Acta Biomaterialia 12 (2015) 352–361

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Mechanical properties and cytocompatibility of oxygen-modified b-type Ti–Cr alloys for spinal fixation devices Huihong Liu a,⇑, Mitsuo Niinomi a, Masaaki Nakai a, Ken Cho a, Kengo Narita a, Mustafa Sß en b,c, Hitoshi Shiku b, Tomokazu Matsue b,d a

Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan Graduate School of Environmental Studies, Tohoku University, Sendai 980-8579, Japan Department of Biomedical Engineering, Faculty of Engineering and Architecture, Katip Celebi University, Izmir 35620, Turkey d WPI-Advanced Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan b c

a r t i c l e

i n f o

Article history: Received 24 June 2014 Received in revised form 5 October 2014 Accepted 15 October 2014 Available online 23 October 2014 Keywords: Spinal fixation Titanium alloys Changeable Young’s modulus Deformation-induced x phase Cytocompatibility

a b s t r a c t In this study, various amounts of oxygen were added to Ti–10Cr (mass%) alloys. It is expected that a large changeable Young’s modulus, caused by a deformation-induced x-phase transformation, can be achieved in Ti–10Cr–O alloys by the appropriate oxygen addition. This ‘‘changeable Young’s modulus’’ property can satisfy the otherwise conflicting requirements for use in spinal implant rods: high and low moduli are preferred by surgeons and patients, respectively. The influence of oxygen on the microstructures and mechanical properties of the alloys was examined, as well as the bending springback and cytocompatibility of the optimized alloy. Among the Ti–10Cr–O alloys, Ti–10Cr–0.2O (mass%) alloy shows the largest changeable Young’s modulus following cold rolling for a constant reduction ratio. This is the result of two competing factors: increased apparent b-lattice stability and decreased amounts of athermal x phase, both of which are caused by oxygen addition. The most favorable balance of these factors for the deformation-induced x-phase transformation occurred at an oxygen concentration of 0.2 mass%. Ti–10Cr– 0.2O alloy not only exhibits high tensile strength and acceptable elongation, but also possesses a good combination of high bending strength, acceptable bending springback and great cytocompatibility. Therefore, Ti–10Cr–0.2O alloy is a potential material for use in spinal fixture devices. Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction The x phase, with a non-close-packed hexagonal structure, is a thermodynamically metastable phase typically found in group IV transition metal (titanium, zirconium and hafnium) base alloys that contain appropriate contents of body-centered cubic (bcc) bphase stabilizing elements [1]. The x phase can be classified according to the process by which it is formed into one of three categories: (i) an athermal x phase formed upon rapid quenching from the b-phase field [2,3]; (ii) an isothermal x phase precipitated during an isothermal aging treatment in the temperature range 373–773 K [4]; or (iii) a deformation-induced x phase formed by the application of stress and/or strain [5–7]. Traditionally, the x phase in Ti alloys is considered a brittle phase that has a deleterious effect on mechanical properties. Recently, however, metastable x-phase materials have attracted a great deal of attention, not only because the x-phase transformation exemplifies an interesting ⇑ Corresponding author. E-mail address: [email protected] (H. Liu). http://dx.doi.org/10.1016/j.actbio.2014.10.014 1742-7061/Ó 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

class of phase transformations, but also because of the marked effect that these phases have on the physical and mechanical properties of the alloys [1,4,8–14]. Nakai et al. [9] determined that the fatigue limit of the biomedical Ti–29Nb–13Ta–4.6Zr (TNTZ) alloy could be enhanced by controlling the amount of isothermal x phase precipitated by a short aging. Kim et al. [10] reported that the precipitation of x phase by aging (isothermal x phase) is effective in increasing the critical stress for slip deformation and that this results in stable and excellent superelasticity in Ti–Nb binary alloys. An application proposed recently for materials that undergo deformation-induced x-phase transformations is in the development of better spinal fixation surgery methods [11–13]. According to surgeons specializing in spinal diseases, low springback is desirable in the implant rod so that it can be handled more easily by surgeons who are required to manipulate the shape of the rod to conform to the curvature of the spine within the limited space inside the patient’s body [15]. Both the bending strength and Young’s modulus of a material determine the degree of springback, and, for two rods with a given strength, it is the rod with the higher

353

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

Young’s modulus that will give the smaller springback. Thus, implant materials with high Young’s modulus are preferable for surgeons from the viewpoint of suppression of springback. On the other hand, low-modulus rods are still required in order to prevent stress shielding, which benefits patients after surgery [16]. b-Ti alloys have been developed for widespread biomedical applications because of their favorable mechanical properties, excellent biocompatibility and corrosion resistance, as well as their low Young’s moduli [17–19]. For certain metastable b-Ti alloys, non-equilibrium phases, such as a0 , a00 and x, would probably appear during deformation [20,21]. Since the Young’s modulus of the x phase is higher than that of the b phase [22], the localized, deformation-induced x-phase transformation within the deformed part of a b-Ti alloy component would increase the Young’s modulus of only the deformed region, which is beneficial for suppressing springback. Meanwhile, a low Young’s modulus is still maintained in the non-deformed part because no x-phase transformation occurs in the non-deformed part, which is favorable for preventing stress shielding. Thus, a ‘‘changeable Young’s modulus’’ or ‘‘selftunable Young’s modulus’’ could be achieved in the implant rod; this confers on the material the capability to satisfy the conflicting requirements of surgeons and patients. This changeable Young’s modulus effect has been investigated in b-type binary Ti–Cr alloys intended for spinal fixation devices in a previous report [12]. In that study, the Ti–12Cr alloy showed a large increase in Young’s modulus by cold rolling through deformation-induced x-phase transformation. It has been reported [23] that the decreased b-phase stability by reducing the content of b-stabilizer alloying element can promote the total amount of the x phase (including both the athermal and deformationinduced x phase) in a Ti alloy. Based on these views, a lower Cr concentration of 10 mass% is determined, which promotes the total amount of the x phase, and the oxygen (a biocompatible element), which can suppress the formation of the athermal x phase [24], is added; thus, an overall enhancement on deformation-induced x-phase transformation can be expected. Even though a larger amount of athermal x phase would be retained upon quenching in Ti–10Cr alloy compared with Ti–12Cr alloy because of the lower b-phase stability, it is reasonable to expect that the athermal x phase could be suppressed in Ti–10Cr–O alloys by optimizing the oxygen concentration; this would lead to an alloy with high deformation-induced x-phase content. In this study, the influence of oxygen content on the microstructural evolution and mechanical properties, such as tensile properties and changeable Young’s modulus, in Ti–10Cr–O alloys was systematically examined. Moreover, the bending springback and cytocompatibility of the optimized alloy were investigated to assess the suitability of the material for use in spinal fixation devices.

2. Materials and methods 2.1. Materials preparation A series of ternary Ti–10Cr–xO alloys (mass%, x = 0.06, 0.2, 0.4, 0.6) was prepared by cold-crucible levitation melting in a highpurity argon atmosphere. The oxygen concentration was controlled by the appropriate addition of TiO2 with a purity of 99.9%. The chemical compositions of all the fabricated alloys were determined by conventional chemical and gas analyses; the metallic elements were determined using an inductively coupled plasma optical emission spectrometry, oxygen was determined using a helium carrier fusion-infrared absorption method, and nitrogen was determined using a helium carrier fusion-thermal conductivity method. The analyzed chemical compositions of all the alloys are listed in Table 1, and are close to the nominal values. All the ingots were

homogenized in an argon atmosphere at 1373 K for 21.6 ks, followed by quenching in ice water. The ingots were then hot forged and subsequently hot rolled into plates to a reduction ratio of 70% at 1273 K in an argon atmosphere, followed by cooling in air. The rolled plates were solution-treated at 1123 or 1173 K (which temperature is selected depends on the b transus temperature, which varies with the oxygen content) for 3.6 ks in vacuum, after which they were quenched in ice water. To evaluate the mechanical properties of the alloys after deformation, a 10% reduction by cold rolling was imparted to the solution-treated plates; this is intended as a mechanical alternative to bending. The solution treatment and cold rolling are respectively denoted ‘‘ST’’ and ‘‘CR’’. 2.2. Microstructural analysis The phase constitutions of all the alloys were characterized by X-ray diffraction (XRD), using a Bruker D8 Discover 2D X-ray diffractometer with Cu Ka radiation at an accelerating voltage of 40 kV and a current of 40 mA. A two-dimensional detector with a 2h resolution of 0.02° was used. The microstructures were examined by optical microscopy (OM; Olympus BX51), electron backscattered diffraction (EBSD; Quanta 200 3D SEM-TSL) and transmission electron microscopy (TEM; JEOL JEM-2000EX). For the OM and EBSD observations, the specimens were polished to a mirror finish and then etched using an aqueous solution of HF (1 vol.%) and HNO3 (0.5 vol.%) for 1 min. The specimens for TEM observations were first mechanically polished to a thickness of 40 lm, after which the thinned specimens were dimple-ground to a level of 10 lm. Finally, the dimpled specimens were further thinned by ion milling to obtain a foil. TEM observations were conducted at an accelerating voltage of 200 kV. 2.3. Mechanical evaluation The mechanical properties of all the fabricated alloys were evaluated using resonant Young’s modulus measurements and tensile tests. For the modulus measurements, rectangular plate specimens with dimensions 40 mm  10 mm  1.5 mm were cut from the ST and CR plates with a wire electric-discharge machine. The specimens were mechanically polished with SiC waterproof emery paper of up to 2400 grit along the longitudinal directions, which are parallel to the rolling direction, and then the Young’s modulus measurements were conducted at room temperature using a freeresonance method. The specimens for tensile tests were also cut by a wire electricdischarge machine, and had a thickness of 1.5 mm, a width of 3 mm and a gauge length of 13 mm. The longitudinal axes of these specimens were also parallel to the rolling direction. The cut specimens were polished using SiC waterproof papers of up to 2400 grit, and then subjected to the tensile tests at a cross-head speed of 8.33  106 m s1 at room temperature on an Instron-type machine (Shimadzu AGS-20kNG). Strain gauges were attached to the gauge sections of the specimens. A three-point bending loading/unloading test was performed on the optimized alloy subjected to ST at room temperature using the same Instron-type tester at a cross-head speed of 8.33  106 m s1

Table 1 Nominal and analyzed chemical compositions of Ti–10Cr–O alloys (mass%). Alloy

Ti

Cr

O

N

Ti–10Cr–0.06O Ti–10Cr–0.2O Ti–10Cr–0.4O Ti–10Cr–0.6O

Bal. Bal. Bal. Bal.

10.06 9.98 10.23 9.98

0.062 0.214 0.416 0.620

0.0071 0.0035 0.0046 0.0081

354

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

in order to evaluate the springback. A developed biomedical TNTZ alloy [18], which has been investigated for use as the production of implant rods for spinal fixture [25], was chosen as the comparison material. As-received 50% cold-rolled TNTZ plates were solution treated at 1063 K for 3.6 ks in vacuum followed by ice-water quenching. The specimens with a length of 40 mm, a width of 5 mm and a thickness of 1.8 mm were utilized for the bending test. The distance between the two parallel supports for the specimen is 30 mm. The specimen was bent until a deflection of 3.0 mm was achieved, followed by unloading.

The proliferation activity of cells was evaluated in conjunction with starting cell number. Parametric multiple group comparisons were performed by ANOVA. The level of significance was set at a P < 0.05. The morphology of the cells cultured on the specimens was observed using SEM, for which the cells were fixed on the specimen surfaces with a 4% paraformaldehyde solution for more than 20 min and then sputter-coated with platinum. 3. Results and discussion 3.1. Microstructures

2.4. Evaluation of cytocompatibility The cytocompatibility of the optimized alloy subjected to ST was evaluated and compared to those of a-type commercially pure (CP) Ti and (a + b)-type Ti–6Al–4V extra-low interstitials (ELI) alloy (Ti64 ELI), which are commonly used in biomedical applications. The CP Ti and Ti64 ELI are as-received 2 mm thick sheet produced by the Kobe Steel, Ltd. and the VSMPO-AVISMA Corporation, respectively. Since the spinal fixation implant rod must be deformed by bending during operation, it is also necessary to evaluate the cytocompatibility of the deformed rod. Hence, the cytocompatibility of the optimized alloy subjected to CR, which is regarded as the deformed alloy, was also examined. The specimens were cut into disks with a diameter of 10 mm and a thickness of 2 mm. The surface of each specimen, upon which the cells were cultured, was mechanically polished using up to 2400 grit SiC waterproof papers and, finally, a colloidal SiO2 suspension. The specimens were first cleaned and sterilized in an autoclave at 394 K under 1 atm for 15 min. The sterilized specimens were then placed individually into a 24-well plate (n = 3) prior to the cell culture. Human osteoblast cells (cell type: hFOB) were kindly provided by Dr. Xuetao Shi (Tohoku University, Japan) and seeded on each specimen surface at a cell density of 1.5  105 cells ml1 in aliquots of 100 ll, i.e. the number of seed cells was 15,000 to evaluate the cell attachment rate. Dulbecco’s Modified Eagle’s Medium (DMEM)/F12 containing 10% fetal bovine serum (FBS) and 50 lg ml1 penicillin/streptomycin was used as the culture medium. The cell cultures were incubated at 307 K in a 5% CO2 humidified atmosphere for 7 h. Following the cell culture, floating dead and unattached cells were flushed along with the medium by washing the sample surface with PBS three times. Afterward, the cells attached to the surface were lifted using Trypsin–EDTA solution. The numbers of attached cells were counted using C-Chip. For the evaluation of the cell proliferation, osteoblast cells were seeded at a cell density of 2  104 cells ml1 in aliquots of 100 ll, i.e. the number of seed cells was 2000. The cell cultures were incubated at 307 K under a 5% CO2 humidified atmosphere for 6 days. The cells were harvested as described above for the cell attachment test. The numbers of the living cells were finally counted by C-Chip.

Fig. 1 shows the XRD profiles of all the prepared alloys subjected to both ST and CR. The patterns indicate that only the peaks associated with the bcc b phase are observed in both ST and CR conditions, while no peaks related to any other phase are detected. Fig. 2 shows optical micrographs of all the prepared alloys subjected to both ST and CR [14]. The microstructures of all the ST alloys are comprised of equiaxed grains with an average diameter of several hundred micrometers, as shown in Fig. 2a–d. For the CR alloys, straight, band-like structures are observed inside the grains in all the alloys (Fig. 2a0 –d0 ). As the oxygen concentration increases, the volume fraction of the band-like structures seems to decrease gradually. The substantial band-like structures are only observed in the Ti–10Cr–0.06O–CR and Ti–10Cr–0.2O–CR alloys. An EBSD analysis was performed on all the CR alloys in order to identify the band-like structures. Fig. 3 shows the EBSD results for the Ti–10Cr–0.06O–CR and Ti–10Cr–0.6O–CR alloys. Massive band-like structures are observed in the orientation image map (OIM) of Ti–10Cr–0.06O–CR alloy (Fig. 3a). Both the corresponding grain boundary map (Fig. 3b), in which the green lines conform to the boundaries of the band-like structures, and the plot of misorientation angle vs. distance (Fig. 3c) along the arrow marked in the OIM suggest that the misorientation angle between the band-like structures and the matrix is 50.5° around the b direction. Thus, the band-like structures formed in the Ti–10Cr–0.06O–CR alloy are identified as {332} mechanical twins, based on the views expressed in the literature [26–29]. For Ti–10Cr–0.6O– CR alloy, only a few band-like structures are observed in the OIM (Fig. 3a0 ). The red lines observed in the corresponding grain boundary map (Fig. 4b0 ) are shown to match the boundaries of the band-like structures, which indicates that the band-like structures possess a misorientation angle of 10° to the matrix around the b direction. This result is supported by the plot of misorientation angle as a function of distance (Fig. 3c0 ) along the arrow seen in the OIM. The plot also reveals that the band-like structures are comprised of several parts with different crystal orientations, including both continuous and discontinuous crystal rotations. Therefore, the band-like structures observed in the Ti–10Cr–0.6O–CR alloy are far from the {332} mechanical

Fig. 1. XRD profiles of Ti–10Cr–O alloys subjected to (a) solution treatment (ST) and (b) cold rolling (CR).

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

355

Fig. 2. Optical micrographs of Ti–10Cr–O alloys subjected to ST and CR [14]: (a) Ti–10Cr–0.06O–ST [14], (b) Ti–10Cr–0.2O–ST, (c) Ti–10Cr–0.4O–ST, (d) Ti–10Cr–0.6O–ST, (a0 ) Ti–10Cr–0.06O–CR, (b0 ) Ti–10Cr–0.2O–CR, (c0 ) Ti–10Cr–0.4O–CR, and (d0 ) Ti–10Cr–0.6O–CR.

twins that are often observed in deformed b-Ti alloys [26–29]. Based on the characteristics of these band-like structures, they seem to be a kind of deformation band, termed the ‘‘kink band’’, based on the previous study [30]. The same EBSD analysis was also performed on the Ti–10Cr–0.2O–CR and Ti–10Cr–0.4O–CR alloys. It was concluded that with increased oxygen concentration, the band-like structures induced by CR in the Ti–10Cr–O alloys change from {332} mechanical twins to apparent kink bands. Fig. 4 shows the selected-area electron diffraction (SAED) patterns of all the prepared alloys subjected to ST and CR [14]. For ST alloys, weak spots and circular diffuse streaks, both of which are related to the athermal x phase [1,5,24], are detected in the SAED pattern of Ti–10Cr–0.06O alloy in addition to b spots (Fig. 4a); this suggests that there is a small volume fraction of

the athermal x phase formed in the Ti–10Cr–0.06O alloy upon quenching. As the oxygen concentration increases, only athermal-x-phase-associated circular diffuse streaks combining with the b spots are detected in the SAED patterns of the Ti–10Cr–0.2O, Ti–10Cr–0.4O and Ti–10Cr–0.6O alloys (Fig. 4b–d). The intensities of these circular diffuse streaks decrease gradually in the order 0.2, 0.4 and 0.6 mass% oxygen. These results suggest that the athermal x phase formed upon quenching in Ti–10Cr–O alloys can be suppressed by oxygen addition, which is consistent with previous studies [24,31]. Following CR, the intensities of the x reflections in all the examined alloys increase compared with those in the ST condition, and the SAED patterns of all the alloys show clear spots corresponding to the x phase (in combination with spots derived from the b phase), as shown in Fig. 4a0 –d0 . These

Fig. 3. EBSD analysis of (a–c) Ti–10Cr–0.06O–CR and (a0 –c0 ) Ti–10Cr–0.6O–CR alloys: (a) orientation image map (OIM) of Ti–10Cr–0.06O–CR alloy; (b) corresponding grain boundary (GB) map, showing boundaries of band-like structures (green lines) possessing a misorientation angle of 50.5° around the b direction with matrix; (c) misorientations relative to the first point along the arrow marked in (a); (a0 ) OIM of Ti–10Cr–0.6O–CR alloy; (b0 ) corresponding GB map showing boundaries of band-like structures (red lines) with a misorientation angle of 10° around the b direction with matrix; (c0 ) misorientations relative to the first point along the arrow marked in (a0 ).

356

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

Fig. 4. Selected-area electron diffraction patterns of Ti–10Cr–O alloys subjected to ST and CR: (a) Ti–10Cr–0.06O–ST [14], (b) Ti–10Cr–0.2O–ST, (c) Ti–10Cr–0.4O–ST, (d) Ti–10Cr–0.6O–ST, (a0 ) Ti–10Cr–0.06O–CR, (b0 ) Ti–10Cr–0.2O–CR, (c0 ) Ti–10Cr–0.4O–CR, (d0 ) Ti–10Cr–0.6O–CR. The beam direction is parallel to [110] b.

results indicate that the deformation-induced x-phase transformation occurs in all the examined alloys during CR.

modulus along with the large change in Young’s modulus possible make Ti–10Cr–0.2O the preferred alloy for spinal fixation devices.

3.2. Mechanical properties

3.2.2. Tensile properties Fig. 6 shows the tensile properties of all the prepared alloys subjected to both ST and CR. With an increase in oxygen concentration, the tensile strength increases, while the elongation decreases for both ST and CR alloys. This is attributed to the solid-solutionstrengthening effect caused by oxygen addition. After CR, the tensile strength of all the examined alloys increases, whereas the elongation decreases, compared to the ST alloys. This is because of the internal stress field and other deformation effects, such as tangled

3.2.1. Young’s modulus The results of the Young’s modulus measurements performed on all the prepared alloys subjected to ST and CR are shown in Fig. 5a [14]. For ST conditions, the moduli of the alloys first decrease as the oxygen concentration increases from 0.06 to 0.2 mass% and then increase gradually with the oxygen concentration. After CR, all the examined alloys exhibit a higher modulus than those in the ST condition. Since the formation of the deformation-induced x phase was detected by TEM analysis in all the examined alloys subjected to CR, the increase in the moduli of these alloys by CR is attributable to the deformation-induced xphase transformation. With an increase in oxygen concentration from 0.06 to 0.6 mass%, a variation tendency on the degree of increase in modulus by CR is observed; it first increases and then decreases gradually, attaining a peak value of 16 GPa at an oxygen content of 0.2 mass% (see Fig. 5b). In this study, the Ti–10Cr– 0.2O alloy exhibits the lowest Young’s modulus of 78 GPa among all the examined alloys in the ST condition. This value is also much lower than those of SUS316L stainless steel (200 GPa), a-type CP Ti (105 GPa) and (a + b)-type Ti64 ELI alloys (110 GPa) [32,33] that are commonly used in biomedicine. Furthermore, Ti–10Cr– 0.2O alloy also shows the largest change in Young’s modulus induced by the 10% reduction by CR. The low original Young’s

(a)

90

80

70

60

50

ST CR

Ti-1 0

Cr-

0.06 O

ST CR

ST CR

ST CR

Ti-1 Ti-1 Ti-1 0C r 0C r 0Cr -0.6 -0.4 -0.2 O O O

Degree of change in Young’s modulus, E/GPa

Young's modulus, E/GPa

100

Fig. 6. Tensile strength (rb), 0.2% proof stress (r0.2) and elongation of Ti–10Cr–O alloys subjected to ST and CR.

20

(b)

10

0.06O 0 Oxygen

0.2O

0.4O

0.6O

Fig. 5. (a) Young’s moduli of Ti–10Cr–O alloys subjected to ST and CR [14], and (b) corresponding dependence of the degree of change in Young’s modulus by CR on oxygen concentration.

357

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

dislocations and/or deformation-induced phase transformations, caused by CR. The Ti–10Cr–0.2O alloy exhibits a high tensile strength: 1050 MPa in the ST condition and 1180 MPa in the CR condition. These values are higher than those of the commonly used SUS 316L stainless steel (580 MPa), CP Ti (450 MPa), Ti64 ELI (920 MPa) [12,33,34] and TNTZ alloy (510 MPa) [18] that has been developed as a biocompatible metallic material for biomedical applications. The elongation of the Ti–10Cr–0.2O alloy is found to be 10% in the ST condition. Based on the report [34] that the elongation of conventional biomedical titanium alloys is between 10% and 20%, the Ti–10Cr–0.2O–ST alloy shows an acceptable elongation for biomedical applications. Based on these results and the Young’s modulus measurements, the Ti–10Cr– 0.2O alloy is considered acceptable for use in spinal fixation devices. 3.2.3. Bending springback Since surgeons must manipulate the implant rods by bending these to conform to the patients’ spinal curvatures during spinal fixation surgery, three-point bending loading/unloading tests were performed to intuitively evaluate the springback performance of the Ti–10Cr–0.2O–ST alloy. Fig. 7a shows the typical stress–deflection curves of Ti–10Cr–0.2O–ST alloy alongside that of TNTZ–ST alloy. The schematic illustration of quantitative calculation for springback is shown in Fig. 7b. The bent specimen curvature is considered approximately as a circular arc (dotted arc) although the actual curvature seems more like a polygonal line (grey solid line). The geometrical relationship between the bent and original specimen curvatures is illustrated. The springback, g, is defined as the following equation:

ðh  h0 Þ=h

h ¼ 2½p  2arctanðL=DÞ

ð2Þ

h0 ¼ 2½p  2arctanðL=D0 Þ

ð3Þ

Thus, the springback of both the alloys, g, can be obtained based on the above-mentioned equations. The mean values and standard deviations of bending strength and springback of both the alloys are listed in Table 2. The results show that Ti–10Cr–0.2O–ST alloy exhibits a much higher bending strength, almost three times that of the TNTZ-ST. Generally, materials with higher bending strength possess a greater ability to maintain their shape after being manipulated to a fixed deformation. Furthermore, a greater strength allows for smaller devices to produce equivalent forces after spinal fixture surgery. Hence, the Ti–10Cr–0.2O–ST alloy, which shows a much higher bending strength, is more preferable for practical spinal fixation use from the viewpoint of strength. On the other hand, the bending strength and Young’s modulus of a material determine its degree of springback. If the moduli of two materials are the same, the material with the higher strength exhibits a larger springback. Based on the results, it is shown that even though the Ti–10Cr–0.2O–ST alloy possesses a much higher

ð1Þ

where h is the corresponding central angle for the curvature when the specimen is bent to the deflection of 3 mm, and h0 is the central angle for the unloaded specimen’s curvature. It is known that the L is the half distance between two parallel supports for the specimen (15 mm), while the D (3 mm) and D0 are the deflection values of the 3 mm bent specimen and unloaded specimen, respectively. Hence, the h and h0 can be calculated according to the following equations:

Fig. 8. Osteoblast cell attachment rates after cells culturing for 7 h on CP Ti, Ti64 ELI, Ti–10Cr–0.2O–ST and Ti–10Cr–0.2O–CR alloys.

(a)

(b)

Fig. 7. (a) Typical stress–deflection curves of Ti–10Cr–0.2O–ST and TNTZ–ST alloys obtained using three-point bending loading–unloading test with a deflection of 3.0 mm; (b) schematic illustration of calculation for springback.

Table 2 Mean values and standard deviations of bending stress, r and springback, g for Ti–10Cr–0.2O–ST and TNTZ–ST alloys. Springback, g

Bending stress, r (MPa) Ti–10Cr–0.2O

Mean value 1919

Standard deviation 39

Mean value 0.6006

Standard deviation 0.0042

TNTZ

Mean value 737

Standard deviation 10

Mean value 0.5385

Standard deviation 0.0065

358

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

bending strength than the TNTZ–ST alloy, it still exhibits a comparable springback. Thus, in terms of both strength and springback, the Ti–10Cr–0.2O–ST alloy is better suited to spinal fixation devices than the TNTZ–ST alloy. 3.3. Cytocompatibility Fig. 8 shows the osteoblast cell attachment rates (number of attached cells divided by the number of initially seeded cells) after cell culturing for 7 h on CP Ti, Ti64 ELI, Ti–10Cr–0.2O–ST and Ti–10Cr–0.2O–CR alloys. Fig. 8 shows that both the Ti–10Cr– 0.2O–ST and Ti–10Cr–0.2O–CR alloys possess cell attachment rates comparable to those of CP Ti and Ti64 ELI, which are currently widely used in practical biomedical applications. This suggests that the Ti–10Cr–0.2O alloy has potential to support cell attachment as well as the CP Ti and Ti64 ELI do in both ST and CR conditions. Fig. 9a shows the osteoblast cell numbers after a culture time of 6 days on CP Ti, Ti64 ELI, Ti–10Cr–0.2O–ST and Ti–10Cr–0.2O–CR alloys. All the cells cultured on these materials show significant growth compared with the initial state (2000 cells). The cultured cells on the Ti–10Cr–0.2O–ST and Ti–10Cr–0.2O–CR alloys are comparable in number with those on the CP Ti and Ti64 ELI, suggesting that the Ti–10Cr–0.2O alloy supports cell proliferation at least as well as the commonly used CP Ti and Ti64 ELI materials in both ST and CR conditions. An SEM image displaying the morphology of the cells cultured on Ti–10Cr–0.2O–ST alloy is shown in Fig. 9b. The cells exhibit a flattened cellular morphology with a high level of attachment, which indicates that the cells can spread well on the surface of Ti–10Cr–0.2O–ST components. This result indicates that the Ti–10Cr–0.2O–ST alloy shows great biocompatibility, supporting cell adhesion and spreading. 4. Discussion In the X-ray diffractograms of all the examined Ti–10Cr–O alloys, the deformation-induced x phase and oxide layer are not identified. The reason for this is considered to be that the size and volume fraction of deformation-induced x-phase particles are generally very small and the oxide layer formed on the surface of the alloy is so thin. Moreover, since the oxygen contents in the present study are all within the solubility of oxygen in titanium, no oxide phase forms in the examined Ti–10Cr–O alloys, and hence no oxide phase can be identified in the diffractograms. Based on the OM and EBSD analyses, {332} mechanical twins with multiple variants are dominant in the Ti–10Cr–0.06O alloy reduced 10% using CR. With the increased oxygen concentration, the formation of {332} deformation twins is suppressed and the deformation-induced band structures changes from {332} twins to apparent kink bands. A phase stability index

(a)

diagram on Ti alloys has been proposed in previous reports [18,22]. In this diagram, electronic structures are calculated for the alloyed elements in Ti, and two alloying parameters are theoretically determined: the bond order (Bo), and the metal d-orbital energy level (Md). This Bo–Md diagram can be used to predict and develop new Ti alloys showing specific improved properties by adjusting the phase stability of the alloys via optimization of the alloying additions. The Ti–10Cr alloy with a Bo and Md of 2.789 and 2.357, respectively, exists in the twin region of this Bo–Md diagram, which suggests that mechanical twinning can be expected in this alloy during deformation. Hence, {332} mechanical twins develop in Ti–10Cr (with a low oxygen concentration of 0.06 mass%) by CR, consistent with the theoretical prediction, and this is attributed to the b-lattice instability of the alloy related to the lattice modulation [35]. As the oxygen concentration increases, solute oxygen atoms in the crystal lattice would generate crystallographic restraint and disturb the atomic rearrangements that are required for the lattice shear accompanying twinning. Therefore, the formation of twins is suppressed by oxygen addition, and the main deformation band-structures change from the mechanical twins to the apparent kink bands that are reportedly related to the dislocation slip [36,37]. The mechanical properties of a material are closely related to the deformation mechanisms. Among the alloys examined, Ti–10Cr–0.06O exhibits a low yield stress and a large elongation in the ST condition. This is attributable to deformation twinning, which not only decreases the yield stress of the alloy, but also increases the pathways for dislocation glide and reduces the effective gliding distances, thereby reducing the local stress concentration and enhancing the ductility. As the mechanical twinning is suppressed by oxygen addition, dislocation slip becomes the main plastic deformation mode in the alloys, and this leads to an increased yield stress and a decreased elongation owing to the local stress concentration that develops rapidly and is caused by the tangled dislocations. The Young’s modulus data of all the alloys demonstrate that the moduli of the alloys first decrease as the oxygen content increases to 0.2 mass% and then increase gradually with the oxygen concentration in ST conditions. This phenomenon can be reasonably explained as follows. As the oxygen concentration increases from 0.06 to 0.2 mass%, the formation of the athermal x phase is suppressed by the 0.2 mass% oxygen addition in the Ti–10Cr–0.2O alloy. Even though the modulus of the Ti–10Cr–0.2O alloy increases with oxygen content via solid-solution strengthening, the suppression of the athermal x phase plays a larger role, and the modulus decreases overall. Thus, the Young’s modulus of Ti–10Cr–0.2O–ST alloy is lower than that of Ti–10Cr–0.06O–ST alloy. As the oxygen concentration further increases to 0.4 and then 0.6 mass%, there is almost no athermal x phase formed in

(b)

Fig. 9. (a) Cell numbers after culturing for 6 days on CP Ti, Ti64 ELI, Ti–10Cr–0.2O–ST, and Ti–10Cr–0.2O–CR alloys and (b) SEM image of cells cultured on Ti–10Cr–0.2O–ST alloy.

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

all the Ti–10Cr–(0.2, 0.4, 0.6)O–ST alloys. In this range of oxygen content, solid-solution strengthening by oxygen dominates, and the modulus increases gradually. After 10% reduction by CR, the largest change in modulus is obtained at an oxygen concentration of 0.2 mass%. Since the change in modulus is determined by the volume fraction of the deformation-induced x phase [12], it is concluded that the deformation-induced x-phase transformation is dominant in the Ti–10Cr–0.2O alloy compared with the others. Two competing factors related to oxygen content are considered to impact the deformation-induced x-phase transformation in Ti–10Cr–O alloys. One factor is the increased apparent b-lattice stability due to the solid solution of oxygen atoms. Since the tetrahedral sites of the bcc b lattice in Ti–10Cr–O alloys can accommodate a radius of 0.42 Å, while the octahedral sites can accommodate radii of 0.21 and 0.91 Å along and directions, respectively, an O atom (atomic radius 0.65 Å) will preferentially occupy the octahedral sites in Ti–10Cr–O alloys. The distribution of the octahedral sites in a b-bcc lattice of Ti viewed along the direction is schematically illustrated in Fig. 10. The solid blue points correspond to the Ti atoms. Except for these solid blue points, all of the points drawn represent the octahedral sites distributed on the different {110} layers in the b-bcc Ti lattice viewed from the direction. The O atoms will occupy these octahedral sites randomly. The occupancy of these octahedral sites depends on the oxygen concentration. It has been proposed [1] that the formation of the x phase, including both the athermal and deformationinduced x phase, involves an athermal mechanism called ‘‘lattice collapse’’, i.e. the x lattice can be obtained from the b lattice by collapsing one pair of neighboring {111} planes into the intermediate position, leaving the adjacent {111} planes unaltered. The schematic in Fig. 10 shows the {111} planes collapsing via atomic displacement towards direction, indicated by the arrows. Thus, O atoms, which occupy the octahedral sites randomly, will introduce distortion in the b lattice, thereby increasing the resistance to the atomic displacement towards the direction and suppressing the formation of the x lattice. Therefore, the degree of deformation-induced x-phase transformation in Ti–10Cr–O alloys gradually decreases with increasing oxygen concentration. The other factor is the decreased amount of the athermal x phase; this also has an effect on deformation-induced x-phase transformation. Previous studies [38–41] have reported that vacancies, areas with local strain and areas with compositional partitioning are the possible nucleation sites for the x phase. Thus,

(111)

[001]

[112]

359

it is reasonable that the suppression of the athermal x phase will leave more possible nucleation sites for deformation-induced xphase transformation, thereby enhancing this transformation. Among the Ti–10Cr–O alloys, the amount of athermal x phase first decreases significantly as oxygen content increases to 0.2 mass%, and then remains at a nearly constant low value as oxygen content increases from 0.2 to 0.6 mass% (almost no athermal x phase formed in this range of oxygen content). Thus, the degree of deformation-induced x-phase transformation shows the opposite trend, first increasing sharply as oxygen content increases to 0.2 mass%, and then keeping a nearly constant high value as the oxygen content increases further. Due to a combination of both these factors, the overall degree of deformation-induced x-phase transformation would reach a peak value at an oxygen concentration of 0.2 mass%, and then fall gradually with further increases in the oxygen content. This trend is consistent with the trend in the change in Young’s modulus. Thus, it is concluded that these two competing factors reach an optimal balance for producing the deformation-induced x-phase transformation at an oxygen concentration of 0.2 mass%. As de Fontaine et al. reported [24], the athermal x-phase transformation is completely reversible during cooling and reheating. Based on previous studies [42,43], a reverse x- to b-phase transformation could also be observed under extremely high static pressure or high temperature. In the present study, we have not found any evidence to suggest that the deformation-induced x-phase transformation is reversible under deformation in the Ti–10Cr–O alloys. The bending springback properties and cytocompatibility of the Ti–10Cr–0.2O–ST alloy show that it possesses a good combination of high bending strength, acceptable springback, and great cytocompatibility. It is known that chromium in titanium alloys increases the passivation tendency of the alloys [44]. It is believed that a stable passive film, titanium oxide, may easily form on the surface of Ti–10Cr–0.2O–ST alloy because of the chromium and oxygen in solid solution. Hsu et al. [45] has reported that a passive film consisting of TiO2 and Cr2O3 can form on the surface of the Ti– Cr alloys. Thus, the stable, passivating film formed on the Ti–10Cr– 0.2O–ST alloy may create a high resistance to corrosion as reported in the literature [46–48], and prevent metal ions, such as Cr6+, that are the main cause of the chromium toxicity [49], from releasing out of the surface, thus contributing to the great cytocompatibility of the Ti–10Cr–0.2O–ST alloy. It should be noted that the Ti–10Cr– 0.2O alloy also shows a comparably great cytocompatibility in the CR condition. This indicates that the deformation effect by CR, including dislocation structures, mechanical twins and deformation-induced x phases, does not show a deleterious effect on the cytocompatibility of Ti–10Cr–0.2O–CR alloy. It has been reported that the uneven distribution of the alloying elements in the phases results in poor corrosion resistance of Ti alloys [50]. Hence, the deformation-induced x phase, which shows a coherent interface with the b matrix and does not involve any compositional partitioning during its formation, seems likely not to reduce the corrosion resistance of the alloy, and this seems a reasonable explanation for why it does not reduce the cytocompatibility.

[110] [111]

5. Conclusions

[110] Octahedral interstice Ti

Octahedral interstice Octahedral interstice

(110) Layer 1 (110) Layer 0 1/2 d{110} (110) Layer -1

Fig. 10. Schematic illustrations of the lattice collapse mechanism for x-lattice formation via atomic shuffling towards direction in a b lattice, in which the distribution of octahedral interstices is schematically indicated.

In this study, the effect of oxygen content on the microstructure and mechanical properties of Ti–10Cr–O alloys was investigated. It was expected that a large change in Young’s modulus related to the formation of the deformation-induced x phase can be achieved in Ti–10Cr–O alloys by optimization of oxygen concentration. Moreover, the bending springback and cytocompatibility of the optimized alloy were evaluated for use in spinal fixation devices. The following conclusions were obtained:

360

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361

(1) As oxygen concentration increases, the nature of the band structures in Ti–10Cr–O alloys induced by cold rolling change from {332} mechanical twins to apparent kink bands. (2) The Young’s moduli of all the prepared alloys increase following cold rolling, and this change is attributed to the deformation-induced x-phase transformation. (3) The Ti–10Cr–0.2O alloy shows the largest change in Young’s modulus among the alloys examined here. This is the result of two competing factors: increased apparent b-lattice stability and decreased amounts of the athermal x phase. Both of these are caused by oxygen addition, and the most favorable balance for the production of the deformation-induced x phase was at an oxygen concentration of 0.2 mass%. (4) The Ti–10Cr–0.2O alloy shows a high tensile strength (1050 MPa) and an acceptable elongation (10%) in the solution-treated condition. (5) The Ti–10Cr–0.2O alloy possesses a good combination of high bending strength, acceptable springback and great cytocompatibility in the solution-treated condition. This alloy also shows a comparably great cytocompatibility in the cold-rolled condition. Therefore, the Ti–10Cr–0.2O alloy is a suitable structural material for spinal fixture from both the viewpoints of mechanical biocompatibility and cytocompatibility.

Acknowledgements This work was supported in part by the Industrial Technology Research Grant Program in 2009 from the New Energy and Industrial Technology Development Organization (NEDO), Japan, and Grant-in-Aid for Scientific Research (A) and Young Scientists (A) from the Japan Society for the Promotion of Science (JSPS), Japan. Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 3 and 7–10 are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/ j.actbio.2014.10.014. References [1] Banerjee S, Mukhopadhyay P. Transformations related to omega structures. In: Cahn RW, editor. Phase transformations: examples from titanium and zirconium alloys. Oxford: Elsevier; 2007. p. 474–550. [2] Sass SL. The structure and decomposition of Zr and Ti bcc. solid solutions. J Less Comm Metals 1972;28:157–73. [3] Dawson CW, Sass SL. The as-quenched form of the omega phase in Zr-Nb alloys. Metall Trans 1970;1:2225–33. [4] Hickman BS. The formation of omega phase in titanium and zirconium alloys: a review. J Mater Sci 1969;4:554–63. [5] Hanada S, Izumi O. Deformation behaviour of retained b phase in b-eutectoid Ti-Cr alloys. J Mater Sci 1986;21:4131–9. [6] Liu HH, Niinomi M, Nakai M, Hieda J, Cho K. Deformation-induced changeable Young’s modulus with high strength in b-type Ti-Cr-O alloys for spinal fixture. J Mech Behav Biomed 2014;30:205–13. [7] Kuan TS, Ahrens RR, Sass SL. The stress-induced omega phase transformation in Ti–V alloys. Metall Trans A 1975;6:1767–74. [8] Williams JC, de Fontaine D, Paton NE. The x-phase as an example of an unusual shear transformation. Metall Trans 1973;4:2701–8. [9] Nakai M, Niinomi M, Oneda T. Improvement in fatigue strength of biomedical b-type Ti–Nb–Ta–Zr alloy while maintaining low Young’s modulus through optimizing x-phase precipitation. Metall Mater Trans A 2012;43:294–302. [10] Kim HY, Ikehara Y, Kim JI, Hosoda H, Miyazaki S. Acta Mater 2006;54:2419–29. [11] Nakai M, Niinomi M, Zhao XF, Zhao XL. Self-adjustment of Young’s modulus in biomedical titanium alloys during orthopaedic operation. Mater Lett 2011;65:688–90. [12] Zhao XF, Niinomi M, Nakai M, Hieda J, Ishimoto T, Nakano T. Optimization of Cr content of metastable b-type Ti–Cr alloys with changeable Young’s modulus for spinal fixation applications. Acta Biomater 2012;8:2392–400.

[13] Zhao XL, Niinomi M, Nakai M, Miyamoto G, Furuhara T. Microstructures and mechanical properties of metastable Ti–30Zr–(Cr, Mo) alloys with changeable Young’s modulus for spinal fixation applications. Acta Biomater 2011;7: 3230–6. [14] Liu HH, Niinomi M, Nakai M, Hieda J, Cho K. Changeable Young’s modulus with large elongation-to-failure in b-type titanium alloys for spinal fixation applications. Scripta Mater 2014;82:29–32. [15] Steib JP, Dumas R, Mitton D, Skalli W. Surgical correction of scoliosis by in situ contouring: a detorsion analysis. Spine 2004;29:193–9. [16] Sumitomo N, Noritake K, Hattori T, Morikawa K, Niwa S, Sato K, et al. Experiment study on fracture fixation with low rigidity titanium alloy. J Mater Sci 2008;19:1581–6. [17] Hao YL, Li SJ, Sun SY, Yang R. Effect of Zr and Sn on Young’s modulus and superelasticity of Ti–Nb-based alloys. Mater Sci Eng, A 2006;441:112–8. [18] Kuroda D, Niinomi M, Morinaga M, Kato Y, Yashiro T. Design and mechanical properties of new b type titanium alloys for implant materials. Mater Sci Eng, A 1998;243:244–9. [19] Matsumoto H, Watanabe S, Hanada S. Beta TiNbSn alloys with low Young’s modulus and high strength. Mater Trans 2005;46:1070–8. [20] Wood RM. Martensitic alpha and omega phases as deformation products in a titanium-15% molybdenum alloy. Acta Metall 1963;11:907–14. [21] Talling RJ, Dashwood RJ, Jackson M, Dye D. On the mechanism of superelasticity in gum metal. Acta Mater 2009;57:1188–98. [22] Abdel-Hady M, Hinoshita K, Morinaga M. General approach to phase stability and elastic properties of b-type Ti-alloys using electronic parameters. Scripta Mater 2006;55:477–80. [23] Zhao XF, Niinomi M, Nakai M, Hieda J. Effect of deformation-induced x phase on the mechanical properties of metastable b-type Ti–V alloys. Mater Trans 2012;53:1379–84. [24] de Fontaine D, Paton NE, Williams JC. The omega phase transformation in titanium alloys as an example of displacement controlled reactions. Acta Metall 1971;19:1153–62. [25] Tamura T, Kozuka S, Oribe K, Nakai M, Niinomi M. Evaluation of Ti–Nb–Ta–Zr as new medical implants. In: Niinomi M, Akiyama S, Hagiwara M, Ikeda M, Maruyama K, editors. Ti-2007 science and technology. Sendai: The Japan Institute of Metals; 2007. p. 1441–4. [26] Bertrand E, Castany P, Peron I, Gloriant T. Twinning system selection in a metastable b-titanium alloy by Schmid factor analysis. Scripta Mater 2011;64:1110–3. [27] Furuta T, Kuramoto S, Hwang JH, Nishino K, Saito T. Elastic deformation behavior of multi-functional Ti–Nb–Ta–Zr–O alloys. Mater Trans 2005;46:3001–7. [28] Hanada S, Izumi O. Correlation of tensile properties, deformation modes, and phase stability in commercial b-phase titanium alloys. Metall Trans A 1987;18:265–71. [29] Liu HH, Niinomi M, Nakai M, Hieda J, Cho K. Bending springback behavior related to deformation-induced phase transformations in Ti–12Cr and Ti– 29Nb–13Ta–4.6Zr alloys for spinal fixation applications. J Mech Behav Biomed 2014;34:66–74. [30] Yang Y, Wu SQ, Li GP, Li YL, Lu YF, Yang K, et al. Evolution of deformation mechanisms of Ti–22.4Nb–0.73Ta–2Zr–1.34O alloy during straining. Acta Mater 2010;58:2778–87. [31] Qazi JI, Marquardt B, Allard LF, Rack HJ. Phase transformations in Ti–35Nb– 7Zr–5Ta–(0.06-0.68)O alloys. Mater Sci Eng, C 2005;25:389–97. [32] Niinomi M. Recent research and development in titanium alloys for biomedical applications and healthcare goods. Sci Technol Adv Mater 2003;4:445–54. [33] Pilliar RM. Modern metal processing for improved load-bearing surgical implants. Biomaterials 1991;12:95–100. [34] Niinomi M. Mechanical properties of biomedical titanium alloys. Mater Sci Eng, A 1998;243:231–6. [35] Tobe H, Kim HY, Inamura T, Hosoda H, Miyazaki S. Origin of 332 twinning in metastable b-Ti alloys. Acta Mater 2014;64:345–55. [36] Hao YL, Li SJ, Sun SY, Zheng CY, Yang R. Elastic deformation behaviour of Ti– 24Nb–4Zr–7.9Sn for biomedical applications. Acta Biomater 2007;3:277–86. [37] Barsoum MW, Farber L, EI-Raghy T. Dislocations, kink bands, and roomtemperature plasticity of Ti3SiC2. Metall Mater Trans A 1999;30:1727–38. [38] Devaraj A, Nag S, Srinivasan R, Williams REA, Banerjee S, Banerjee R, et al. Experimental evidence of concurrent compositional and structural instabilities leading to x precipitation in titanium–molybdenum alloys. Acta Mater 2012;60:596–609. [39] Nag S, Devaraj A, Srinivasan R, Williams REA, Gupta N, Viswanathan GB, et al. Novel mixed-mode phase transition involving a composition-dependent displacive component. Phys Rev Lett 2011;106:245701. [40] Sukedai E, Shimoda M, Nishizawa H, Nako Y. Nucleation behaviour of b to x phase transformations in b-Type Ti–Mo Alloys. Mater Trans 2011;52: 324–30. [41] Tane M, Nakano T, Kuramoto S, Niinomi M, Takesue N, Nakajima H. x transformation in cold-worked Ti–Nb–Ta–Zr–O alloys with low body-centered cubic phase stability and its correlation with their elastic properties. Acta Mater 2013;61:139–50. [42] Dey GK, Tewari R, Banerjee S, Jyoti G, Gupta SC, Joshi KD, et al. Formation of a shock deformation induced x phase in Zr20Nb alloy. Acta Mater 2004;52: 5243–54. [43] Sikka SK, Vohra YK, Chidambaram R. Omega phase in materials. Prog Mater Sci 1982;27:245–310.

H. Liu et al. / Acta Biomaterialia 12 (2015) 352–361 [44] Donachie MJ. Titanium: a technical guide. 2nd ed. Materials Park, OH: ASM International; 2000. p. 126. [45] Hsu HC, Wu SC, Wang CF, Ho WF. Electrochemical behavior of Ti–Cr alloys in artificial saliva. J Alloy Compd 2009;487:439–44. [46] Hanawa T, Ota M. Characterization of surface film formed on titanium in electrolyte using XPS. Appl Surf Sci 1992;55:269–76. [47] Ong JL, Lucas LC, Raikar GN, Connatser R, Gregory JC. Spectroscopic characterization of passivated titanium in a physiologic solution. J Mater Sci – Mater Med 1995;6:113–9.

361

[48] Hanawa T, Asami K, Asaoka K. Repassivation of titanium and surface oxide film regenerated in simulated bioliquid. J Biomed Mater Res 1998;40:530–8. [49] Shanker AK, Cervantes C, Loza-Tavera H, Avudainayagam S. Chromium toxicity in plants. Environ Int 2005;31:739–53. [50] Geetha M, Kamachi Mudali U, Gogia AK, Asokamani R, Geetha M, Baldev Raj. Influence of microstructure and alloying elements on corrosion behavior of Ti– 13Nb–13Zr alloy. Corros Sci 2004;46:877–92.