Mechanical, forming and biological properties of Ti-Fe-Zr-Y alloys prepared by 3D printing

1. Introduction

3D printing (3DP) is a process of laser-based additive manufacturing to make a metal component directly from a virtual 3-D CAD model [[1], [2], [3], [4]]. It can significantly reduce the “concept to product” lead time by eliminating several intermediate steps, and provides a new approach for fabricating the customized medical implants [[5], [6], [7], [8], [9], [10], [11]]. Thus extensive researches have been carried out to investigate 3DP medical implants [[12], [13], [14], [15]]. Till now, most of these related studies focused on the conventional Ti-based alloys (such as Ti-6Al-4 V and Ti-6Al-7Nb) [11]. However, the application of these alloys will be limited in the future, because they contain toxic vanadium and aluminum. What’s more, their Young's moduli are much higher than that of natural bone [[16], [17], [18]]. New β-type titanium alloys exhibit excellent biocompatibility and low Young's modulus [19]. Since these alloys undergo wider temperature range of solidification during 3DP process, it is easy to exert adverse impacts on forming precision and quality of products. In addition, strength and wear resistance are insufficient [20,21], because the substitutional solid solution strengthening is limited.

3DP biomaterials not only have high mechanical properties and good biocompatibility, but also possess excellent formability. It would be worth to develop high strength and ductile eutectic microstructure, since such alloys have single melting temperature. With this respect, the binary Ti-Fe eutectic alloy having a novel combination of high mechanical properties and good formability is potential candidacy for a 3DP biomaterial [[22], [23], [24]]. But it still has limitations. One is that the alloy is easily oxidized forming harmful Ti₄Fe₂O phase. The other is that the Young's modulus of the alloy (149-154 GPa) is higher [25]. To develop it into a 3DP biomaterial, the most important thing is to improve oxygen removal ability and to decrease Young's modulus. For this purpose, Y and Zr elements were added to the binary Ti-Fe eutectic alloy. Based on a “cluster-plus-glue-atom” model, a dual cluster formula for the Ti-Fe-Zr-Y alloys was proposed as [Ti_24-xFe₁₀Zr_x]_0.7Y_0.3. When the alloys conforming to the composition formula were produced by 3DP, there were β-Ti, TiFe and Zr₂Fe phases in the alloys. The volume fraction of the β-Ti increases and that of the TiFe decreases with increasing Zr addition, while that of the Zr₂Fe phase first increases and then decreases, with the highest fraction obtained at 5.86 at.% Zr addition. It is worth noting that no Ti₄Fe₂O oxide is found in the studied alloys due to purifying effect of Y element. There was a sequential structure change from hypereutectic to eutectic to hypoeutectic with the increasing Zr addition. When the addition of Zr less than 5.86 at.%, there are two structure constituents in the alloys: primary TiFe dendrites and (β-Ti + TiFe + Zr₂Fe) eutectic, of which the primary dendrites become less in amount, shorter in length and thinner in width with the increase of Zr addition. When the addition of Zr reaches 5.86 at.%, a fully (β-Ti + TiFe + Zr₂Fe) eutectic structure is obtained. Above such an additive amount, the microstructure develops into hypoeutectic structure consisting of primary β-Ti dendrites plus (β-Ti + TiFe + Zr₂Fe) eutectic, showing obvious increase in number and length of primary β-Ti dendrites with even more Zr addition. The optimized Ti-Fe-Zr-Y eutectic alloy not only exhibits a high hardness and a low Young's modulus resulting from grain refinement and dissolution of Zr in the β-Ti solid solution, but also has good corrosion resistance in Hank’s solution due to “enveloping effect” of a dispersed eutectic structure against corrosion [26]. Continuing our investigations of the Ti-Fe-Zr-Y system, the influences of Zr additions on the mechanical, forming, and biological properties of the alloys were analyzed in the present work.

2. Experimental procedure

2.1. Materials

Pure titanium rolled plate (Baoji Titanium Industry Co. Ltd., 99.70% purity) with dimension of 20 mm × 20 mm × 10 mm was chosen as substrate. Master alloys with nominal compositions of Ti_68.91Fe_29.32Zr_1.47Y_0.30, Ti_67.45Fe_29.32Zr_2.93Y_0.30, Ti_64.52Fe_29.32Zr_5.86Y_0.30, Ti_61.58Fe_29.32Zr_8.79Y_0.30, Ti_58.65Fe_29.32Zr_11.72Y_0.30 and Ti_70.5Fe_29.5 (atomic percentage) were prepared by arc-melting the mixtures of 99.99% Ti, 99.90% Fe, 99.99% Zr and 99.90% Y (wt%) in an argon atmosphere, and then were ground into spherical powder with size of 30-100 μm by ball grinder (DECO-PBM-V-2 l), which were chosen as 3DP materials. The Ti-6Al-4 V alloy used in the experiment was also provided by the above company, the impurity content of the alloy is lower than 1.045 wt%.

2.2. Laser 3D printing

A 6 kW continuous-wave CO₂ laser unit (DL-HL-T5000B, China) was used for producing deposited layers in the argon protection environment. Laser manufacturing parameters play crucial role in controlling the quality and the microstructure of deposited layers. Our previous research showed that when the energy density of laser was less than 93.7 J/mm³, the deposited layers were not dense owing to less energy per unite area to melt powder. With the increasing of the energy density, the relative density of the deposited layers presented the trend of first increasing and then decreasing, with the highest relative density obtained at 125 J/mm³. Therefore the energy density of 125 J/mm³ was appropriate to produce the deposited layer with high relative density and free defects. What’s more, the same combined laser process parameters could still produce different microstructure if individual parameters were different. Based on orthogonal tests, the optimized laser processing parameters were adopted as follows: laser power 2.6 kW, laser beam diameter 6 mm, scanning velocity 9 mm/s, preset layer thickness 1 mm, overlapping 30%, argon flow rate 7.0 L/min. Cylindrical specimens of about 5.0 mm in diameter and 10.0 mm in height were prepared using scanning strategies of cross-hatching.

2.3. Testings of mechanical and physical properties

Reciprocating friction-wear test was performed using a CETR UMT-2 testing machine. A Si₃N₄ ball with a diameter of 5.96 mm and a hardness of 1500 HV was selected as the wear couple. The experiment was performed at a normal load of 5 N, a sliding speed of 2.0 mm/s, and a wear time of 30 min. Worn surface morphologies were observed using a JEOL-5600LV scanning electron microscopy (SEM), and composition was analyzed using a EPMA-1720 electron probe microanalyzer (EPMA) equipped with four wavelength dispersive spectrometers and tungsten filament. The used analyzing crystals were LiF and PET (pentaerythritol).

The compressive property was measured with an Instron-type testing machine at a strain rate of 0.036 mm/min. The specimens for compressive testing were 6 mm long rectangular parallelepiped with 3 mm × 3 mm cross-section. The surface roughness was tested using NV5022 surface profiler. The density was measured with Mettler Toledo XS64 densimeter.

2.4. Apatite-forming ability

For evaluation of apatite growth on surface of the alloys, the cylindrical specimens (10 mm in diameter and 2 mm in height) were soaked for 4 d, in 10 ml of a simulated body fluid (SBF) solution at 37 °C, which contained ion concentrations nearly equal to those of human blood plasma with respect to Na⁺, K⁺, Ca²⁺, Mg²⁺, Cl^-, HCO₃^-, HPO₄^2-, and SO₄^2- concentrations. After a predetermined soaking time of 4 d, the specimens were removed from the SBF solution, rinsed with deionized water, and then dried at 40 °C. Surface morphologies and composition of sediments were analyzed by SEM and EPMA.

2.5. In vitro experiments: cytotoxicity testing of the alloys

2.5.1. Preparation of extracts from alloys for cytotoxicity testing

Firstly, the samples were sterilized in an autoclave, which was sealed and heated at 120 °C for 3 h. Secondly, the samples were respectively placed in several 35 mm diameter petri dishes to extract samples for cytotoxicity testing, then 3 ml DMEM culture medium (hereafter referred to as serum-free medium, SFM) was added to each petri dish. At the end of the extraction period, the aqueous extract was filtered by a disposable needle filter.

2.5.2. Cell culture and relative proliferation rate

Mouse fibroblasts (L-929 cells) were used to evaluate the cytotoxicity of the alloys. On day 1, l-929 cells were seeded in a 96-well culture plate at a concentration of 5 × 10³ cells/ml in DMEM containing 10% FBS (fetal bovine serum) with 100 μl of cell suspension added to each well and then the cells were cultured at 37 °C in a saturated humidity atmosphere of 5% CO₂ for 24 h. On day 2, the original culture medium in each well was respectively replaced by 100 μl dilution with concentration of 100%, 50%, 10% and 1% that diluted by DMEM (dulbecco's modified eagles medium). The SFM medium was as a negative control and the SFM medium with 0.64% phenol was as a positive control. All plates were incubated at 37 °C for 3 d. On day 4, 20 μl of thiazolyl blue tetrazolium blue reagent (MTT, T0793-5 G, Amresco) at a concentration of 5 mg/ml was added to each well and incubated for 4 h. Then the morphology of cells was checked using an inverted microscope (DMI 1, Leica). After the morphology observation, 150 μl DMSO was added to per well. The optical density (OD) was read spectrophotometrically at 490 nm by molecular devices (SpectraMax 190). The cell relative growth rate (RGR) is the ratio between the OD of experimental group and negative control. The RGR was converted into 0-5 material toxicity levels according to GB/T 16,886.5 standard. 0 and 1 levels are nontoxic, while 2-5 levels are toxic.

2.6. Cell adhesion and proliferation

Firstly, the samples were placed in a 24-well culture plate and immersed in 1 ml DMEM with 10% FBS. After that, the l-929 cells were seeded at a density of 1 × 10⁵ cells/hole, and a cell culture plate without sample was also included as a control. The cells were cultured at 37 °C in a humidified atmosphere with 5% CO₂ for 30, 60 and 120 min. Secondly, after the cell culturing for 30, 60 and 120 min, the wells were rinsed twice with PBS. After that, 200 μl MTT (5 mg/ml) and 800 μl DMEM were added into the wells and the cells were incubated at 37 °C for 4 h. Thirdly, after 4 h, the supernatant was abandoned, 1 ml DMSO was added to the wells and shook for 10 min to dissolve the formazan product. Then the OD of the solving liquid was read spectrophotometrically at 490 nm by molecular devices.

Except the cell inoculation density was 2 × 10⁴ cells/hole, the culture time of l-929 cells was 1, 4, and 8 d, the other process of cell proliferation was the same as that of cell adhesion.

3. Results and discussion

3.1. Tribological properties

Fig. 1 shows changing curves of the friction coefficient and the worn volume of the alloys with the amount of Zr addition. It can be found that the friction coefficient and the worn volume gradually decrease with increasing Zr addition to 5.86 at.%, and they increase afterwards. This suggests that the alloy with 5.86 at.% Zr addition has the best anti-friction property and the highest wear resistance among the studied alloys. Moreover, the studied alloys exhibit enhanced tribological properties compared with the binary Ti_70.5Fe_29.5 eutectic alloy (its friction coefficient and the worn volume were measured to be 0.881 and 0.0782 mm³, respectively) and Ti-6Al-4 V alloy (its friction coefficient and the worn volume were measured to be 0.665 and 0.0757 mm³, respectively).

In order to investigate the wear mechanism, the worn surface morphologies of the alloys are observed by SEM. As shown in Fig. 2(a), some adhesive substance (indicated by arrows) appears on the worn surface of the hypereutectic alloy with 1.47 at.% Zr addition besides plowing grooves induced by abrasive wear. EPMA analysis reveals that the adhesive substance is mostly composed of Ti and Fe, whose average composition is Ti_51.48Fe_48.36Zr_0.16 (atomic percentage), indicating that material transfer of TiFe phase occurs during the wear process. This is because the alloy has high fraction of TiFe (about 60 vol.%), while TiFe has high adhesive tendency with the Si₃N₄ counterpart [27], which easily causes junctions at the contacting asperities of the TiFe phase and the Si₃N₄ counterpart under a compressive load. As sliding continues, the junctions will be broken under shearing stress [28], producing material transfer and desquamating pits. At the same time, since some of hard debris joins the wear process as wear particles, abrasive wear with plowing groove feature occurs on the worn surface [29]. Similar worn surface morphology is also observed in the hypereutectic alloy with 2.93 at.% Zr addition, and the adhesive substance has similar composition (Ti_52.78Fe_47.01Zr_0.21) to that of above alloy. But the transferred materials decrease in number, and the plowing grooves become broad and deep (Fig. 2(b)), being attributed to the reduced fraction (about 35 vol.%) of the TiFe phase and the decreased hardness of the alloy. The variation trend continues to the alloy with 5.86 at.% Zr addition due to the formation of fine eutectic structure (Fig. 2(c)), the composition of the adhesive substance is measured to be Ti_50.38Fe_49.29Zr_0.33 (atomic percentage). In the case of the hypoeutectic alloy with 8.79 at.% Zr addition, the material transfer phenomenon (indicated by arrows) is enhanced again (Fig. 2(d)). Unlike the hypereutectic alloys, the transferred materials having average composition of Ti_87.05Fe_12.48Zr_0.47 (atomic percentage) are rich in Ti as revealed by EPMA, indicating that the alloy undergoes a totally different material transfer process. A possible explanation for this is that accompanied by the increase of the tough β-Ti phase (about 70 vol.%), plastic deformation may occur in the subsurface of the alloy under the sustained contact stress [30,31], which easily leads to the sharp increase of dislocation density [32], inducing the nucleation and growth of crack [33]. As sliding continues, some of materials will be separated from the surface. The crack initiation plays a leading role during this stress fatigue wear process, the lower hardness tends to induce the initiation of crack. Therefore, material transfer becomes severer for the hypoeutectic alloy with 11.72 at.% Zr addition owing to even higher fraction of the β-Ti phase (about 80 vol.%). Meanwhile, the plowing grooves become broader and deeper (Fig. 2(e)), but the change in composition of adhesive substance is not significant, the average composition is Ti_85.32Fe_14.07Zr_0.61 (atomic percentage). As for the binary Ti_70.5Fe_29.5 eutectic alloy, besides the desquamating pits and the plowing grooves, a large crack (indicated by arrows) propagating along an angle of about 45° to the sliding direction was clearly observed on the worn surface (Fig. 2(f)). The formation of the crack is mostly attributed to the existence of the Ti₄Fe₂O oxide at eutectic colony boundaries, which acted as stress-concentrating site, leading to initiation and propagation of crack. This can be used to explain why the binary eutectic alloy has low tribological properties compared with the studied alloys. As for the Ti-6Al-4 V alloy, there are a lot of wide and deep plowing grooves and debris on the worn surface induced by severe abrasive wear (Fig. 2(g)). This makes the tribological properties of Ti-6Al-4 V alloy are lower than that of the studied alloys.

3.2. Compression performance

The stress-strain curves of the Ti-Fe-Zr-Y alloys under compressive test are shown in Fig. 3. The mechanical test data are summarized in Table 1. It is interesting to note that the hypereutectic alloys with the addition of Zr less than 5.86 at.% exhibit an excellent combination of mechanical strength and ductility. The high strength values result from the dissolution of Zr and Fe in the β-Ti phase (solution strengthening) and the formation of a hard primary TiFe phase [26]. The rounded dendritic morphology of TiFe compound is believed to be one of the factors improving ductility of the alloys, which acts as efficient barriers for shear strain and cracks propagation [[34], [35], [36]]. One should also mention that the hypereutectic alloy with 1.47 at.% Zr addition, having about two times higher volume fraction of the primary TiFe phase than that with 2.93 at.% Zr addition, exhibits the highest strength and the best ductility among the studied alloys, indicating that the increased fraction of primary TiFe phase is beneficial to improve the mechanical properties of the alloys in the experimental composition range. In the case of the eutectic alloy with 5.86 at.% Zr addition, there is decrease in strength and ductility compared to the hypereutectic alloys, owing to an absence of the hard primary TiFe dendrites and high fraction of the brittle Zr₂Fe. The decreasing trend in strength and ductility continues to the alloys with the addition of Zr more than 5.86 at.%, resulting from the formation of hypoeutectic structure having the soft carcass of β-Ti solid solution in the eutectic matrix. Here it is worth mentioning that the strength and ductility of the studied alloys are superior to that of binary Ti_70.5Fe_29.5 eutectic alloy as shown in Table 1. While because of there is no obvious compression fracture for Ti-6Al-4 V alloy, it is impossible to draw an accurate conclusion which alloy's compression property is better between the studied alloys and Ti-6Al-4 V alloy. But due to the compression strength of Ti-6Al-4 V alloy is similar to that of tensile strength, we can infer the strength of the studied alloys (1410-2156 MPa) are higher than that of Ti-6Al-4 V alloy (its tensile strength was measured to be 967 MPa).

The fractographic observation displays that the fracture of the hypereutectic alloys with 1.47 at.% and 2.93 at.% Zr additions occurs under both normal and shear stresses. Vector normal to the compressive fracture surface is inclined at different degrees with the load direction varying from 90° to 45°. At high magnification, the alloys exhibit irregular fracture surfaces that are similar to that of fatigue failure of Ti-based alloy (Fig. 4(a) and (b)). One should mention that crack deflection can be clearly observed in eutectic colony, leading to tortuous zigzag-shaped crack path. But the crack does not cross the eutectic TiFe phase as indicated by arrows in Fig. 4(a) and (b), which indicates that TiFe phase is an efficient barrier for crack propagation. The fracture of the eutectic alloy with 5.86 at.% Zr addition occurs along the plane of maximum normal stress without showing shear fracture. The fracture surface mainly displays cleavage facets which correspond to the eutectic β-Ti phase. The presence of rod shape TiFe in the smooth cleavage surface can be observed (Fig. 4(c)). This suggests that the β-Ti phase is prone to initiate cleavage. Possibly the stress concentration raised from the dislocation pile-up stress at the primary β-Ti/TiFe interface is relieved by slip transfer. It is believed that the stress required to initiating cleavage in the primary β-Ti phase is lower than the pile-up stress required to activating the slip first[37]. Similar fracture behavior is also observed in the hypoeutectic alloys with 8.79 at.% and 11.72 at.% Zr additions, but cleavage facets become smoother and larger due to the increases in size and number of the primary β-Ti phase (Fig. 4(d) and (e)). In the case of the binary Ti_70.5Fe_29.5 eutectic alloy, there is also clear evidence of radial cracking at the β-Ti/Ti₄Fe₂O interface (Fig. 4(f)), where acts as stress-concentrating site. It is believed that the brittle oxides offer no resistance to crack growth. Once the crack is formed in the case of the crack-nucleation controlled fracture, it immediately propagates to cause the alloy fracture.

3.3. Surface roughness and density

In checking the surface roughness, the top surface and the side wall were tested for each alloy. From a few initial measurements it was found that the roughness of top surface was approximately four percent greater than that of side wall. Since the largest roughness on each alloy was of primary interest, measurements were only taken on the top surface. Fig. 5 shows the typical 3D profiles taken from the top surfaces of the alloys. The data reveals that the surface roughness (R_a) gradually decreases with the increasing Zr addition to 5.86 at.%, and it increases afterwards. This suggests that the alloy with 5.86 at.% Zr addition has the best formability, as it has an eutectic structure and undergoes narrower solidification range than hypereutectic and hypoeutectic alloys, which will endow it with a high liquid flowability [[38], [39], [40]]. Moreover, the surface roughness of the multi-component eutectic alloy is comparable with that of the binary Ti_70.5Fe_29.5 eutectic alloy (R_a = 4.721 μm), because the purifying effect of Y restrains the formation of Ti₄Fe₂O oxides, leading to viscosity of the melt decreased, despite the fact that the multi-component eutectic alloy undergoes wider solidification temperature range than the binary eutectic alloy. Therefore, the solidification range and purity of the melt are the two crucial factors for the formability of the alloys.

Fig. 6 presents changing curve of densities of the alloys with Zr added amount. It is clear that the densities of the alloys monotonously increase with the increasing Zr addition, and are higher than that (5.69 g/cm³) of the binary Ti_70.5Fe_29.5 eutectic alloy. Internal quality and composition are two main factors affecting the density of a material. In current practice, it is believed that the internal quality has little effect on the densities of the alloys, because these alloys are dense and defect-free. Therefore, the densities of the alloys can be associated with the composition. Zr element is designed to replace Ti element, comparatively speaking, the atomic weight of the former (91.22) is much higher than that of the latter (47.87). Therefore, the increased Zr amount and the decreased Ti amount inevitably increase the densities of the alloys, as the amounts of other elements are fixed. Moreover, combined with the strength data listed in Table 1, one can draw a conclusion that the strength/density ratio of the alloys decreases with the increasing Zr addition.

3.4. SBF bioactivity

Fig. 7 shows the surface morphologies of the alloys after immersion in SBF solution for four days. It is clear that only a small number of granular sediments, growing in the form of aggregates, occur on the local surface of the hypereutectic alloy with 1.47 at.% Zr addition (Fig. 7(a)). EPMA analysis reveals that sediments contain Ca, P, C, O, Ti, Fe, and a small amount of Na and S, among which the atomic ratio of Ca and Ti is 0.056. Thereby it can be inferred that the sediments may be composed of apatite containing carbonate ions [[41], [42], [43], [44], [45]]. An entirely different surface morphology is observed for the hypereutectic alloy with 2.93 at.% Zr addition. The surface is almost fully covered by the sediment layer with furrow-like grain boundaries (Fig. 7(b)). Meanwhile, the atomic ratio of Ca and Ti in the layer increases to 0.275, as evident from EPMA analysis. Similar surface morphology is also observed in the case of the eutectic alloy with 5.86 at.% Zr addition, but the difference is that the sediment layer becomes thick, and the morphology of grain boundary changes from furrow-like to net-like (Fig. 7(c)). The increased thickness makes the layer low in Ti content, so that the atomic ratio of Ca and Ti is further increased to 0.475. However, the sediment layer becomes thin for the hypoeutectic alloy with 8.79 at.% Zr addition, the microstructure under the sediment layer is clearly observed (Fig. 7(d)). Meanwhile, the grain boundary changes into furrow-like shape again. As a result, the atomic ratio of Ca and Ti in the sediment layer is decreased to 0.066. The decreasing trend in the atomic ratio of Ca and Ti continues to the hypoeutectic alloy with 11.72 at.% Zr addition (its value is 0.047). On the other hand the sediment layer becomes loose and rough (Fig. 7(e)). From the Ca/Ti atomic ratio and surface morphology information, it can be concluded that the eutectic alloy with 5.86 at.% Zr addition has the best apatite deposition ability among the studied alloys. As for the binary Ti_70.5Fe_29.5 eutectic alloy, the sediment layer is not continuous and becomes thin, on which some flocculent sediment is clearly observed (Fig. 7(f)). As a result, the atomic ratio of Ca and Ti decreases to 0.032, much lower than those of the studied alloys. With regard to the Ti-6Al-4 V alloy, some fine sediments, growing in the form of partition aggregates, occur on the local surface of the Ti-6Al-4 V alloy (Fig. 7(g)). As a consequence, the atomic ratio of Ca and Ti decreases to 0.0394, much lower than those of the studied alloys. The composition of the sediments on the surface of the alloys is listed in Table 2.

In previous literatures [43,46], the mechanism of apatite deposition on Ti alloy is that hydroxyl groups (OH^-) are adsorbed from SBF to form Ti-OH groups. When the pH is approximately 7.4, the Ti-OH groups are negatively charged, due to the presence of deprotonated acidic hydroxides. Calcium ions (Ca²⁺) are then adsorb from the SBF to the alloy surface, as a result of the negatively charged surface, so that HOPO₄^2- or H₂PO₄^- can easily react with the as-adsorbed Ca²⁺ to finally produce calcium phosphate. Therefore, the formation of Ti-OH groups on the Ti alloy is crucial for apatite deposition. In current practice, the β-Ti and TiFe phases can generate galvanic microcells in SBF, of which β-Ti with the high reactivity is acted as the anode with respect to TiFe with the low reactivity. Thus, the corrosion attack initiates from the β-Ti phase, leading to the dissolution of the β-Ti phase and formation of Ti-OH groups. The finer grain is, the larger anode to cathode area ratio is, and the more Ti-OH groups will be formed. This explains why the eutectic alloy with 5.86 at.% Zr addition, having the finest grain, exhibits the best ability of apatite deposition. Although the binary Ti_70.5Fe_29.5 eutectic alloy has fine eutectic microstructure, the existence of Ti₄Fe₂O oxides at eutectic colony boundary weakens the deposition ability of apatite.

3.5. Cytotoxicity

Observation of cell morphology shows that l-929 fibroblasts in the extracts of the studied alloys adhere to the wall and display the morphological characteristics of spindle and polyon. Such a representative morphology is shown in Fig. 8(a), which is similar to that in the SFM negative control (Fig. 8(b)). While most of cells in positive control group become small and round with karyopyknosis, being a poisoning morphology (Fig. 8(c)).

In the further MTT assay, large amounts of needle-like crystals are observed in different concentrations of extracts of each alloy. As a representative, the morphology of the crystals in different concentrations of the eutectic alloy with 5.86 at.% Zr addition is shown in Fig. 9. It can be found that the crystals shown in Fig. 9(a)-(d) are very similar to those in the negative control (Fig. 9(e)). In contrast, no crystals are observed in positive control (Fig. 9(f)), because the cytotoxicity of phenol leads to cell death. Table 3 lists OD value, relative growth rate, and toxicity level of the studied alloys, Ti_70.5Fe_29.5 alloy and Ti-6Al-4 V alloy. Statistical results display that OD value and relative growth rate of the studied alloys, Ti_70.5Fe_29.5 alloy and Ti-6Al-4 V alloy have no obvious difference with that of the negative control. The toxicity level of the studied alloys, Ti_70.5Fe_29.5 alloy and Ti-6Al-4 V alloy in different concentrations of extracts is 0 or 1. Thus one can draw the conclusion that the studied alloys, Ti_70.5Fe_29.5 alloy and Ti-6Al-4 V alloy have no cytotoxicity. While the positive control group is cytotoxic as its toxicity level is 3, associating with the cytotoxicity of phenol.

3.6. Cell adhesion and proliferation

Cell adhesion assay shows that round cells have adhered to the surfaces of alloys after being incubated for 30 min, and show increasing trend in number with the extension of incubation time. Such a typical change is shown in Fig. 10. One should mention that the cell adhesive levels on the hypereutectic, eutectic and hypoeutectic alloys are very similar. Statistic analysis reveals that the OD values of the cells adhered to the studied alloys have no significant difference with that of Ti_70.5Fe_29.5 alloy, Ti-6Al-4 V alloy and control group as shown in Table 4. The result means that the studied alloys can promote the early adhesion of cells, and are beneficial to growth of the fibroblast cells.

In the further cell proliferation assay, one can see that l-929 fibroblast cells on the studied alloys have morphological characteristics of spindle after being incubated for 1 d. With the extension of incubation time, the cells exhibit significant increase in number, and almost fully cover the whole surface of the alloys. A typical example is shown in Fig. 11. Statistic analysis reveals that the OD values of the cells proliferated on the studied alloys have no significant difference with that of Ti_70.5Fe_29.5 alloy, Ti-6Al-4 V alloy and control group (Table 5) indicating that l-929 fibroblast cells have excellent proliferation ability on the studied alloys.

4. Conclusions

The influences of Zr on mechanical, forming, and biological properties of the Ti-Fe-Zr-Y alloys were investigated in detail. The main results are summarized as follows:

(1)With the increasing of Zr addition, the surface roughness, friction coefficient and worn volume of the alloys first decrease and then increase, with the lowest values obtained at the eutectic alloy with 5.86 at. % Zr addition, while ultimate compression strength and specific strength gradually decrease.

(2)The ability of apatite deposition on the studied alloys is closely related to grain size: the finer grain of the alloys is, the higher the ability of apatite deposition is, i.e., the eutectic alloy with 5.86 at.% Zr addition has the best ability of apatite deposition among the studied alloys.

(3)The studied alloys have no cytotoxicity. They can promote the early adhesion of cells, and are beneficial to growth of the fibroblast cells.

Acknowledgement

This work was supported financially by the National Natural Science Foundation of China (No. 51371041).

The authors have declared that no competing interests exist.

---