Fabrication of a β-based titanium alloy for biomedical applications

. The aim of this study was to produce a titanium-based alloy with mainly β-phase and reduced Young’s modulus for biomedical applications. Alloys Ti-Nb x -Ta 5 -Zr 5 ( x = 20, 30, 40 at.%Nb) were prepared by arc melting then solution annealed at 950℃ for 1 h, and aged at 480℃ for 12 h. Optical microscopy showed mixtures of dendritic and needle-like microstructures before and after heat treatment in all alloys. X-ray diffraction (XRD) identified β-phase in all alloys. Small fractions of orthorhombic martensite (αʹ) and ω-phase were also detected by XRD which decreased after ageing. Alloy Ti-Nb 20 -Ta 5 -Zr 5 had the lowest Young’s modulus, derived from nanoindentation hardness of 69.8 ± 7.2 GPa in the as cast condition. There was no significant change in elastic modulus of the alloy after ageing (70.8 ± 6.8 GPa). As-cast Ti-Nb 30 -Ta 5 -Zr 5 had the highest elastic modulus of 94.7 ± 3.0 GPa. The elastic modulus decreased to 84.4 ± 0.32 GPa after heat treatment.


Introduction
The development of titanium materials for biomedical applications is currently an area of active research world-wide and many serious attempts are made every year to improve materials properties in this field [1]. Beta-based titanium (βTi) alloys are being developed to replace the high elastic modulus commercial (alpha (α) and duplex (α + β) titanium alloys such as commercially pure Ti (CP-Ti) and Ti-6Al-4V in the biomedical sector [2]. The moduli of CP-Ti and Ti-6Al-4V lie between 100-110 GPa [5], which is significantly higher than that of human cortical bone (10-40 GPa) [3]. The high mismatch in elastic moduli of these alloys relative to the human cortical bone is caused by high amounts of α phase, which can lead to osteoporosis and poor osseointegration [4]. The βTi alloys have lower Young's modulus, high strength, superior bio-corrosion resistance and excellent biocompatibility [1,6]. Their elastic moduli can be significantly reduced by adjusting the concentration of β stabilising elements such as Nb, Ta, Zr, Mo, etc. [7].
The β-type Ti-based alloys have been extensively developed and examples include Ti-15Mo, Ti-13Nb-13Zr, Ti-12Mo-6Zr-2Fe, Ti-35Nb-5Ta-7Zr and Ti-29Nb-13Ta-4.6Zr [8][9][10]. Amongst these alloys, Ti-Nb-Ta-Zr alloys have lower Young's modulus in the range of 48 -55 GPa, which is about 50% of that of conventional CP-Ti and Ti-6Al-4V alloys [11][12][13]. However, the lowest Young's modulus reported so far in bulk Ti-based alloys developed for biomedical applications is 40 GPa for the Ti-35Nb-4Sn alloy [14]. The alloying elements Nb and Ta stabilise the β phase and lowering the elastic modulus [15]. Mohammed et al. [1] reported the β phase as the largest contributor to the reduction of the Young's modulus because it has the lowest modulus, 35.29 GPa, than the other phases: α, hexagonal martensite (αʹ), orthorhombic martensite (αʹʹ) and omega (ω) in these Ti alloys. Niobium, acting as a βphase stabiliser and a biocompatible element, has attracted much attention and it has been added to many β-type Ti-based alloys and near β-type Ti-based alloys [2]. Zirconium, which is a neutral element, enhances strength and improves elasticity while suppressing the precipitation of omega (ω) phase when dissolved in titanium [9,16]. Tantalum is expected to contribute to the stabilisation of the β phase and improve mechanical performance [9]. The aim of this work was to fabricate a β Ti-Nb-Ta-Zr alloy with reduced Young's modulus by arc melting and heat treatment, targeted for biomedical applications.

Experimental methods
The Ti-Nbx-Ta5-Zr5 (x = 20, 30, 40 at% Nb) alloys were produced by button arc melting on a water-cooled copper hearth using pure Ti, Nb, Ta and Zr metal powders as raw materials. The as-cast ingots were solution annealed under argon atmosphere at 950℃ for 1 h and quenched, then aged for 12 h at 480℃ followed by furnace cooling for homogenisation and precipitation hardening.
The as-cast and heat treated buttons were analysed for phases using optical microscopy (OM Leica DMI5000 M) and X-ray diffraction (EMPYREAN diffractometer system). The hardnesses and Young's moduli were measured by a Vickers micro-hardness tester (Future Tech. Corp., FM-700) and a nano-indenter (Anton Paar, TTX-NHT 3 ). The buttons were cut, ground and polished, then etched in a 10 vol.% HF, 10 vol.% HNO3 + glycerol solution to reveal the microstructures. X-ray diffraction measurements were carried out at 45 kV and 40 mA using monochromatic Cu Kα radiation (λ = 0.17890 nm). Nano-indentation was done at 400 mN load at a dwell time of 20 seconds. Vicker's micro-hardness test was done at 500 gf at a dwell time of 15 seconds.

Microstructural analysis
The microstructure of the alloys before and after heat treatment are shown in Figure 1. All the alloys had β phase in the as-cast and heat treated conditions. The β stabilising elements and treatment conditions contributed to the formation of dendritic and basket-weave microstructures [17] Figure 1a shows a dendritic structure with some acicular martensite (αʹ), a basket-weave structure, in the as-cast condition. According to ImageJ analysis, increasing niobium content resulted in higher volume fraction of β phase, Table 1. For 40 at.% Nb, less alpha was seen in the as-cast condition. Ageing gave regions of α phase within the main β phase for most alloys, except for Ti-Nb30-Ta5-Zr5, which only had β-rich dendrites and some areas of acicular martensite. Nasakina et al. [18] and Elias et al. [19] obtained similar morphologies of the microstructures of alloy Ti-Nb-Ta-Zr produced by arc melting. Ageing treatment resulted in increased amount of β phase except for alloy Ti-Nb40-Ta5-Zr5.   Figure 2 shows the XRD patterns of the samples in the as-cast and heat treated conditions. Diffraction peaks corresponded to the β phase for all samples with small amounts of ω and α phases, which appeared to diminish after ageing. The XRD results agreed with the microstructures in Figure 1 even though αʹ and ω phases were not identified in the microstructures. Larger intensity peaks of β phase were observed in Ti-Nb20-Ta5-Zr5 after heat treatment. Small fractions of α phase were detected in aged alloys Ti-Nb30-Ta5-Zr5 and Ti-Nb40-Ta5-Zr5 which shows its precipitation [20], although the Ti-Nb30-Ta5-Zr5 OM microstructure does not show the phase. As-cast alloys Ti-Nb20-Ta5-Zr5 and Ti-Nb40-Ta5-Zr5 had similar peaks with higher background counts in the range 42<2θ<50º which were indexed as hexagonal martensite (αʹ) and ω phase.   Figure 3 shows the Vickers micro-hardness results of the alloys. Alloy Ti-Nb30-Ta5-Zr5 had the highest hardness of 398 ± 18.6 HV, as shown in this figure and Table 2. However, the hardness decreased to 350 ± 3.3 HV after heat treatment due to decreased α phase. Alloy Ti-Nb20-Ta5-Zr5 had a slight increase in hardness after heat treatment and this can be attributed to the presence of α phase detected during XRD analysis.    Table 3 shows the Young's modulus of the alloys in their respective conditions. Alloy Ti-Nb20-Ta5-Zr5 had the lowest Young's modulus of 69.8 ± 7.2 GPa in the as-cast condition although there was a large standard deviation. The modulus increased slightly after heat treatment, which was attributed to precipitation of α phase (Figure 1b) even though the phase was not detected by XRD, so must have been less than 4 vol.%. Alloy Ti-Nb30-Ta5-Zr5 had the highest Young's modulus of 94.7 ± 3.0 GPa in the as-cast condition. The modulus reduced to 84.4 ± 0.3 GPa after heat treatment and this can be attributed to a decrease in α phase (Figure 1d). Figure 4 shows nano-indentation test results for as-cast and heat treated samples, with the curves showing the elastic behaviour (EIT). Heat treated Ti-Nb40-Ta5-Zr5 alloy followed the same load-displacement recovery path as the as-cast alloy, with only 2% shift after treatment (%shift = 100 x (new penetration depth -old penetration)/old penetration depth). The loaddisplacement curve of alloy Ti-Nb30-Ta5-Zr5 shifted to the right by 16% after heat treatment, indicating softening. A 6% shift to the left for alloy Ti-Nb20-Ta5-Zr5 was observed after heat treatment, hence a slight increase in hardness. The nano-indentation EIT for all the alloys was in agreement with Vicker's micro-hardness response. Table 3. Elastic modulus of the alloys before and after heat treatment.

Alloy
Elastic Young's modulus (GPa) As-cast HT Ti-Nb20- Ta5 The alloys had mainly β dendrites with small areas of α. Secondary phases such as αʹ (hexagonal), αʹʹ (orthorhombic) and ω phase were not identified under optical microscopy. However, XRD identified the ω phase. The XRD results also showed a high background between 2 = 42 to 50º which were indexed as αʹ and ω phases. These findings require further analysis by scanning electron microscopy (SEM) and more XRD analyses. Alloy Ti-Nb20-Ta5-Zr5 had the desired microstructure and lower Young's modulus. Ageing improved the hardness of the alloy although the precipitation of α phase was not observed.
Ageing resulted in increased amount of β phase for alloys Ti-Nb20-Ta5-Zr5 and Ti-Nb30-Ta5-Zr5. Alloy Ti-Nb40-Ta5-Zr5 undergone precipitation of α phase. • The XRD results confirmed the major β phase in all alloys before and after heat treatment, showing the alloys were successful as  alloys. Small fractions of α, αʹ and ω phases were detected in the samples by XRD and decreased after heat treatment. Alloy Ti-Nb20-Ta5-Zr5 had lower Young's modulus (69.8 ± 7.2 -70.8 ± 6.8 GPa), closer to that of the human cortical bone (10-40 GPa), and good hardness. • Decrease in hardness and elastic Young's modulus for alloy Ti-Nb30-Ta5-Zr5 was attributed to the dissolution of α phase after heat treatment even though a small amount of the phase was detected during XRD analysis. Alloy Ti-Nb40-Ta5-Zr5 has increased hardness and Young's modulus, which was attributed to precipitation of α phase after ageing treatment.