Additively manufactured Ti-6Al-4V/Ta bimetal: from neutron diffraction residual stress analysis to wear and bioproperty testings

Raw materials

The powder feedstocks used in this study were pre-alloyed, gas-atomized Ti-6Al-4V powder and gas-atomized Ta powder [Table 2]. The particle size distributions were tested using a Malvern LS-609 laser particle size analyzer, with Dv (50) values of 35.2 μm for the Ti-6Al-4V powder and 27.9 μm for the Ta powder. scanning electron microscope (SEM) images and particle size distribution results are shown in Figure 1A-D.

Figure 1. Powder morphology and particle size: (A) Morphological characteristics of Ti-6Al-4V powder; (B) Ti-6Al-4V particle size distribution map; (C) Morphological characteristics of Ta powder; and (D) Ta particle size distribution map.

PBF-LB process and heat treatment

The metal printer used was an SLM Solution ® 125HL. The bimetallic printing process was conducted as follows: (a) Ti-6Al-4V was printed first using fixed parameters, i.e., laser power of 275 W, scanning speed of 1,100 mm/s, hatch distance of 0.12 mm, and layer thickness of 0.03 mm; (b) Printing was then paused to replace the Ti-6Al-4V powder with Ta powder; (c) Approximately 1 mm of Ta was subsequently printed on top of the already printed Ti-6Al-4V. For Ta printing, the layer thickness (0.03 mm) and hatch distance (0.08 mm) were kept constant, while the other parameters were varied for optimization [Table 3]. The corresponding energy density was calculated using the following Equation 1^[29]:

where P represents laser power, v for laser scanning speed, h for hatch distance, and s for layer thickness.

To release residual stress and study the effect of heat treatment on the as-printed samples, three different heat treatments were applied: 550 °C, 750 °C, and 950 °C, each for 2h. The samples were heated at a rate of 10 °C/min and subsequently cooled within the furnace. The rationale for choosing these temperatures is as follows: For pure Ta, its stress relief temperature is normally around 900 °C, while the recrystallization temperature is between 980 ~ 1,200 °C^[30]. For the Ti-6Al-4V alloy, the β-transus temperature is ~ 980 °C, the α + β phase range is 600-980 °C, and the martensite phase transformation starts at ~ 610 °C^[31].

Microstructural characterization and mechanical properties testing

The dimensions of the tensile samples are shown in Figure 2A. A schematic of the bimetal is shown in Figure 2B, where a Ta layer (~ 1 mm thick) is printed on the top of a Ti-6Al-4V substrate (~ 2 mm thick). Room temperature tensile tests were performed using a universal testing machine (Instron3382) at a strain rate of 0.2 mm/min. Three samples were tested for each group, and the results were averaged.

Figure 2. (A) Sample dimensions for the tensile test; and (B) setup for the nanohardness and modulus tests.

Microhardness was measured using an FM-ARS9000 Vickers microhardness tester. A total of 26 points were selected on each sample for testing, with a spacing of 0.1 mm between points [Figure 2B]. The test load was 0.98 N (equivalent to 100 gf), with a holding time of 15 s. The average value was calculated from three repeated measurements. Nanohardness testing was conducted using a TI950 nanoindenter (Hysitron, USA) with a load of 100 mN. The loading and unloading times were 5 s each, with a holding time of 2 s. Three experimental results were averaged. Young’s modulus was also measured using the same equipment. The specific area used for the modulus test is indicated in Figure 2B.

Friction and wear tests were conducted at room temperature using a ball-on-disc dry friction and wear method. The friction pair consisted of a Ф5.556 mm Si₃N₄ ceramic ball rotating at a speed of 200 r/min. The applied load was 6 N, and friction time was 15 min. The average value was used after repeating the test three times. For the study of the friction marks, a laser confocal microscope (LEXT OLS4000, Olympus, Japan) was used. Each mark was measured at three different positions, and the average value was used.

An optical microscope (OM, model LEICADM 2700M) was used to image porosity, and the result was subsequently analyzed using Image J software. A SEM (Zeiss Merlin, Germany) was used for microstructural analysis. An energy dispersive spectrometer (EDS) was used for compositional analysis, using a voltage of 20 kV and a current of 3000 pA. For XRD (MiniFlex600) analysis, the target material was Cu target Kα, with a voltage of 40 kV, a current of 15 mA, a step size of 0.02, and a scanning speed of 9°/min.

To construct the binary phase diagram of Ti-Ta, the ThermoCalc platform was used with the TCTI3:TT Ti-Alloys v3.1 database.

Neutron diffraction analysis of the residual stress

Neutron diffraction measurements were performed on the general purpose powder diffractometer, BL18 beamline, in the China spallation neutron source. For the residual stress analysis, the gauge volume is controlled by the four-blade slits, neutron Kirkpatrick-Baez mirror, and radial collimators, and the general gauge volume is φ2 × 2 mm³. A decoupled poisoned hydrogen moderator generates neutron pulses with a frequency of 25 Hz. 90 ± 20.5° ⁶LiF/ZnS scintillation detectors (4 mm × 4 mm) were used for data collection. It took about 2 h to measure the stress of each point. More details about the beamline can be found in published references^[32,33]. Two types of samples were analyzed in the neutron diffraction analysis: One was the as-printed Ti-6Al-4V/Ta bimetal, with dimensions of 10 mm × 9.5 mm × 5 mm; the other was the pure printed Ta, with dimensions of 10 mm × 5 mm × 3.5 mm. The samples were directly cut from the building substrate without any other polish process.

Electrochemical corrosion test

Electrochemical corrosion tests were conducted using a CHI760E (Shanghai Chenhua) electrochemical workstation. The sample to be tested served as the working electrode, with platinum as the counter electrode and a saturated calomel electrode (SCE) as the reference. During the open circuit potential (OCP) test, the scanning rate was set to 1 mV/s, with a total scanning duration of 7,200 s. For polarization curve measurements, the potential was scanned from -0.5 V to 1.5 V (vs. OCP) at a scan rate of 1 mV/s. To avoid the influence of the Ti-6Al-4V matrix, the remaining surface of the bimetal sample was sealed with resin and tape, leaving only a 1 mm² area of the test surface exposed. The sample was placed in a constant-temperature water bath at 37 °C to simulate the temperature of simulated body fluid (SBF) temperature. The composition of the SBF is shown in Table 4.

Biocompatibility testing

Ф10 mm bimetal discs were polished with 2,000 # sand paper to remove the surface oxide layer, then subjected to ultrasonic cleaning for 10 min in anhydrous alcohol with purities of 75% and 100%, respectively. The samples were sterilized using high-temperature and high-pressure treatment. MG-63 cells (human osteosarcoma cells, ATCC) were seeded onto the samples at a density of 4 × 10⁴ cells/sample for 2 h, and then fixed at 4°C for 2 h using 2.5% glutaraldehyde (C₅H₈O₂). The sample surfaces were washed with phosphate buffer solution and dehydrated in a graded ethanol series (10%, 30%, 50%, 70%, 80%, 90%, and 100%), with each step lasting 10 min. The samples were then dried using a freeze-drying machine, followed by platinum coating for observation of cell adhesion.

MG-63 cells were seeded onto the samples in a 24-well tissue culture polystyrene plate at a density of 4 × 10⁴ cells/sample, with each well containing 500 μL of culture medium. Cell proliferation was assessed using a CCK-8 assay (Shanghai Yisheng Biotechnology Co., Ltd, China) after incubation for 24, 72, and 120 h. The absorbance (optical density) of each well was measured at 450 nm using a microplate reader and software of Microplate Manager® 6, Version 6.3 (iMark, BIO-RAD, China). All biomedical data were analyzed using one-way ANOVA and are expressed as the mean ± standard deviation from at least three independent experiments. Statistical significance was defined as P < 0.05. The data were analyzed using SPSS 17.0.

Comsol molten pool simulation

COMSOL Multiphysic6.0 commercial software was used to simulate temperature field characteristics within molten pools ^[34,35]. Key parameters for the simulation, such as specific heat capacity (C_p), thermal conductivity (K), density (ρ), etc., were obtained from the literature^[3,27,28] [Table 5]. Detailed model descriptions are summarized in the Supplementary Materials.

Overall research flowchart

Figure 3 summarizes all the important steps involved in this study.

Figure 3. Overall research flow chart of the study.

Printing parameter optimization for the Ti-6Al-4V/Ta bimetal

Figure 4A and B shows the printing results obtained for the bimetal samples manufactured using laser power of 150 W and 200 W, respectively, with varying scanning speeds. It should be noted that the parameter optimizations mentioned here apply specifically to the top Ta layer, while the printing parameters for the bottom Ti-6Al-4V layer remained unchanged throughout the process. Figure 4 and Table 3 suggest that when the energy input is insufficient for the bimetal, such as ~ 220 J/mm³, a large number of pores (> 10%) exist in the microstructure. Decreasing the scanning speed and increasing the energy density reduced the porosity to a low level of ~ 2%. The best results were obtained when the laser power was 200 W and the printing speed was 330 mm/s, achieving a relative density of ~ 99.5% and porosity of ~ 0.5%. A further higher energy density (e.g., > 260 J/mm³) should lead to a lowered density, as it is known that the Ti-6Al-4V alone has the best energy density of ~ 70 J/mm^3[1] and too high energy input tends to cause evaporation in the Ti-6Al-4V part.

Figure 4. (A) Porosity and energy density of the bimetal when printed using 150 W laser power; and (B) porosity and energy density of the bimetal when printed using 200 W laser. The error bars represent the standard deviation.

The relative density results were verified using cross-sectional OM images, shown in Figure 5. The figure reveals that the samples produced using a laser power of 200 W and scanning speed of 330 mm/s had the lowest number of defects, consistent with the density measured in Figure 4.

Figure 5. OM images for the bimetal printed using different parameters: (A) 150 W and 240 mm/s; (B) 150 W and 260 mm/s; (C) 150 W and 280 mm/s; (D) 200 W and 330 mm/s; (E) 200 W and 350 mm/s; and (F) 200 W and 370 mm/s. OM: Optical microscope.

The as-printed bimetal was further characterized using OM and low-magnification SEM, with the resultant images shown in Figure 6A and B, respectively. The results indicate a good surface finish and a pore-free bonding interface between the Ti-6Al-4V matrix and the Ta layer.

Figure 6. (A) Appearance of the as-printed bimetal; and (B) SEM image showing the interface between the Ti-6Al-4V layer and the Ta layer. SEM: Scanning electron microscope.

To understand the bimetal printing process, a molten pool simulation was performed using the optimized printing parameters described above. The results are presented in Figure 7, where Figure 7A and C shows the data for Ti-6Al-4V and Figure 7B and D displays the results for Ta. The highest temperature achieved by the laser beam was ~ 3,500 K, exceeding the melting points of both Ti (1,940 K) and Ta (3,269 K), thereby ensuring metallurgical bonding. Since Ti-6Al-4V was printed first, followed by Ta, heat conduction from the subsequently printed Ta layer was reduced, leading to a slightly prolonged molten pool duration at high temperatures.

Figure 7. Molten pool simulation results for (A) and (C) Ti-6Al-4V; and (B) and (D) Ta.

Microstructure and mechanical properties of as-printed Ti-6Al-4V/Ta bimetal

Microstructural analysis

The microstructure of the printed bimetal was investigated using SEM with energy dispersive X-ray (SEM-EDX) spectroscopy, with a particular focus on the interface. The results are shown in Figure 8. The diffusion zone between the two distinctive metals was about 800 μm, as shown in Figure 8A and B. In this zone, all the constituent elements, namely Ti, Al, and V from the printed Ti-6Al-4V and Ta from the printed Ta, can be detected, indicating significant mutual diffusion between the two metals. However, Figure 8C also reveals the presence of a small amount of residual Ta powder, which is likely attributable to the high melting point of Ta (3,269 K).

Figure 8. EDS spectra of the bimetallic interface binding zone: (A) scanned area; (B) line scan energy spectrum; and (C) surface scan energy spectrum. EDS: Energy dispersive spectrometer.

The SEM results were verified through XRD analysis [Figure 9]. It is evident that α'/α-Ti is the main phase for the printed Ti-6Al-4V, although some β phases (β-Ti) are also visible. Within the pure Ta side, a single bcc-structured phase (bcc-Ta) was observed, while all three phases were observed at the bimetal interface. The enlarged inset in Figure 9 further indicates that, compared to the pure Ta phase, the Ta detected in the interface area was slightly shifted toward a higher angle, indicating a reduced unit cell parameter possibly resulting from the partial replacement of Ta by other elements, particularly Ti. Another possible reason for the shifted peak is due to residual stress in the alloy.

Figure 9. XRD spectra of the Ti-6Al-4V, Ta, and Ti-6Al-4V/Ta. XRD: X-ray diffractometer.

Mechanical properties

Figure 10A shows the distribution of Vickers hardness across the entire sample. This Vickers hardness of the printed Ta was 247-254 HV0.1, while that of the matrix Ti-6Al-4V was 337-340 HV0.1, which are similar to previously reported results^[27-28,36]. The Vickers hardness of the bonding zone was within the range of 279-290 HV0.1, likely due to the inter-diffusion between Ti-6Al-4V and Ta. Figure 10B shows the tensile stress–strain curve obtained for the Ti-6Al-4V/Ta bimetallic sample, demonstrating a tensile strength of ~934 MPa and a fracture elongation of 4.17%. The theoretical tensile strength R_mj of the bimetal can be calculated using Equation 2^[37]:

Figure 10. Mechanical properties of the as-printed bimetal using the best printing parameters: (A) Hardness* (The error bars represent the standard deviation); and (B) Tensile property; (C) Overall fracture morphology of the as-printed bimetal; (D-F) Enlarged images showing the fracture morphology at Ta layer, interfacial bonding region, and Ti6Al4V matrix, respectively. *Error limit and average value are shown for Figure 10A, where error limit is expressed as line segemant and average value is shown as square spot.

(2) $$ R_{m j}=\frac{R_{m 1} \times t_{1}+R_{m 2} \times t_{2}}{t_{1}+t_{2}} \\ $$

where R_m1 and R_m2 are the lower-limit standard values of the tensile strength of the substrate and added layer, respectively, and T₁ and T₂ are the thicknesses of the substrate and added layer, respectively. Based on the known thickness of each metal and the international standard for single Ti-6Al-4V and single Ta, the theoretical tensile strength of the Ti-6Al-4V/Ta bimetallic structure was calculated to be approximately 686 MPa^[38,39]. The difference between the experimental and theoretical values may be caused by the fine microstructure and residual stress induced by the high cooling rate of the PBF-LB process. Thus, the higher strength of the as-printed Ti-6Al-4V/Ta bimetal (934 MPa) obtained here demonstrates the advantages of PBF-LB in the development of high-strength biomedical materials. The fracture surface was also characterized by SEM to determine the fracture mechanism. As shown in Figure 10C, the overall fracture surface shows distinct morphologies from the Ta layer to the Ti-6Al-4V matrix. The Ta layer shows a dissociation fracture model [Figure 10D], while the interfacial and Ti-6Al-4V matrix present ductile dimples [Figure 10E and F]. The compressive stress within the Ta/Ti6Al4V interface region may contribute to the bonding strength and thus improve the ductility.

Residual stress by neutron diffraction of as-printed Ti-6Al-4V/Ta bimetal

Neutron diffraction was used to evaluate the residual stress of the as-printed bimetal^[32]. The pure Ta was analyzed first, followed by the as-printed bimetal.

Figure 11 shows the results for the Ta part, with Figure 11A showing the analysis points (Point 1 and Point 2) inside the as-printed Ta. In the middle of the sample, tensile residual stresses of 140-180 MPa were present in the X-, Y-, and Z- directions. Figure 11B shows the original experimental data, based on which residual stress has been determined in each direction^[33,34] [Table 6].

Figure 11. (A) Sample dimensions for the neutron diffraction analysis, with measurement locations indicated; and (B) Residual stress distributions in the X, Y, and Z directions.

A more comprehensive neutron diffraction analysis was conducted on the as-printed bimetal. Figure 12 shows the results for the analysis points inside the Ta alone (Point 1), in the interface bonding area (Point 2), and inside the Ti-6Al-4V (Point 3, which is closer to the interface, and Point 4, which is away from the interface) [Table 6]. The results show a few notable findings:

Figure 12. (A) Sample dimensions for the neutron diffraction analysis, with measurement locations indicated, and the corresponding results table; and (B) Residual stresses in the X, Y and Z directions.

(1) The residual stress measured in the Ta part of the bimetal (Point 1) was similar to or slightly higher than that of the as-printed pure Ta.

(2) A compressive residual stress of ~ (30-70) MPa was recorded near the interface of the as-printed bimetal (Point 2), differing greatly from the higher tensile stress (~ 100 MPa) observed in the pure Ta (Point 1 and Figure 12).

(3) A high residual tensile stress (~ 350-390 MPa) was detected around Point 3, that is, inside the as-printed Ti-6Al-4V.

Microstructure and properties of heat-treated Ti-6Al-4V/Ta bimetal

Microstructure and interface characterization of heat-treated Ti-6Al-4V/Ta bimetal

As demonstrated by the results in the above section, there is a high level of residual stress in the as-printed bimetal, which is also nonhomogeneous and likely to induce premature fracture. As for most other as-printed materials, this can be rectified by heat treatment. The bimetal samples were subjected to various heat treatment regimens. Figure 13A shows the XRD patterns of the Ta side of the Ti-6Al-4V/Ta bimetal in the as-printed state (T0) and after heat treatment at 550 °C (T1), 750 °C (T2), and 950 °C (T3). Diffraction peaks corresponding to bcc-Ta were detected in all samples. When the heat treatment temperature increased from 550 °C to 950 °C, the diffraction peak of bcc-Ta shifted significantly to the right, most likely due to a reduction in residual stress. In addition, with an increase in the heat treatment temperature, the diffraction peak width of the bcc-Ta phase increased slightly, indicating an enlarged grain size after heat treatment. Figure 13B shows the XRD patterns of the area around the Ti-6Al-4V/Ta interface. The (0002)_α-Ti diffraction peak shifted left as the heat treatment temperature increased, suggesting heat treatment may have caused Ta atoms to diffuse into the Ti matrix.

Figure 13. XRD patterns of bimetallic materials under different heat treatments: (A) the Ta side of the Ti-6Al-4V/Ta bimetal; and (B) The Ti-6Al-4V/Ta interface area. XRD: X-ray diffractometer.

Nanohardness and tensile properties of heat-treated Ti-6Al-4V/Ta bimetal

The nanoindentation hardness of the interface area of the Ti-6Al-4V/Ta bimetals in the as-printed and heat-treated states was evaluated, and the results are shown in Figure 14. The nanohardness of the Ta decreased from ~ 3.4 GPa in the as-printed state to ~ 2.7 GPa after heat treatment. The nanohardness of the Ti-6Al-4V matrix also decreased. The 550 °C heat treatment case differed from the others, exhibiting a high nanohardness of up to ~ 4.4 GPa. This is likely attributed to the residual stress detected in the local area.

Figure 14. Nanohardness mapping results for the bimetal under different heat treatments: (A) As-printed; (B) 550 °C; (C) 750 °C; and (D) 950 °C.

Figure 15 shows the stress-strain curves of the as-printed and heat-treated bimetals. Heat treatment at 550 °C resulted in the highest tensile strength (~ 1,044 MPa) but the lowest elongation (2.82%). Higher heat treatment temperatures reduced the tensile strength and increased the elongation at break. At temperatures of 750 °C and 950 °C, the corresponding strengths are 908 MPa and 860 MPa, with elongations of 5.24% and 6.81%, respectively. As shown in Table 7^[40], the performance of the heat-treated Ti-6Al-4V/Ta bimetal was close to that of the annealed Ti-6Al-4V.

Figure 15. Room temperature tensile curves of the bimetal under different heat treatments.

Table 7

Results of bimetallic room temperature tensile test^[40]

	T0	T1	T2	T3	SLM Ti-6Al-4V[40]
Status	As-printed	550°C	750°C	950°C	600°C	800°C	900°C
Tensile strength /MPa	938	1044	908	860	959	802	689.5
Elongation/%	4.17	2.82	5.24	6.81	8.3	6.8	6.5

Wear resistance of heat-treated Ti-6Al-4V/Ta bimetal

Room-temperature friction and wear tests were conducted on the as-printed and heat-treated bimetals, and the results were compared with those for as-printed Ti-6Al-4V. Figure 16A shows the wear track depth and width curves of the different samples. The average wear rate was calculated using the wear volume of the sample and the reciprocating stroke during the wear test, as shown in Equation 3^[41].

Figure 16. Wear properties of the bimetal under different heat treatments: (A) wear profile and (B) average wear rate. The error bars represent the standard deviation.

(3) $$ K_{v}=\frac{V_{\text {uear }}}{S F_{N}} \\ $$

where K_v represents the wear rate (mm³/Nm); V_wear is the wear volume (mm³); S is the travel distance (m); and F_N is the applied normal load (N).

Figure 16B compares the wear rates of the as-printed and heat-treated samples with each other and with that of the base Ti64 alloy. The wear rate of the sample heat-treated at 550 °C was virtually identical to that of the as-printed bimetal, measured at 4.02 × 10^-5 mm³/Nm and 3.94 × 10^-5 mm³/Nm, respectively. As the heat treatment temperature increased, the average wear rate rose to 6.39 × 10^-5 mm³/Nm and 9.43 × 10^-5 mm³/Nm for samples treated at 750 °C and 950 °C, respectively. Compared to the as-printed Ti-6Al-4V alloy, all bimetal samples exhibited significantly improved wear resistance.

Corrosion resistance of the heat-treated Ti-6Al-4V/Ta bimetal

Figure 17A and B shows the electrochemical corrosion behavior of the different bimetals, as well as that of the as-printed Ti-6Al-4V, measured using the controlled potential method in an SBF solution at 37 °C. The corrosion rate can be indirectly assessed using the self-corrosion current density, which follows Faraday's law, as shown in Equation 4^[42,43]:

Figure 17. (A) Tafel curve of the potentiodynamic polarization curves of the bimetallic in the SBF solution and under different heat treatments; and (B) corrosion voltage and corrosion current. SBF: Simulated body fluid.

(4) $$ S_{\text {corr }}=\frac{I_{\text {corr }} \times A}{F \times n \times \rho} \\ $$

where S_corr is the metal corrosion rate, I_corr is the corrosion current density, A is the relative atomic mass of the metal, F is the Faraday constant, n is the number of electrons lost from the metal, and ρ is the density of the metal.

The self-corrosion current density was determined by extrapolating the Tafel region of the polarization curve. When using Tafel fitting in Zview, key parameters such as the corrosion potential (E_corr), anodic and cathodic Tafel slopes (β_a/β_c), and initial corrosion current (I_corr) should be selected. The precise values were obtained through iterative convergence using the least squares method. Table 8 lists the self-corrosion current densities and voltages after fitting and converting the Tafel curves. Compared to the as-printed Ti-6Al-4V, the as-printed bimetal (T0) already exhibited superior corrosion resistance. Specifically, its E_corr increased from -0.34 V to -0.31 V, while the I_corr significantly decreased from 2.29 × 10^-6 A/cm² to 4.57 × 10^-7 A/cm². Correspondingly, the corrosion rate dropped markedly from 2.02 × 10^-3 mm/y to 3.24 × 10^-4 mm/y, indicating that Ta inherently possesses better electrochemical stability in corrosive environments.

Table 8

Polarization results of the different samples

	T0	T1	T2	T3	As-printed Ti-6Al-4V
Ecorr(V)	-0.31	-0.19	-0.17	-0.29	-0.34
Icorr (A/cm2)	4.57 × 10-7	3.81 × 10-7	2.88 × 10-7	4.27 × 10-7	2.29 × 10-6
Scorr(mm/y)	3.24 × 10-4	2.71 × 10-4	2.04 × 10-4	3.03 × 10-4	2.02 × 10-3

In general, heat treatment improved the corrosion resistance of the bimetallic samples, resulting in a decrease in corrosion current. With increasing heat treatment temperatures, the corrosion resistance of Ta improved progressively in the T1 (550 °C) and T2 (750 °C) samples. Among them, T2 exhibited the best performance, with the highest E_corr (-0.17 V), the lowest I_corr (2.88 × 10^-7 A/cm²), and the lowest corrosion rate (2.04 × 10^-4 mm/y), reflecting optimal electrochemical stability. In contrast, at an even higher temperature (T3, 950 °C), the I_corr of the T3 sample slightly increased, with a corresponding corrosion rate of 3.03 × 10^-4 mm/y. Although marginally inferior to T2, its performance still surpassed that of the untreated sample. A possible reason is that excessively high heat treatment temperature may cause inward diffusion of O atoms from the surface oxide layer, leaving corrosion channels in their wake and thereby slightly reducing the sample’s resistance to corrosion.

In vitro biocompatibility of the heat-treated Ti-6Al-4V/Ta bimetal

The initial adhesion of MG-63 cells on the various printed bimetal samples over 2 h was assessed using SEM, as shown in Figure 18A-D. Most MG-63 cells, except for the printed Ti-6Al-4V/Ta, exhibited a flat spherical/prismatic shape with a diameter of approximately 30 μm. More notably, the MG-63 cells had many filamentous filopodia extensions on the surfaces of the bimetallic samples, whereas fewer filamentous filopodia extensions were observed on the Ti64 sample. The cell dispersion was similar for all samples, but cell-to-cell communication and adhesion were more pronounced on the bimetallic sample and bimetallic sample treated at 950 °C. It is well established that cell adhesion to material surfaces is regulated by both surface chemistry and topology^[44].

Figure 18. SEM images showing cell adhesion after 2 h of culture on bimetal and as-printed Ti-6Al-4V under different heat treatments: (A) as-printed Ti-6Al-4V; (B) As-printed Ti-6Al-4V/Ta; (C) Heat-treated at 550 °C; and (D) heat-treated at 950 °C; (E) Proliferation of MG-63 cells on the bimetals under different heat treatments after 24 h, 72 h, and 120 h of culture (*P < 0.05) (For each condition, five samples were studied). The error bars represent the standard deviation. SEM: Scanning electron microscope.

The in vitro cytocompatibility of all samples was investigated by measuring the cellular activity after culturing the MG-63 cells for 24 h, 72 h, and 120 h. The results are shown in Figure 18E. In comparison with the Ti-6Al-4V sample, the bimetallic samples showed higher cell proliferation rates at 72 h and 120 h, indicating that they significantly promote cell proliferation and have excellent biocompatibility. However, the biocompatibility of the bimetal samples decreased after heat treatment, likely due to changes in the physical and chemical properties of the surface oxide layer, which may have reduced their ability to interact with cells. Nevertheless, these results suggest that bimetals exhibit superior biocompatibility compared to Ti-6Al-4V.

Discussion

The above experiments indicate that when printed using optimized parameters, the Ti-6Al-4V/Ta bimetal achieves a series of favorable properties, including high fracture strength, superior wear and corrosion resistance, and improved biocompatibility. Compared with pure Ti-6Al-4V, the Ta layer plays a crucial role in improving wear and corrosion resistance, as well as biocompatibility. On one hand, the improved wear resistance by the Ta layer can be explained by the formation of tantalum oxides during localized short-term heating caused by frictional interaction with the counter body. These oxides reduce friction and thereby contribute to improved wear resistance^[45]. On the other hand, the superior corrosion resistance is also related to tantalum oxide, which is more chemically stable than titanium oxide. The tantalum oxide layer can prevent Cl⁻ ions in the SBF solution from penetrating the passive film and initiating pitting corrosion^[46]. Moreover, the improved biocompatibility can be attributed to the intrinsic excellent biocompatibility of element Ta. The Ta layer also acts as a barrier, preventing the diffusion of cytotoxic Al and V ions into surrounding tissues or tissue fluids. These improved properties are closely related to the unique structural characteristics of the Ti-6Al-4V/Ta bimetal, which are strongly influenced by residual stress and interfacial bonding.

Residual stress

Generally speaking, residual stress due to thermal gradient is proportional to (ΔT·α·E), where ΔT is the temperature gradient, α is the coefficient of thermal expansion, and E is the Young’s modulus^[47]. In the above section, a change from tensile to compressive residual stress was observed in the as-printed Ta alone and the interface area in the as-printed bimetal. Based thereon, the following analysis is made:

(1) The absolute value of the residual stress in the interface area region of the bimetal was reduced compared to the plain Ta. This is attributed to two reasons: the lowered temperature gradient and a lower Young’s modulus. Regarding the former, the Ta part of the bimetal was printed on top of a freshly printed Ti-6Al-4V substrate. The newly deposited alloy is likely to have been at a higher temperature than the substrate of a normal printing condition (often approximately 200 °C). Regarding the latter, the inter-diffusion of Ta atoms and Ti/Al/V atoms during printing is likely to have reduced the Young’s modulus of the Ta part of the bimetal because of the lower Young’s modulus of Ti-6Al-4V.

(2) The change from tensile to compressive residual stress is most likely caused by changes in the lattice parameters, which affect plane distance and, eventually, the residual stress status. This is also possible to have been exacerbated by the fact that Ta has an atomic radius of ~ 217 ppm and Ti has an atomic radius of ~ 187 ppm^[48]. The diffusion of Ti atoms into the Ta lattices and the replacement of Ta atoms decreased the lattice parameters, thereby changing the residual stress from tensile to compressive.

Interface bonding in bimetal

As mentioned before (e.g., Figure 7), a COMSOL simulation was conducted to determine the optimal printing parameters for the Ti-6Al-4V/Ta bimetal. The resulting parameters were predicted to achieve a molten pool temperature of up to ~ 3,500 K, which, from an energy perspective, is high enough to remelt the Ta powder and create the Ti-6Al-4V/Ta bimetal. Theoretical analysis of the depth of the molten pool can also assist in understanding the bimetal formation. This was done using the model developed by Tang et al.^[49]:

(5) $$ D=\frac{W}{2}=\sqrt{\frac{2 A P}{\operatorname{en} \rho C\left(T_{m}-T_{0}\right) v}} $$

where W and D are the melt pool width and depth, respectively; A is the laser absorptivity (assumed to be 0.374 for the Ta powder bed); ρ is the material density; C is the material’s thermal capacity; v is scanning speed; T_m is the material’s melting temperature; and T₀ is the substrate temperature.

Under the optimized parameters (i.e., laser power of 200 W and scanning speed of 0.33 m/s), the corresponding D and W values were calculated as: D = 87.4 μm and W = 174.8 μm.

The following observations were made from the molten pool simulation results.

(1) Since the layer thickness used in the study is only 30 μm, the ~ 80 μm molten pool depth ensures that the very few first layers of Ta that have been deposited on top of the Ti-6Al-4V matrix have a good chance to inter-diffuse and bond with the high-temperature Ti phase (the β-Ti phase).

(2) In the printing experiments, the width of the interface area was recorded to be approximately 800 μm (see the microstructural analysis in Figure 8). This was much higher than the simulated molten pool depth, implying that a full layer of Ta was created only after approximately 26-layer powder depositions. This is understandable, considering that the vacancy of the powder bed can normally be ~ 30 vol.%, leaving a large amount of open space to be filled by the next layer of material. Furthermore, Ta has a much higher density than Ti, which also contributes to the downward movement of Ta atoms and the enlargement of the bonding area.

(3) Ti and Ta have infinite solubility in one another [Figure 19], which is often beneficial for forming good metallurgical bonds without brittle intermetallics.

Figure 19. Binary phase diagram of Ti-Ta.

(4) Compressive residual stress defected in the interface area of the as-printed bimetal may have also contributed to a good bonding [Figure 13].