Determination of the Elastic Properties at Aging of Medical Ti-6Al-4V ELI Alloy by Ultrasonic Velocity Measurements

In this paper, we report the experimental data of the elastic properties of the Young’s modulus and shear modulus based on the variation in the ultrasonic velocity parameter during the microstructural evolution in a Ti-6Al-4V alloy with two—bimodal and Widmanstätten—varying microstructures. The two different initial microstructures were treated thermally by aging at 515, 545, and 575°C at different times from 1 min to 576 h to induce a precipitation process. Ultrasonic measurements of shear and longitudinal wave velocities, scanning electron microscopy image processing, optical microscopy, and microhardness measurements were performed, establishing a direct correlation with the measurements of the ultrasonic velocity and the elastic properties developed during the thermal treatment by artificial aging. The results of the ultrasonic velocity show a very clear trend as the aging time progresses, which is affected by precipitation of Ti3Al particles inside the α phase. In this way, we can find, in a fast and efficient way, the elastic properties developed during the heat treatment by long-term aging, since the presence of these precipitates hardens the material microstructure, thus affecting the final mechanical properties.


INTRODUCTION
The use of Ti6Al4V alloy in the aerospace and biomaterials industry [13] has been led by the demand for further development of characterization techniques to figure the microstructureproperty relations in this metallic alloy. It is also of main concern to examine phase transformations that occur during processing in order to design microstructures that possess excellent properties specific to these applications. On the biomedical field, metal implants must have a mechanical compatibility with the substituted bone, which is achieved through the combination of low Youngs modulus, high flexural and fatigue strengths. In the case of the medical Ti6Al4V ELI alloy, when this titanium alloy is overaged at 500600°C, precipitation of nanometer-sized phases such as α 2 (Ti 3 Al) can be observed in the titanium alloy matrix, increasing the mechanical properties of the metallic material [4 7].
On the other hand, the behavior of ultrasonic wave propagation in metals provides information on the microstructure, mechanical and physical properties of the material under investigation. The measurements of the two main ultrasonic parameters such as attenuation and velocity can be used to estimate the material microstructural changes and properties [8]. The ultrasonic wave velocity is related to the elastic constants and density of the medium while the ultrasonic wave attenuation stands on microstructure and crystalline defects [911]. In the present research study, we applied the contact ultrasonic pulse-echo technique using buffer media to calculate the ultrasonic propagation velocity in the artificially aged Ti6Al4V ELI alloy. Combined with the theory of diffraction correction of the ultrasonic velocity in solid media, the influence of the two microstructures (bimodal and acicular) on the ultrasonic propagation velocity in different aging conditions was discussed. In addition, hardness measurements and microstructural examinations using optical and scanning electron microscopy were carried out by relating elastic properties such as Youngs modulus and shear modulus to the observed microstructure.

Material
The material used in this study was a medical Ti 6Al4V ELI alloy in the form of a plate 6.76 mm thick, and its exact chemical composition is 6.10Al3.93V 0.17.e0.12O0.009C0.005N0.002HTi (wt %). This alloy plate was subjected to different heat treatments to obtainWidmanstätten and bimodal microstructures. The Widmanstätten microstructure was obtained by holding the material at 1020°C for 20 min (above the β transformation temperature), then it was cooled down to 850°C and held for 2 h, followed by water quenching. The bimodal microstructure was obtained by holding at 950°C (α + β region) for 1 h, followed by water cooling/quenching. These two microstructures were overaged at 515, 545, and 575°C for different aging times ranging from 1 min to 576 h to promote the formation of very fine α 2 (Ti 3 Al) precipitates.

Microscopic Techniques
.or optical and scanning electron microscopy inspection, heat-treated specimens were mounted in bakelite, ground, and polished down to a 1 µm diamond cloth. Microstructural features were unveiled using Krolls reagent (H 2 O 100 ml, H. 100 ml, HNO 3 100 ml). The main microstructural features for the medical Ti6Al4V ELI alloy were calculated using an image analyzer. To evaluate microhardness variations amongst the specimens, Vickers hardness measurements were carried out by applying a load of 500 g for 30 s to the unaged and aged specimens. Prior to a microhardness test, the specimens were cross-sectioned and metallographically prepared in a similar fashion as for the optical microscopy observation but using a much finer finishing approach with a colloidal silica solution (0.05 µm).

Ultrasonic Technique
Ultrasonic compressional and transverse velocities were calculated by measuring the travel time for the ultrasonic waves through the titanium specimens. .igure 1 illustrates a schematic diagram of the ultrasonic measurements. A broadband longitudinal transducer (Aerotech) with a central frequency of 10 MHz and element diameter of 6 mm was used to generate compressional ultrasonic waves. A broadband shear transducer (Aerotech) with a central frequency of 5 MHz and element diameter of 12 mm was operated to induce transverse ultrasonic waves. In order to evade the nonuniform amplitude near the transducer field, a buffer rod (fused silica) was used to add a time delay between the excitation pulse and echoes returning from the measured specimen. The ultrasonic transducer was placed on the specimen and excited by a Panametrics 5072PR broadband pulser. The ultrasonic signal was digitized and averaged by a LeCroyWavesurfer 432 oscilloscope and then sent to the computer for data processing.
In experimental measurements, the beam diffraction usually generated considerable additions to the ultrasonic velocity parameter as measured by the pulseecho technique [1214]. In order to prevent the effect of diffraction on the echo waveforms, it was necessary to realize corrections to the two echo waveforms. .irst, a normalized distance S is calculated for each echo where z is the propagation distance, a is the transducer radius, and λ is the wavelength. The phase shift leads to a travel time diffraction correction as t = t′ + ∆t, (2) where t′ is the measured travel time between the echoes n and m (first and second), t is the corrected travel time, and ∆t is the travel time correction. The expression for the travel time correction is given by where the phase φ is calculated at the normalized dis-  (1 )(1 2 ) 1 where E is the Youngs modulus (GPa), G is the shear modulus (GPa), σ is the Poisson ratio, c C is the compressional velocity (m/s), s C is the transverse velocity (m/s), and ρ is the density of the specimen (g/cm 3 ).

RESULTS AND DISCUSSION
.igure 2 shows optical micrographs of the Widmanstätten and bimodal microstructures. In the Widmanstätten microstructure, α-phase grains (white/bright) develop along prior β grain boundaries (black/dark) and colonies of the lath-type β and α lamellar structure are observed inside prior β grains (.ig. 2a). The colony size, thickness of α β platelets, volume fraction of the α phase, volume fraction of the β phase were measured to be 973 µm, 5 µm, 56%, and 44%, respectively. The bimodal microstructure consists of equiaxed α grains (white/bright) and tempered martensite with a small amount of residual β phase (black/dark). The α-grain size was measured to be 10 µm, and volume fractions of the α and β phases were measured to be 52 and 48%, respectively, as shown in .ig. 2b.
Table summarizes the main microstructural features in the unaged and aged conditions for the two microstructures of the Ti6Al4V ELI alloy after holding for 2, 144 and 576 h at the three different aging temperatures. It is observed from the analysis data that the aging treatment changes the percent of the α and β phases present initially in the unaged condition for both the Ti6Al4V ELI alloy microstructures (Widmanstätten and bimodal). A slight increase in the percent of the α phase is noticed.
An understanding of the two Ti6Al4V ELI microstructures is essential to interpret the results on ultrasonic velocities as well as the elastic and shear modulus data. There are some differences in the high-magnification scanning electron microscopy micrographs of the aged acicular and bimodal microstructures as shown in .igs. 3 and 4. .ine α 2 -phases 50200 nm in size are homogeneously distributed in α phases (black/ dark) in the acicular microstructure (see .igs. 3d3f). In the Widmanstätten microstructure, the platelet size increases significantly with aging, and thus a higher number density of nucleation sites available for α 2 precipitates is formed (.igs. 3a3c). On the other hand, .ig. 4 also shows that the amount of precipitates in-creases with respect to the aging time for the bimodal microstructure (see .igs. 4d4f). .igures 4a and 4d correspond to the aging time 2 h, in which the presence of α 2 precipitates was observed in grain boundaries of the α phase (black/dark). At 144 h (.igs. 4b and 4e), the precipitates are more homogeneous within and at the α-phase boundary. The size and morphology of the intermetallic α 2 precipitates remain constant. .or the aging time 576 h (.igs. 4c and 4f ), these α 2 precipitates are found in a greater number, being located mainly in large groups (clusters) within the α matrix, as shown in .ig. 4f. The precipitation of these α 2 (Ti 3 Al) nanoparticles induces hardening of the Ti 6Al4V matrix [16].
The alloy element separation effect has the aftermath that α lamellae produced from the β-phase upon cooling have a lower concentration of elements (oxy-.ig. 3. Scanning electron microscopy micrographs of the overaged acicular microstructure at the 1000× (ac) and 30 000× (df) magnification for 575°C at the three different aging times 2, 144, and 576 h, respectively, showing very fine α 2 phases homogeneously distributed in the α phase. Etched with Krolls reagent. gen), which induces age-hardening by the formation of coherent α 2 particles. In this stage, the temperature parameter is more relevant than the time parameter since the temperature being either below or above the α 2 solvus temperature figures whether age-hardening by α 2 particles takes place in the α phase or not. In the medical Ti6Al4V ELI alloy, the α 2 solvus temperature is ~600°C, which indicates that aging at 500°C will precipitate α 2 particles while the heat treatment at 600°C or above will be only a stress relieving treatment. In this research work, the three overaging temperatures (515, 545, and 575°C) will induce age-hardening by Ti 3 Al particles.
.igure 5 presents the Vickers microhardness modification in the two initial medical Ti6Al4V ELI mi-crostructuresWidmanstätten and bimodalas a function of different aging times at 515, 545, and 575°C. The microhardness of the bimodal microstructure is greater than that of the Widmanstätten microstructure. This is probably due to the presence of a huge volume fraction of the tempered martensite in the bimodal microstructure. The microhardness results for the medical Ti6Al4V ELI alloys (.ig. 5) show an important tendency to increase with aging times for both the microstructures. This phenomenon is attributed to the precipitation of α 2 (Ti 3 Al) particles inside the α phase. This is due to a greater volume fraction of the α phase after aging in the microstructures and thus a higher number density of nucleation sites available for α 2 precipitates. The precipitation of these nanoparticles .ig. 4. Scanning electron microscopy bright-field images of the overaged bimodal microstructure at the 1000× (ac) and 30 000× (df) magnification for 575°C by the three different aging times 2, 144, and 576 h, respectively, showing very fine α 2 phases homogeneously distributed in the α phase. Etched with Krolls reagent. occurs mainly at the boundary and within the α 2 -phase, resulting in a microhardness increment. After 300 h, the hardness data tend to remain constant with increasing the aging time due to the thickening of α β platelets (Widmanstätten) and α grains (bimodal) because of precipitates of smaller sizes.
.igure 6 shows the variations in the Youngs and shear moduli in the Widmanstätten and bimodal microstructures with different aging parameters at the medical TiAl4V ELI alloy. The Youngs and shear moduli were calculated from Eqs. (4) and (5), respectively, by using the diffraction correction of the ultrasonic compressional and transverse velocities, ρ = 4.43 g/cm 3 and σ = 0.35. The Youngs modulus of the bimodal microstructure is larger than that of the Widmanstätten microstructure, as shown in .ig. 6a. The shear modulus of the bimodal microstructure is also larger than that of the Widmanstätten microstructure, as shown in .ig. 6b. In general, the ultrasonic velocities were found to be lower in the coarse grain medium [10]. In the Widmanstätten microstructure, α β platelets have different orientations, forming colonies (see .ig. 2a). Strongly preferential orientations of these coarse colonies increase the morphological anisotropy while equiaxed α grains (bimodal microstructure) decrease it. Due to the inverse relation between ultrasonic velocity and density, an increase in the ultrasonic velocity must be related to an increase in the Youngs modulus and shear modulus [1719].
.rom these results, it is evident that the ultrasonic velocity is very sensitive to the stages of the precipitation process in both the Ti6Al4V ELI microstructures in different aging conditions. This is attributed to the fact that the ultrasonic velocities depend on the elastic properties of the material, which are very sensitive to any phase transformation and degree of mor-phological anisotropy. The first stage of precipitation (i.e. incubation) influences the Youngs and shear moduli of the material and in turn the compressional and transverse ultrasonic velocities. The depletion of precipitate-forming elements from the matrix to the α phase induced by aging leads to an increase in the Youngs modulus and shear modulus of the two varying medical Ti6Al4V ELI alloy microstructures, i.e. bimodal and Widmanstätten, respectively.

CONCLUSIONS
The precipitation behavior in the two varying medical Ti6Al4V ELI alloy microstructures (Widmanstätten and bimodal) has been studied using the microscopic and ultrasonic techniques. The study clearly reveals that the change of the ultrasonic compressional and transverse velocities arises mainly from the contrast in the elastic properties of the α and β phases which are affected by the degree of lattice distortion and misorientation. The ultrasonic compressional and transverse velocities are more sensitive to the early stage of precipitation whereas the effect of precipitation on hardness can be perceived only after the precipitates earn a minimum size to influence the move-.ig. 6. Influence of artificial aging in a medical TiAl4V ELI alloy on the Youngs (a) and shear (b) modulus with different microstructures (color online). ment of dislocations. These observations are consistent with the electron microscopy studies. The ultrasonic diffraction correction velocity method can be used to calculate the Youngs and shear moduli in a medical Ti6Al4V ELI alloy. The ultrasonic velocity parameter is very sensitive to microstructural changes; it can be used reliably for the monitoring of precipitates in aging processes of metallic alloys. .inally, the proposed nondestructive ultrasonic technique demonstrates an important assessment in the biomedical field and its applications. Moreover, it could be extrapolated to the phase transformation in more complex metallic and nonmetallic alloys. However, the identification of small microstructural changes with the strong effect on the mechanical properties requires new inspection methods, which should consider additional physical variables.

.UNDING
This work was performed at the Universidad Michoacana de San Nicolas de Hidalgo (UMSNH, Mexico) and partially funded by the Consejo Nacional de Ciencia y Tecnologia (CONACYT, Mexico) under Projects CB-2015/256013 and Prodep/CA-140.