corrosion potential and corrosion current densities are summarized in Table 2. The Ti–Nb–Hf alloy presents a good corrosion resistance as it was expected from the high corrosion resistance reported for Hf [23] and Nb [24]. The results are in agreement with the good corrosion resistance of β-type titanium alloys such as Ti–22Nb, Ti–22Nb–6Zr [25] and Ti–29Nb–13Ta–4.6Zr [26]. The Ti25Nb16Hf alloy (95% C.R.) displays the highest corrosion resistance as revealed by its least corrosion current density value of 1.52 μA/cm2. It is worth highlighting that the cold rolled alloy is formed by a β-nanocrystalline structure. This outcome is in agreement with the best corrosion resistance reported in literature for nanocrystalline alloys [11,12]. R. Mishra et al. [12] reported a decrease in the corrosion current density from 83.18 to 24.15 μA/cm2, corresponding to bulk and nanocrystalline nickel, respectively, in experiments conducted in 1 mol/L H2SO4. The enhanced corrosion resistance of the nanocrystalline alloys has been attributed to the high surface/volume ratio due to the presence of a large number of grain boundaries. The higher diffusivity improves the formation of a protective layer [11]. The Ti25Nb16Hf alloy (0% C.R.) displayed a good corrosion resistance, with a corrosion current Table 2 Electrochemical measurement results in Hanks' solution at 37 °C.
1.0 4.8 2.9 1.8
Material
iCORR (μA/cm2)
ECORR (mV)
EOCP (mV)
Ti25Nb16Hf (95% C.R.) Ti25Nb16Hf (0% C.R.) Ti grade II
1.52 ± 0.25 2.18 ± 0.33 2.08 ± 0.26
984 ± 20 1283 ± 25 979 ± 18
−241 ± 3 −341 ± 4 −330 ± 3
694
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
Within the period of measurement, it was observed that the Ti25Nb16Hf (95% C.R.) alloy displayed the highest EOCP value, indicating thus the possibility of thermodynamic stability. The Ti25Nb16Hf (0% C.R.) and Ti grade II had the lowest EOCP value, which indicates that they have a higher thermodynamic tendency to corrode in the studied medium. 3.4. Cytotoxicity evaluation
Fig. 3. Cytotoxicity test of MG63 cells to the substrate of Ti25Nb16Hf alloy. Data represent means and ± standard deviation. The dashed line corresponds to the negative control response.
density value similar to that of Ti grade II. From the corrosion potential results we can observe that the Ti25Nb16Hf alloy (95% C.R.) and Ti grade II displayed an ECORR more cathodic than the Ti25Nb16Hf (0% C.R.) sample.
Fig. 3 shows the results of the cytotoxicity test for the Ti25Nb16Hf alloy samples. The graph presents the percentage of viability of cells cultured in the pure extract and in the dilution of the medium that were in contact with the alloy during 72 h of immersion. A 100% of cell viability can be seen in the extracts of the alloy with no significant differences with the results of the culture TCPS plate used as a negative control. It confirms that the new alloy has an excellent biocompatibility in vitro as the alloying elements present in the material did not induce cytotoxic effect, as it has been reported by literature [14,22]. 3.5. Cell adhesion Ti25Nb16Hf (95% C.R.) alloy was selected for the cell adhesion test due to its low elastic modulus and to its high corrosion resistance. Immunofluorescence images (Fig. 4) show the cells adhering to the new alloy and to the control materials after 4 h of culture. MG63 cells
Fig. 4. Double immunofluorescence staining (blue—nucleus, red—actin filaments and green—focal points) of MG63 adhered cells after 4 h of culture corresponding to the Ti25Nb16Hf alloy, Ti grade II and TCPS substrates. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).
M. González et al. / Materials Science and Engineering C 42 (2014) 691–695
695
were stained to visualize focal contacts (green), cytoskeletal structures of actin filaments (red) and nuclei (blue). The observation of the cell morphology shows a good polygon-like adherent growth and good cell spreading with well-defined clusters of focal adhesion points in the different surfaces studied. The figure shows a lower number of adhering cells on the Ti25Nb16Hf specimen, compared to the control materials. The higher cell recruitment corresponds to the TCPS surface. Cell number quantification (Fig. 5) demonstrates that the new alloy induces a lower cell recruitment compared to Ti grade II, with significant difference. The number of attached cells on the new alloy is about half the value determined for TCPS control. Results are in accordance with the fibroblast cell adhesion studies reported by Wang B L et al. [13] who concluded that the number of adhered cells was lower on Ti–Nb–Hf alloy than on pure Ti. 3.6. Cell proliferation Cell number studies were carried out in cell cultures of 1, 3, 7 and 14 days and show that the MG63 cells proliferated properly on the surface of the new alloy (Fig. 6). Although the cell adhesion after 4 h was lower in the new alloy, there were no differences in the proliferation behavior of the MG63 cells after cultures of 1, 3 and 7 days. The differences in the number of cells after 14 days of incubation were statistically lower in the Ti grade II surface than in the Ti–Nb–Hf alloy. The cells proliferated continually in all the substrates, as indicated by the test showing the non-cytotoxic nature of the Ti25Nb16Hf alloy. The results are in agreement with the Arciniegas et al. [27] studies which showed similar proliferation behavior between the Ti41.2Nb6.1Zr and Ti grade II substrates after cultures of 1 and 9 days. 4. Conclusions The low elastic modulus of the Ti25Nb16Hf alloy obtained by tensile testing confirms the results of the previous nanoindentation studies. The elastic modulus of the new alloy decreases after cold rolling. The 95% C.R. alloy shows a satisfactory balance between low elastic modulus (42 GPa), high strength (870 MPa) and high elastic deformability. The developed alloy exhibits a good corrosion resistance, similar to that of Ti grade II. The cell cultures show the excellent cell response in terms of cytotoxicity evaluation, cell adhesion and proliferation. As a whole, these outcomes demonstrate the potentialities of the novel Ti alloy with mechanical properties similar to those of cortical
Fig. 5. Number of cells/cm2 attached to the Ti25Nb16Hf alloy and control substrates after 4 h of culture.
Fig. 6. Proliferation number of MG63 cells cultured on each sample for periods of 1, 3, 7 and 14 days.
bone and with a good osteoblast cell response. This material could be considered a good candidate to be used in implant devices for replacing failed hard tissue. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.msec.2014.06.010. References [1] M. Niinomi, Mater. Sci. Eng. A 243 (1998) 231–236. [2] M.I.Z. Ridzwan, S. Shuib, A.Y. Hassan, A.A. Shokri, M.N. Mohamad Ibrahim, J. Med. Sci. 7 (2007) 460–467. [3] T.W. Duering, K.N. Melton, D. Stoeckel, C.M. Wayman, Engineering aspects of shape memory alloys, Butterworth-Heinemann Ltd, 1990. [4] Y.L. Hao, Titanium alloy with extra-low modulus and superelasticity and its producing method and processing thereof, US 2007/0137742 a1, 2007. [5] T. Saito, T. Furuta, J.H. Hwang, S. Kuramoto, K. Nishino, N. Suzuki, R. Cheng, A. Yamada, K. Ito, Y. Seno, T. Nonaka, H. Ikehata, N. Nagasako, C. Iwamoto, Y. Ikuhara, T. Sakuma, Science 300 (2003) 464–467. [6] Y.L. Hao, S.J. Li, S.Y. Sun, C.Y. Zheng, R. Yang, Acta Biomater. 3 (2008) 277–286. [7] H. Matsumoto, S. Watanabe, S. Hanada, Mater. Trans. 46 (2005) 1070–1078. [8] M. Niinomi, M. Nakai, J. Hieda, Acta Biomater. 8 (2012) 3888–3903. [9] M. González, J. Peña, J.M. Manero, M. Arciniegas, F.J.J. Gil, Mater. Eng. Perform. 18 (2009) 506–510. [10] M. González, F.J. Gil, J.M. Manero, J.J. Peña, Mater. Eng. Perform. 20 (2011) 653–657. [11] D. Zander, U. Köster, Mater. Sci. Eng. A 375–377 (2004) 53–59. [12] R. Mishra, R. Balasubramaniam, Corr. Sci. 46 (2004) 3019–3029. [13] B.L. Wang, L. Li, Y.F. Zheng, Biomed. Mater. 5 (2010) 1–8. [14] A. Balakrishnan, B.C. Lee, T.N. Kim, B.B. Panigrahi, Trends Biomater, Artif. Organs 2 (2008) 58–64. [15] M. Arciniegas, J. Peña, J.M. Manero, J.C. Paniagua, F.J. Gil, Phil. Mag. 88 (2008) 2529–2548. [16] M. Niinomi, Sci. Technol. Adv. Mater. 4 (2003) 445–454. [17] M. Long, H.J. Rack, Biomaterials 19 (1998) 1621–1639. [18] ASTM-E8-04, Standard test methods for tension testing of metallic materials, annual book of ASTM standards, American Society for Testing and Materials, Philadelphia, PA, 2004. [19] American Society for Testing, Materials, Annual Book of ASTM Standard, ASTM, Philadelphia, PA, 1978. 817. [20] J. Schiøtz, F.D. Di Tolla, K.W. Jacobsen, Nature 391 (1998) 561–563. [21] Y.L. Hao, S.J. Li, Y. Sun, C.Y. Zheng, Q.M. Hu, R. Yang, Appl. Phys. Lett. 87 (9) (2005) 1–3. [22] H. Matsuno, A. Yokoyama, F. Watari, M. Uo, T. Kawasaki, Biomaterials 22 (2001) 1253–1262. [23] S. Mohammadi, M. Esposito, M. Cucu, L.E. Ericson, Thomsen P, J. Mater. Sci. Mater. Med. 12 (2001) 603–611. [24] D.L. Moffat, D.C. Larbalestier, Metall. Mater. Trans. A 19 (7) (1988) 1687–1694. [25] B.L. Wang, Y.F. Zheng, L.C. Zhao, Mater. Corros. 60 (2009) 788–794. [26] Y. Tanaka, M. Nakai, T. Akahori, M. Niinomi, Y. Tsutsumi, H. Doi, T. Hanawa, Corros. Sci. 50 (2008) 2111–2116. [27] M. Arciniegas, J. Peña, F.J. Gi, J.M. Manero, J. Biomed. Mater. Res. B 101 (2013) 709–720.