Fatigue characteristics of bioactive glass-ceramic-coated Ti-29Nb-13Ta-4.6Zr for biomedical application

S. J. Li ^{a, b}, M. Niinomi ^{, , a}, T. Akahori ^a, T. Kasuga ^c, R. Yang ^b and Y. L. Hao <sup>b

a</sup> Department of Production Systems Engineering, Toyohashi University of Technology, 1-1, Hibarigaoka, Tempaku-cho, Toyohashi 441-8580, Japan
^b Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang 110016, People's Republic of China
^c Department of Materials Science, Nagoya Institute of Technology, Nagoya 466-8555, Japan

Received 12 May 2003;  accepted 21 September 2003.  Available online 21 November 2003.
Biomaterials
Volume 25, Issue 17 , August 2004, Pages 3369-3378

1. Abstract

A new surface-coating method by which CaP invert glass is used to improve the bioactivity of titanium alloys has been developed recently. In this method, the powder of CaP invert glass (CaO-P₂O₅-TiO₂-Na₂O) is coated on the surface of titanium alloy samples and heated between 1073 and 1123 K. With this treatment, a calcium phosphate layer mainly containing
-Ca₃(PO₄)₂ phase can be coated easily on titanium alloy samples. In the present study, the effect of this coating process on the fatigue properties of Ti-29Nb-13Ta-4.6Zr, a new metastable
alloy for biomedical applications, has been investigated. The fatigue endurance limit of the coated alloy was found to be about 15% higher than that of uncoated alloy, as a result of the formation of a hard (
+
) layer and a small amount of the
phase during the coating process. The coating exhibits excellent adhesion to the substrate during the tensile and fatigue tests. Subsequent ageing at 673 K for 259.2 ks greatly improves the fatigue resistance of the coated alloy due to isothermal
phase precipitation, and does not have obvious detrimental effect on the coating properties.

Author Keywords: Author Keywords: Titanium alloys; Biomaterials; Bioactive coating; Fatigue

2. Article Outline

1. Introduction

2. Materials and methods

3. Results

3.1. Microstructure

3.2. Characteristics of calcium phosphate coating

3.3. Tensile properties

3.4. Fatigue properties

4. Discussions

4.1. Factors that influence strength

4.2. Fatigue characteristics of coated alloy

5. Conclusions

Acknowledgements

References

3. 1. Introduction

Titanium alloys have seen increasingly wide use in making load bearing implant parts, and a new metastable
alloy, Ti-29Nb-13Ta-4.6Zr, has recently been developed for biomedical applications [1]. This alloy has low modulus, high strength and toughness, and excellent corrosion resistance. However, similar to other titanium alloys such as Ti-6Al-4V ELI, Ti-5Al-2.5Fe, and Ti-6Al-7Nb for biomedical applications [2, 3 and 4], the alloy also is incapable of forming a chemical bond with bone directly. The failure to join with bone directly causes problems associated with micromotion of the device, which may lead to the final loosing of the implant.

A common method to resolve this problem is physical coating of a thin layer of highly biocompatible calcium phosphate (Ca-P) on the surface of the alloys. These coatings can accelerate the process of bony growth in the vicinity of the prosthesis [5, 6, 7 and 8]. By encouraging the formation of new bone up to and into the surface of a hip prosthesis, the device may be more securely fixed in place. Processes for this purpose include dip coating [9], electron-beam deposition [10], hot isostatic pressing [11], pulsed laser deposition [12] and plasma spraying [13]. Of these methods, plasma spraying is most often used.

Recently, a new surface-coating method using CaP invert glass was developed in order to improve the bioactivity of Ti-29Nb-13Ta-4.6Zr [14]. In this method, the powder of CaP invert glass (CaO-P₂O₅-TiO₂-Na₂O) is coated on the surface of Ti-29Nb-13Ta-4.6Zr and heated between 1073 and 1123 K. By this treatment, a Ca-P layer mainly containing
-Ca₃(PO₄)₂ phase can be coated easily on Ti-29-13Ta-4.6Zr. The coating shows strong joining with the substrate, where a compositionally gradient layer in TiO₂-P₂O₅-Na₂O-CaO system can automatically form on the alloy during heating. The adhesive strength of the coating layer is much higher than that of plasma-sprayed HA coating [15]. By soaking the coating layer in simulated body fluid, a new hydroxyapatite phase can be formed on the surface of the coating layer, suggesting that the bioactivity of Ti-29Nb-13Ta-4.6Zr can be improved with this method. Since the biological compatibility of the implant is improved, the alloy will experience longer lifetime and fatigue will become an important material property. It is therefore important to investigate the effect of the coating process on the fatigue properties of Ti-29Nb-13Ta-4.6Zr. In this study, Ti-29Nb-13Ta-4.6Zr was coated using the new coating method described above and the fatigue properties of such bioactive glass-ceramic-coated Ti-29Nb-13Ta-4.6Zr were investigated.

4. 2. Materials and methods

Ti-29Nb-13Ta-4.6Zr used in this study was fabricated by induction skull melting using pure Ti, Nb, Ta and Zr as raw materials and then hot forged to rods with a diameter of 12 mm. The chemical composition of the as-forged alloys according to wet chemical and gas analysis is shown in Table 1. The bars were sectioned and solution treated at 1063 K for 3.6 ks followed by water quenching (scheme TNTZ in Table 2). Then, these solution-treated specimens were machined to the dimensions given in Fig. 1 for tensile and fatigue tests according to ISO6892-84 and ISO1099-75 specifications, respectively. In order to eliminate the notches formed during the machining process, the specimens were ground with successive SiC papers from 57 to 10 
m and polished mechanically to obtain a `mirror-like' surface using colloidal silica suspension.

Table 1. Chemical composition of forged Ti-29Nb-13Ta-4.6Zr (mass%)

Table 2. Scheme of heat treatment, sandblasting treatment and dip-coating treatment

A: 1063 K, 3.6 ks, WQ, under argon protection; B: sandblasting; C: dip coating; D: 1073 K, 3.6 ks, FC, in air; E: 673 K, 259.2 ks, FC, in vacuum furnace.

(6K)

Fig. 1. Geometries of: (a) tensile, and (b) fatigue specimens employed in this investigation (in mm).

Some of the polished tensile and fatigue specimens were coated with a Ca-P layer containing bioactive
-Ca₃(PO₄)₂ phase using the method outlined in the preceding section (scheme TNTZ1 in Table 2). These specimens were first sandblasted and washed thoroughly using acetone in an ultrasonic bath. Then the specimens were dipped into the glass-powder slurry (mixture of Ca-P₂O₅-TiO₂-Na₂O glass powders and water) and drawn up at a speed of 1.4 mm/s. Subsequently, the dip-coated specimens were dried at 373 K. The specimens with a glass-powder layer were finally heated in air at 1073 K for 3.6 ks and cooled to room temperature in an electric furnace. For comparison, some polished but uncoated fatigue specimens were also heat treated in air at 1073 K for 3.6 ks and furnace cooled to room temperature (scheme TNTZ2 in Table 2).

In order to improve the fatigue resistance of the coated Ti-29Nb-13Ta-4.6Zr, ageing treatment was carried out. Ageing response of the coating properties of Ti-29Nb-13Ta-4.6Zr has been investigated by Yamaguchi [16]. In their study, the coated samples were aged at 673 K for 1.8, 3.6, 10.8, 18.0, 25.2, 43.2, 64.8, 86.4, 172.8 and 259.2 ks, followed by furnace cooling. Subsequently, Vickers hardness measurements were conducted on the surface of the coated samples with a load of 10 kg for 15 s. Because the penetration depth of the indenter is larger than the thickness of coating, the hardness of coated samples shown in Fig. 2 represents a mixture of the hardness of the coating and that of the substrate. A total of 12 microhardness measurements were carried out on each specimen. The highest hardness of the coated sample was achieved by ageing at 673 K for 259.2 ks (see Fig. 2). Based on this study, some coated tensile and fatigue specimens were aged at 673 K for 259 ks followed by cooling in a vacuum furnace to investigate ageing response of fatigue properties of the coated Ti-29Nb-13Ta-4.6Zr (scheme TNTZ3 in Table 2).

(7K)

Fig. 2. Comparison of age hardening characteristics of Ti-29Nb-13Ta-4.6Zr solution treated at 1063 K for 3.6 ks between uncoated samples and those coated with Ca-P glass ceramic [16].

The tensile test was conducted at a cross-head speed of 0.5 mm/min using an Instron type machine. The fatigue properties of these coated specimens were evaluated using an electro-hydraulic-servo testing machine (at a stress ratio, R, of 0.1 and a frequency of 10 Hz). All the tests were carried out in air at room temperature.

The microstructures and fracture surfaces of the specimens were observed using a scanning electron microscope (SEM). The specimens for the microstructural observations were mechanically polished and then etched in a solution consisting of 8 vol% HF, 15 vol% HNO₃ and 77 vol% H₂O. Phase constitutions were examined with an X-ray diffractometer using a CuK
irradiation with an accelerating voltage of 40 kV and a current of 250 mA.

5. 3. Results

3.1. Microstructure

Fig. 3(a) shows an SEM micrograph of (
+
″) microstructure of TNTZ substrate. X-ray diffraction analysis shows the presence of (0 2 0), (1 1 3) and (0 0 4) peaks of orthorhombic
″ martensite formed in the water quenched specimens (see Fig. 4(a)). The sintering treatment at 1073 K for 3.6 ks in air followed by furnace cooling results in the presence of
phase, which probably formed in TNTZ1 substrate during the slow furnace cooling, as revealed by X-ray diffraction analysis (see Fig. 4(b)). After ageing at 673 K for 259.2 ks, a large amount of
phase precipitated isothermally from the
grains in TNTZ3 substrate, as evidenced by stronger diffraction intensity on the X-ray spectrum shown in Fig. 4(c). The
particles formed in both TNTZ1 and TNTZ3 were too fine to be identified during SEM observation. From the SEM images of Figs. 3(b) and (c), therefore, only
grains are observed.

(60K)

Fig. 3. SEM micrographs of: (a) TNTZ, (b) TNTZ1, and (c) TNTZ3 substrates.

(7K)

Fig. 4. X-ray diffraction profiles of: (a) TNTZ, (b) TNTZ1, and (c) TNTZ3 substrates.

3.2. Characteristics of calcium phosphate coating

Typical cross-sectional views of the interface between the coating and substrate are shown in Fig. 5. Under the experimental conditions of the present work, a coating layer with a thickness of about 5-10 
m forms on the substrate. The coating layer adheres to the surface tightly and no serious cracks and defects were found in the layer (see Fig. 5(a)). After ageing at 673 K for 259.2 ks, small microcracks were observed between the coating and substrate (see Fig. 5(b)). The adhesive strength between the coating and the substrate of the studied alloy has been investigated previously using a tensile method modified from ASTM C-633 specification [16]. The results showed that, for specimens with coating about 5 
m in thickness, the adhesive strength is about 25 MPa and ageing at 673 K for 259.2 ks results in the decrease of the adhesive strength to about 20 MPa. This may be due to the adverse effect of microcrack formation or phase transformation of the coating during ageing and subsequent cooling.

(32K)

Fig. 5. Cross-section SEM micrographs showing Ca-P coating, interface and substrate of: (a) TNTZ1, and (b) TNTZ3 before fatigue testing.

Below the coating layer, some needle-like
particles precipitate from the TNTZ1 and TNTZ3 substrate due to the inward diffusion of oxygen which is a strong
stabilizer, and the oxygen diffusion layers are seen clearly from Fig. 6. Hardness test shows that the surface hardness of TNTZ1 and TNTZ3 is higher than that of the substrate (see Fig. 2). Because the ageing treatment has been conducted in vacuum, the thickness of the oxygen diffusion layer of TNTZ3 is almost the same as that of TNTZ1 samples as can be seen from Fig. 6. The SEM observation and hardness test confirm that heat treatment in air at 1073 K for 3.6 ks followed by furnace cooling results in the formation of the hard oxygen diffusion layer below the coating. The surface morphology of the coating is shown in Fig. 7. Glass particles are round-shaped and many of them have been sintered. Many pores of several micrometres in size are observed. X-ray diffraction analysis presented in Fig. 8 confirms that the main constituent of the coating layer is
-Ca₃(PO₄)₂. Some amount of the
-Ca₂P₂O₇ phase was also detected. Comparison of the two X-ray diffraction spectra of Fig. 8 shows that the ageing treatment does not lead to obvious change of the phase constitution of the layer, although some slight changes in the relative intensity of the peaks due to unidentified phase(s) can be noted.

(45K)

Fig. 6. Cross-sectional SEM micrographs showing the precipitation of the
phase from the oxygen diffusion zone below the coating layer of: (a) TNTZ1, and (b) TNTZ3.

(22K)

Fig. 7. SEM surface morphology of TNTZ1.

(8K)

Fig. 8. X-ray diffraction profiles of: (a) TNTZ1, and (b) TNTZ3 coating layers.

3.3. Tensile properties

The tensile properties of the specimens were evaluated in both uncoated and coated conditions as shown in Fig. 9. The ultimate tensile strength and yield strength of TNTZ1 (coated) are much higher than those of TNTZ (as-solution-treated). After ageing treatment at 673 K for 259.2 ks, both the ultimate tensile strength and yield strength increase greatly whereas the elongation decreases. The above results demonstrate that both the coating process conducted in the present study and the subsequent ageing treatment increase strength at the expense of elongation in comparison with the case of as-solution-treated specimens. It should be noted that during the tensile tests, the coating layer well adheres to the substrate even in large deformation conditions (see Fig. 10). A micrograph showing the morphology of the coating surface in the large deformation area on a tensile tested TNTZ1 sample is presented in Fig. 10(b). The coating layer is cracked and displays a ripple-like morphology. However, even in this case, the coating layer has not peeled off from the substrate. The TNTZ3 samples did not experience large deformation during tensile tests. In spite of the presence of small microcracks in the coating, the coating still tightly adheres to the substrate. These observations are consistent with experimental results of Yamaguchi [16], namely, the adhesion of the coating and the substrate is very strong and the ageing treatment does not have obvious effect on the adhesive strength.

(6K)

Fig. 9. Tensile properties of TNTZ, TNTZ1 and TNTZ3.

(51K)

Fig. 10. SEM micrographs of one TNTZ1 specimen after tensile test showing morphologies of coating: (a) near the edge of the tensile fracture surface, and (b) near the necked region of the sample.

3.4. Fatigue properties

In order to elucidate the effect due to the presence of coating on the fatigue properties of Ti-29Nb-13Ta-4.6Zr, fatigue specimens were divided into four groups according to Table 2. Fig. 11 presents the S-N curves of the four-group specimens. It can be seen from the plots that the fatigue endurance limit of TNTZ1 (coated) increases by about 15% over that of TNTZ. Ageing treatment of the coated samples (TNTZ3) leads to a further increase of the fatigue limit by about 65%.

(7K)

Fig. 11. S-N curves of TNTZ, TNTZ1, TNTZ2 and TNTZ3.

The fatigue fracture surfaces of uncoated and coated specimens observed with an SEM are shown in Fig. 12. In the high-cycle fatigue (HCF) regime (N>1×10⁵), the crack initiation site in the uncoated specimen was found to be inside the sample (see Fig. 12(a)). However, the initiation site in TNTZ1 and TNTZ3 changes to the grit-blast pits on the prior surface of the substrate (see Figs. 12(b) and (c)). In the low-cycle fatigue regime, crack initiation sites in both the coated and uncoated specimens were located at the surface of the substrate. For all the coated specimens, no fatigue microcracks were observed in the coating layer. SEM observation confirms that for coated specimens, the critical fatigue cracks formed inside the substrate itself, as opposed to nucleating in the coating and propagating across the interface into the substrate [17, 18, 19 and 20]. The coating layer well adhered to the substrate both during fatigue and during final fracture (see Fig. 12(c)). Even after 10⁷ cycles, the coating layer still adhered to the substrate tightly and no delaminations were observed.

(72K)

Fig. 12. General views of fatigue fracture surfaces of: (a) TNTZ (N=899767), and (b) TNTZ1 (N=557862). An area framed in (b) showing the initiation site of the main crack is shown in detail in (c).

6. 4. Discussions

4.1. Factors that influence strength

The above results show that the ultimate strength and yield strength of TNTZ1 and TNTZ3 (both coated) increased over that of the uncoated TNTZ, especially for the coated and aged TNTZ3. The
-titanium alloys can be strengthened by controlling the amount of the
and
phases [21 and 22], so the increase of strength can be related to the
and
precipitations during the coating process and subsequent ageing treatment. SEM observations and X-ray results indicate that a hard (
+
) layer formed beneath the coating layer during sintering and some
phase formed in TNTZ1 during furnace cooling in air. These
and
particles result in the increase of strength for TNTZ1 over TNTZ. For TNTZ3, a large amount of isothermal
phase precipitated during the subsequent ageing treatment (see Fig. 4); thus the significant increase of strength of TNTZ3 over TNTZ1 as seen from Fig. 9 must be due to the isothermal
precipitation.

A peculiar feature seen in Fig. 2 is that, after ageing at 673 K, the hardness of the coating increases much more rapidly than that of uncoated alloy samples. Because the ageing was conducted in a vacuum furnace, the increase in hardness with ageing time for the uncoated samples can be regarded as solely due to the isothermal precipitation of the
phase. For the coated samples, although the X-ray diffraction patterns of Fig. 8 show some changes of the unidentified phase(s), the possibility that such changes will cause the pronounced variation of hardness as seen in Fig. 2 is small. The reason for the rapid hardness increase must be perceived in relation to changes occurring in the substrate. For the
to
phase transformation, the lattice constants of both the parent and the product phase can be calculated from the X-ray diffraction peaks shown in Fig. 4: A lattice constant of 0.329 nm was obtained for the
phase from Fig. 4(b), whereas for the hexagonal
phase, Fig. 4(c) gives a=0.457 nm and c=0.286 nm. It can be shown on the basis of these lattice constants that there is a contraction of volume by about 3% in the transformed regions. Because the volume fraction of the
phase is appreciable as evidenced by its strong diffraction peaks in Fig. 4(c), the shrinkage of the substrate after ageing cannot be ignored. Such a shrinkage will cause a compressive stress in the coating layer which, understandably, will increase the hardness as seen in Fig. 2. The composite system comprising the stressed coating/substrate pair will also contribute to the increase of strength of TNTZ3.

4.2. Fatigue characteristics of coated alloy

It can be seen from Fig. 11 that the coated TNTZ1 samples have higher fatigue limit than the uncoated TNTZ samples. During the coating process, the specimens were first sandblasted before dip coating and then subjected to a sintering treatment at 1073 K after dip coating. These two treatments may contribute to improved resistance to fatigue crack initiation and/or reduction in crack growth rate of coated TNTZ1 specimen. After the sintering treatment at 1073 K for 3.6 ks in air followed by furnace cooling, the ultimate strength and yield strength of TNTZ1 (coated) increased by about 200 and 400 MPa than that of the uncoated TNTZ alloy, respectively, due to the formation of a hard (
+
) layer beneath the surface of Ti-29Nb-13Ta-4.6Zr as a result of inward diffusion of oxygen, as well as due to the presence of some
phase that formed during furnace cooling. The
particles formed in the
matrix are very fine and it has been suggested that the initiation of the fatigue crack can be prohibited by fine
precipitation [23]. In addition, the hard diffusion layer formed below the coating layer also significantly affects the initiation and early growth of fatigue cracks [24 and 25], and this will be beneficial for fatigue strength. Fatigue crack growth rate, another factor influencing fatigue process, was not quantitatively measured in this study. But previous studies show that the fatigue crack growth rate can be increased by the precipitation of fine
phase [26]. Therefore, it can be deduced that the improvement of the fatigue resistance of TNTZ1 is mainly due to the improved resistance to fatigue crack initiation in TNTZ1 caused by sintering treatment at 1073 K for 3.6 ks in air followed by furnace cooling.

However, comparison of the S-N curves in Fig. 11 also indicates that the fatigue limit of TNTZ1 (coated) is not as good as that of TNTZ2. This deterioration might be due to a surface roughening procedure of the coating process—sandblasting. The effects of sandblasting on the fatigue resistance have been widely studied [27, 28, 29, 30, 31, 32 and 33]. In general, sandblasting is thought to have two effects on the substrate: the surface roughness will be increased and residual stresses will be set up beneath the surface. Of these two factors, the residual stresses are beneficial to fatigue strength whereas the increased surface roughness brings about detrimental effect. In this study, because the sandblasted specimens were heat treated at 1073 K (during sintering of the coating powder) that is well within the temperature range for the stress relieving treatments [33 and 34], the residual stress set up in the surface layer of the substrate during sandblasting will be relieved by the heat treatment. Thus the decreased fatigue resistance should be related to the increased surface roughness caused by sandblasting. SEM examination of the grit-blasted TNTZ surface as illustrated in Fig. 13 reveals deep surface scars caused by the impact of the particles. These surface defects can promote the initiation of fatigue cracks by causing local stress concentrations. Fig. 12 shows that the critical fatigue cracks initiation site in HCF life region for the polished TNTZ (uncoated) specimens is in the interior of the specimens but the initiation site changed to the prior surface of the substrate for coated TNTZ1 alloy. This result is consistent with the study of Forrest et al. [35] on the fatigue properties of the plasma sprayed hydroxyapatite-coated titanium alloy. These authors found that fatigue crack is prone to initiating at the grit-blast pits caused by sandblasting treatment, and this contributes to the reduced fatigue limit of specimens. Although the fatigue limit of TNTZ1 specimens could be reduced to some degree by sandblasting treatment, it is still higher than that of uncoated TNTZ. Therefore, the factor that dominates the fatigue behaviour of TNTZ1 is higher strength due to
precipitation and the formation of a hard layer beneath the coating layer, which will improve the resistance to fatigue crack initiation in TNTZ1 (coated). Subsequent ageing treatment of the coated samples (TNTZ3) greatly increases their ultimate tensile strength and yield strength, and X-ray diffraction results show that large amount of
phase precipitated in the
grains during ageing treatment. In this case, the initiation of the fatigue crack will be more difficult. Accordingly, the fatigue strength also increases to a large degree (see Fig. 11). These results suggest that ageing treatment at the intermediate temperatures provides an effective way to reducing the adverse effect on fatigue resistance due to sandblasting treatment which is necessary for achieving good adhesion of the Ca-P coating.

(31K)

Fig. 13. SEM micrograph of sandblasted TNTZ surface.

It should be noted that all the fatigue tests were performed in air instead of in simulated body fluid environment. Kasuga et al. [14] investigated the change of the studied coating at body temperature in the simulated body fluid and their results showed that hydroxyapatite forms on the surface of the coating after soaking for 10 days. Such change is not expected to affect fatigue limit of the coated alloy because the preceding discussion clearly shows that it is the mechanical and metallurgical processes associated with the coating procedure (including sandblasting, dip-coating as well as solution treatment and furnace cooling in air) rather than the coating per se that have the main effect on the fatigue limit. As to uncoated Ti-29Nb-13Ta-4.6Zr alloy, experimental results of the authors' group showed that the simulated body fluid has no detectable effect on the fatigue limit [36]. Therefore, the fatigue limit of the studied alloy in air is expected to be similar to that in simulated body fluid.

7. 5. Conclusions

In this study, a Ca-P layer containing bioactive
-Ca₃(PO₄)₂ phase was coated on the surface of Ti-29Nb-13Ta-4.6Zr alloy using a technique developed by Kasuga et al. The influence of the coating procedures on the fatigue properties of Ti-29Nb-13Ta-4.6Zr was investigated in detail and the following conclusions can be drawn from this research:

(1) The application of sintering treatment at 1073 K for 3.6 ks followed by furnace cooling in air, as part of the coating process, leads to the formation of some
phase during the slow cooling and the formation of a hard (
+
) layer beneath the surface of Ti-29Nb-13Ta-4.6Zr due to inward diffusion of oxygen. Both factors result in increase in tensile strength of the alloy.

(2) The Ca-P coating layer shows excellent adhesion to the substrate during the tensile and fatigue tests. No delaminations of the coating were observed during the fatigue tests and final fracture.

(3) The fatigue resistance of the coated Ti-29Nb-13Ta-4.6Zr is improved due to the increase in resistance to fatigue crack initiation caused by both the formation of the hard (
+
) layer beneath the coating and the precipitation of
from the substrate during the slow cooling.

(4) Ageing at 673 K for 259.2 ks greatly increases fatigue limit of the coated Ti-29Nb-13Ta-4.6Zr and effectively relieves the detrimental effect of sandblasting on fatigue properties. The ageing treatment does not bring about obvious detrimental effect on the coating properties such as cohesive strength.

8. Acknowledgements

This study is in part supported by NEDO, JSPS, Ministry of Education, Science and Culture (Tokyo, Japan), The Iron and Steel Institute of Japan (Tokyo, Japan), The Light Metal Education Foundation (Osaka, Japan), and Suzuki Foundation (Hamamatsu, Japan). SJL's visit to Toyohashi University of Technology is made possible by a studentship from the Ministry of Education, Science and Culture of Japan (Tokyo, Japan). The work of SJL, RY and YLH at IMR is supported by the Chinese NSF under grants No. 59925103 and 50828101, and the MoST of China (grant No. TG2000067105).

9. References

1. M. Niinomi, Mechanical properties of biomedical titanium alloys. Mater Sci Eng A243 (1998), pp. 231-236. SummaryPlus | Full Text + Links | PDF (330 K)

2. K. Wang, The use of titanium for medical applications in the USA. Mater Sci Eng A213 (1996), pp. 134-137. Abstract | PDF (325 K)

3. Zwicker U, Büehler K, Müeller R, Gustavson LJ. Mechanical properties and tissue reactions of a titanium alloy for implant material. In: Kimura H, Izumi O, editors. Titanium'80: Science and Technology, Proceeding of the Fourth International Conference on Titanium. Japan: Kyoto; May 1980. p. 505-14.

4. M. Semlitsch, F. Staub and H. Webber, Titanium-aluminum-niobium alloy: development for biocompatible, high strength surgical implants. Biomed Tech 30 (1985), pp. 334-339. Abstract-EMBASE | Abstract-MEDLINE  

5. P. Ducheyne, Bioceramics: material characteristics versus in vivo behaviour. J Biomed Mater Res 21 A2 (1987), pp. 219-236. Abstract-MEDLINE  

6. T. Fujiu and M. Ogino, Difference of bone bonding behaviour among surface active glasses and sintered apatite. J Biomed Mater Res 18 7 (1984), pp. 845-859. Abstract-EMBASE | Abstract-Compendex | Abstract-MEDLINE  

7. G. Daculsi, R. LeGeros, E. Nevy, K. Lynch and B. Kerebel, Transformation of biphasic calcium phosphate ceramics in vivo: ultrastructural and physico-chemical characterization. J Biomed Mater Res 23 (1989), pp. 883-894. Abstract-MEDLINE | Abstract-Compendex | Abstract-EMBASE  

8. B.M. Tracy and R.H. Doremus, Direct electron microscopy studies of the bone-hydroxyapatite interface. J Biomed Mater Res 18 (1984), pp. 719-726. Abstract-EMBASE | Abstract-Compendex | Abstract-MEDLINE  

9. T. Li, J. Lee, T. Kobayashi and H. Aoki, Hydroxyapatite coating by dipping method and bone bonding strength. J Mater Sci Mater Med 7 (1993), pp. 355-357.

10. J.M. Choi, Y.M. Kong, S. Kim, H.E. Kim and C.S. Hwang, Formation and characterization of hydroxyapatite coating layer on Ti-based metal implant by electron-beam deposition. J Mater Res 14 7 (1999), pp. 2980-2985.

11. H. Hero, H. Wie, R.B. Jorgensen and L.E. Ruyter, Hydroxyapatite coatings on Ti produced by hot isostatic pressing. J Biomed Mater Res 28 (1994), pp. 343-348. Abstract-EMBASE | Abstract-Compendex  

12. J.M. Femández, G. Sardin, L. Clčries, P. Serra, C. Ferrater and J.L. Morenza, Deposition of hydroxyapatite thin films by excimer laser ablation. Thin Solid Films 317 (1998), pp. 393-396.

13. K. de Groot, R. Geesingk, C.P.A.T. Klein and P. Serekian, Plasma sprayed coatings of hydroxyapatite. J Biomed Mater Res 21 (1987), pp. 1375-1381. Abstract-MEDLINE | Abstract-Compendex  

14. T. Kasuga, T. Mizuno, M. Watanabe, M. Nogami and M. Niinomi, Calcium phosphate invert glass-ceramic coatings joined by self-development of compositionally gradient layers on a titanium alloy. Biomaterials 22 (2001), pp. 577-582. SummaryPlus | Full Text + Links | PDF (844 K)

15. Y.C. Tsui, C. Doyle and T.W. Clyne, Plasma sprayed hydroxyapatite coatings on titanium substrates—Part 1: mechanical properties and residual stress levels. Biomaterials 19 (1998), pp. 2015-2029. Abstract | PDF (676 K)

16. M. Niinomi, T. Akahori, T. Yamaguchi, T. Kasuga, H. Fukui and A. Suzuki, Ageing characteristics and mechanical properties of Ti-29Nb-13Ta-4.6Zr coated with calcium phosphate glass-ceramic for biomedical applications. J Jpn Inst Metals 67 (2003), pp. 604-613. Abstract-INSPEC | Abstract-Compendex  

17. Y. Sugimura and S. Suresh, Fatigue of coated materials. Ceram Coat 44 (1993), pp. 9-14. Abstract-Compendex  

18. S. Suresh, Y. Sugimura and E.K. Tschegg, The growth of a fatigue crack approaching a perpendicularly oriented bimaterial interface. Scr Metall Mater 27 (1992), pp. 1189-1194. Abstract

19. Y. Sugimura, P.G. Lim, F. Shih C and S. Suresh, Fracture normal to a bimaterial interface: effects of plasticity on crack-tip shielding and amplification. Acta Metall Mater 43 (1995), pp. 1157-1169. Abstract | PDF (1180 K)

20. S. Suresh, Y. Sugimura and T. Ogawa, Fatigue cracking in coated materials with brittle surface coatings. Scr Metall Mater 29 (1993), pp. 237-242. Abstract

21. B.S. Hickman, Omega phase precipitation in alloys of titanium with transition metals. Trans Met Soc AIME 254 (1969), pp. 1329-1336.

22. J.C. Williams, B.S. Hickman and H.L. Marcus, The effect of omega phase on the mechanical properties of titanium alloys. Metall Trans 2 (1971), pp. 1913-1919.

23. M. Niinomi, Fatigue performance and cyto-toxicity of low rigidity titanium alloy, Ti-29Nb-13Ta-4.6Zr. Biomaterials 24 (2003), pp. 2673-2683. SummaryPlus | Full Text + Links | PDF (801 K)

24. A. Puskar and S.A. Golovin. Fatigue in materials: cumulative damage processes, Elsevier, Amsterdam (1985).

25. J.A. Bannantine, J.J. Comer and H.L. James. Fundamentals of metal fatigue analysis, Prentice-Hall, New Jersey (1990).

26. G.W. Salgat and D.A. Koss, Fatigue crack propagation in single crystal and polycrystalline, age-hardenable Ti-V alloys. Mater Sci Eng 35 (1978), pp. 263-272. Abstract

27. R.T.R. McGrann, D.J. Greving, J.R. Shandley, E.F. Rybicki, T.L. Kruecke and B.R. Bodger, The effect of coating residual stress on the fatigue life of thermal spray-coated steel and aluminum. Surf Coat Technol 108 (1998), pp. 59-64. SummaryPlus | Full Text + Links | PDF (230 K)

28. Y.C. Tsui and T.W. Clyne, An analytical model for predicting residual stresses in progressively deposited coatings—Part 1: planar geometry. Thin Solid Films 306 (1997), pp. 23-33. Abstract | PDF (800 K)

29. Y.C. Tsui and T.W. Clyne, An analytical model for predicting residual stresses in progressively deposited coatings—Part 2: cylindrical geometry. Thin Solid Films 306 (1997), pp. 34-51. Abstract | PDF (801 K)

30. R. Elsing, O. Knotek and U. Balting, Calculation of residual thermal stresses in plasma-sprayed coatings. Surf Coat Technol 43-44 (1990), pp. 416-425. Abstract

31. S.R. Brown, I.G. Turner and H. Reiter, Residual stress measurement in thermal sprayed hydroxyapatite coatings. J Mater Sci 5 (1994), pp. 756-759. Abstract-EMBASE | Abstract-Compendex  

32. G.R. Leverant, B.S. Langer, A. Yuen and S.W. Hopkins, Surface residual stresses, surface topography and the fatigue behaviour of Ti-6Al-4V. Metall Trans A 10A (1979), pp. 251-257. Abstract-INSPEC | Abstract-Compendex  

33. S.L. Evans and P.J. Gregson, The effect of a plasma-sprayed hydroxyapatite coating on the fatigue properties of Ti-6Al-4V. Mater Lett 16 (1993), pp. 270-274. Abstract

34. A.K. Lynn and K.L. DuQuesnay, Hydroxyapatite-coated Ti-6Al-4V—Part 1: the effect of coating thickness on mechanical fatigue behaviour. Biomaterials 23 (2002), pp. 1937-1946. SummaryPlus | Full Text + Links | PDF (512 K)

35. D.J. Forrest, J.C. Shelton and P.J. Gregson, Fatigue failure in a plasma sprayed hydroxyapatite coated titanium alloy. In: Proceeding of the Fifth World Biomaterials Congress, University of Toronto Press, Toronto (1996), p. 324. Abstract-Compendex  

36. Akahori T, Niinomi M, Yabunaka T, Kuroda D, Fukui H, Suzuki A, Hasegawa J. Fretting fatigue characteristics of biomedical new beta titanium alloy. Proc PRICM4: Fourth Pacific Rim International Conference on Advanced Materials and Processing, 11-15 December, 2001, Honolulu, Hawaii, USA. p. 209-12.

Corresponding author. Tel.: [phone]; fax: [phone]

---