Heat Treatment Degrading the Corrosion Resistance of Selective Laser Melted Ti-6Al-4V Alloy

Cubic samples of the Ti-6Al-4V alloy of a size of 10 mm × 10 mm × 10 mm were manufactured by SLM in an MTT SLM 250 HL machine, which is equipped with a 400 W Yb: YAG fiber laser, at a laser wavelength of 1070 nm, a maximum power of 200 W in continuous laser mode, and a spot size of 80 μm. The SLM-produced samples were produced with a scanning speed of 1250 mm/s. Both the hatch spacing (distance between scan lines) and the layer thickness were 100 μm. The layers were scanned using continuous laser mode according to a zigzag pattern, which was alternated by 90° between each successive layer. This parameter setup guarantees near-full density (i.e. greater than 99%) and good surface quality of the SLM-produced Ti-6Al-4V samples. The chemical composition of the SLM-produced Ti-6Al-4V alloy in wt% was 0.01 C, 0.002 H, 0.14 O, 0.02 N, 0.22 Fe, 6.25 Al, 4.04 V and Ti balance. Fig. 1 schematically presents the picture of a Ti-6Al-4V alloy sample fabricated through SLM technique.

Zoom In Zoom Out Reset image size

Prior to heat-treatment, the side view of the SLM-produced Ti-6Al-4V alloy (i.e. the XZ plane³¹) was chosen as the studied plane, and its surface was abraded with SiC paper down to 2000 grit, cleaned with double-distilled water, ultrasonically cleaned in ethanol, and then dried in air. The SLM-produced Ti-6Al-4V alloy samples were heat treated at 500, 850 or 1000°C for 2 hours in a tube furnace. The heat-treatment followed the following procedures. First, the as-received SLM-produced Ti-6Al-4V samples were enclosed in a quartz tube, and then argon atmosphere was used to purge the oxygen inside the tube to minimize the oxidation of the titanium samples. The samples were then heated at about 3°C/min from room temperature (25°C) to 500, 850 or 1000°C. After having been held for 2 hours at each temperature, the samples were cooled to room temperature inside the tube, known as furnace cooling (at an approximate cooling rate of 1.67°C/min).

Electrochemical measurements including open circuit potential (OCP), potentiodynamic polarization and electrochemical impedance spectroscopy (EIS), were conducted using a Solartron SI1280B electrochemical station in a conventional three-electrode cell. All the alloy samples (including the as-received SLM-produced and heat-treated samples) were mounted in a plastic tube and sealed with epoxy resin with exposure of the side view to the outside. Afterward, all the samples were abraded with SiC paper down to 2000 grit, cleaned with double-distilled water, followed by ultrasonic cleaning in ethanol, and finally dried in air.

In the electrochemical measurements, a platinum sheet was used as the counter electrode, a saturated calomel electrode (SCE) acted as the reference electrode, and the surface of the samples with an area of 1.0 cm² served as the working electrode. Prior to the polarization and EIS tests, the samples were immersed in the electrolyte solution for sufficient time to gain a stable OCP. The potentiodynamic polarization curves were tested in the potential range from −0.5 to +2.5 V versus OCP, with a scan rate of 0.1667 mV/s. EIS tests were conducted at the OCP potentiostatically with a potential perturbation amplitude of 10 mV. The impedance spectra presented as Bode plots were measured in the frequency range from 100 kHz to 10 mHz. All reported potentials in this work were relative to SCE. The ZsimpWin Software was employed to fit the EIS data. All the electrochemical measurements were carried out in 3.5 wt% NaCl solution at room temperature (25 ± 1°C). The electrochemical tests for each sample were performed at least three times for data reproducibility.

The phase compositions of the as-received and heat-treated SLM-produced Ti-6Al-4V samples were characterized by X-ray diffraction (XRD) using a Bruker D8Advance X-ray diffractometer with Cu Kα radiation at room temperature (25°C). The scanning range of 2θ was from 30° to 80°. The Jade 5.0 software was employed to calculate the different phase contents in the SLM-produced Ti-6Al-4V alloys. A Zeiss optical microscope (AxioCam MRc5, Germany) was employed to examine the microstructural characteristics of both the SLM-produced Ti-6Al-4V alloy samples. With regard to the microstructural characterization of the studied samples, a Kroll's solution composed of 10 ml HF, 15 ml HNO₃ and 75 ml H₂O, was prepared as an etchant solution. All the solutions used as electrolyte and etchant solution were prepared with analytical-grade reagents and double-distilled water. After the electrochemical tests, the corresponding samples were experienced electrochemical corrosion beyond the passive region through polarization. The surface morphology of the as-received and heat treated SLM-produced alloys after electrochemical corrosion was characterized using JSM-7800F (JEOL, Tokyo, Japan) field emission scanning electron microscopy (FE-SEM) at an accelerating voltage of 15 kV.

Fig. 2 shows the optical microstructure of the as-received and heat-treated SLM-produced Ti-6Al-4V alloy samples (side view). Fig. 2a shows the microstructure of the as-received SLM-produced alloy. It can be seen that acicular α^' martensite is overwhelmingly distributed in the whole microstructure, accompanying some typical long, columnar prior β grains that grow along the build direction. Furthermore, fine α^' martensite can be also observed inside the prior β grains, as seen from the magnified image in the inset of Fig. 2a. Some pores can also be captured in the microstructure. A detailed description of the microstructure of the SLM-produced Ti-6Al-4V alloy could be referred to our previous studies.^30,31 An explanation of the formation of prior β grains and the existence of pore defects has been discussed previously.¹ As for the heat-treated sample at 500°C, no obvious changes are observed in the microstructure of the SLM-produced sample (Fig. 2b); the prior β grains and fine α^' martensitic phases are still visible in the microstructure. On the other hand, the α^' martensitic phases are partially transformed into the plate-shaped α phase. The microstructure of the sample heat treated at 850°C (still below T_β) undergoes significant changes (Fig. 2c). The fine acicular α^' martensite and columnar prior β grains vanish, and transform into the plate-shaped α phase and the β grains. As the β transus temperature (T_β) of Ti-6Al-4V alloy is about 995°C, a homogeneous 100% β phase would form in the microstructure when heat treated above T_β,¹⁴ while a microstructure of lamellar α + β mixture is formed during the furnace cooling process. For the heat treated temperature at 1000°C (i.e. beyond T_β), the microstructure of the heat-treated sample is substantially different to that treated below T_β; the long columnar β grains disappear, indicating the extensive grain growth of the SLM-produced alloy up to the point of semi-equiaxed β grains. On the other hand, a mixture of lamellar α + β microstructure forms due to the furnace cooling. As seen in the magnified image in the inset of Fig. 2d, the light areas correspond to the α phase, and the dark areas belong to the lamellar transformed β phase. Similar results of the heat-treated effects on the microstructure of SLM-produced Ti-6Al-4V alloys were obtained by Vilaro et al.²⁹ and Vrancken et al.¹⁴

Zoom In Zoom Out Reset image size

Complementary microstructural analysis was performed for the as-received and heat treated SLM-produced Ti-6Al-4V alloy samples. Fig. 3a displays the XRD pattern of the as-received SLM-produced sample; the main peaks of (100)_α^'-Ti, (002)_α^'-Ti, (101)_α^'-Ti, (102)_α^'-Ti, (110)_α^'-Ti, (112)_α^'-Ti, (201)_α^'-Ti are observed. It has been pointed out that the β-Ti phase is hard to detect, or exhibits minority, or is even absent in the SLM-produced Ti-6Al-4V alloy.^14,24 It has also been proposed that the peak near 2θ = 72° could be the α-Ti or β-Ti phase.^38–40 As for the sample heat treated at 500°C (Fig. 3b), the major peaks remain the same as in the as-received sample. When heat treated at 850°C, the (110)_β-Ti peak is present in the XRD pattern. At the same time, the peaks near 2θ = 35°, 39°, 41°, 53.5°, 64°, 77°, 78° correspond to the plate-shaped α phase; the microstructure in Fig. 2c proves that the slim, long plate shape is the α phase. More obviously, the peak of β-Ti near 2θ = 39.5° in the XRD pattern indicates that an amount of β-Ti phase is yielded in the microstructure, so that it can be examined by X-ray. As can be seen in Table I, the phase constituents and their volume fraction (V_f) of as-received and heat-treated SLM-produced Ti-6Al-4V alloys were calculated from the XRD patterns by means of the Jade 5.0 software. The method of calculating the phase content in microstructure is described elsewhere.^41-43 The volume fraction of the β phase in the microstructure of the as-received and heat-treated samples (500°C) are very close, while the volume fraction of the β phase increases evidently after heat-treatment at higher temperatures of 850°C and 1000°C. In addition, the peaks of the α-Ti phase display a strong intensity, which is related to the increasing of the grain size. The results of the XRD patterns are in good agreement with the microstructural studies.

Zoom In Zoom Out Reset image size

Fig. 4 shows the OCP curves of the as-received and heat-treated SLM-produced Ti-6Al-4V alloy samples immersed in 3.5 wt% NaCl solution. The OCPs of both the as-received and heat-treated samples exhibit a positive shift with prolonging the immersion time, suggesting that a passive film TiO₂ formed on the sample surface. It takes approximate 50 h for different samples to gain relatively stable OCP values. The slight OCP change rate of 1.5 mV/h after 50 h immersion does not significantly influence the successive EIS tests. The final OCP value of the as-received SLM-produced Ti-6Al-4V alloy is −79 ± 9.5 mV. The OCPs for the SLM-produced Ti-6Al-4V alloy samples heat treated at 500°C and 850°C exhibit a lower value compared to that of the as-received counterpart. The OCP values for the samples heat treated at 500°C and 850°C are −115 ± 10.6 mV and −155 ± 12.1 mV, respectively. As for the sample heat treated at 1000°C, the OCP exhibits the most negative value of −175 ± 1.7 mV.

Zoom In Zoom Out Reset image size

Fig. 5 shows the potentiodynamic polarization curves of as-received and heat-treated SLM-produced Ti-6Al-4V alloys. The fitted values of corrosion current density (i_corr) and passive current density (i_p) on the basis of the polarization curves from Fig. 5 are summarized in Table II. For the as-received sample, the SLM-produced Ti-6Al-4V alloy exhibits a very low potential-dependent passive current density in the potential range from 650 mV to 1200 mV; i_p of as-received SLM-produced alloy (i_p,A) is about 0.9 ± 0.04 μA cm⁻². Furthermore, it is interesting to note that second passivation is observed at the potential of 1600 mV. In general, the passivation current density (i_p) describes the stability of the passive film and states how much current density could support the film generation period. When heat treated at 500°C, the polarization curves of the SLM-produced Ti-6Al-4V alloy also display a passivated behavior in the potential range between 600 mV and 1250 mV; the i_p,B is 1.3 ± 0.07 μA cm⁻², which is slightly higher than that of the as-received ones. This reveals that the heat-treated sample at 500°C exhibits an inferior passivated behavior compared to the as-SLM-produced sample. Likewise, the alloy sample heat treated at 500°C also has a second passivation beyond the potential of 1500 mV; the value of the second-passivation current density however seems lower than that of the as-received ones. For the SLM-produced Ti-6Al-4V alloy heat treated at 850°C, it shows a similar electrode behavior compared with the samples treated at 500°C; it seems that the sample exhibits a little higher i_p,C of 1.5 ± 0.05 μA cm⁻². The polarization curves for the sample heat treated beyond the β-transus temperature at 1000°C, the polarization curves exhibit a similar behavior as former samples, but without a fixed i_p. The passive film undergoes increasing dissolution during the entire anodic branch.

Zoom In Zoom Out Reset image size

On the other hand, in order to evaluate the corrosion rate of different samples, i_corr is additionally fitted by the Corrshow software to calculate the corrosion rate of the heat-treated samples. Atapour et al.^44,45 and Yilbas et al.⁴⁶ also evaluated the corrosion rate of titanium alloys through the fitting of i_corr. The fitted results of i_corr are listed in Table II. It is clear that the heat-treated samples exhibit a higher corrosion rate compared to the as-received ones; the higher the heat-treatment temperature, the higher the corrosion rate (i_corr). Thus, heat-treatment fails to improve the corrosion resistance of the SLM-produced Ti-6Al-4V alloy, instead of increasing the corrosion rate of the SLM-produced samples. In addition, it should be noted that the values of E_corr obtained from the potentiodynamic polarization curves are more negative compared with that of the OCP values, which is ascribed to the cathodic polarization for 0.5 V during the polarization tests. Such a large cathodic polarization potential removes the oxide film on the sample, leading to decrease in the cathodic kinetics and increase in anodic kinetics, as a result of the negative shift of E_corr compared with the OCP values. But the OCP and E_corr exhibited very close values. The corrosion rate of the sample in solution system is a result of the mixture of the anodic and cathodic reactions, the calculation of the i_corr is also based on the anodic and cathodic branches of the polarization curve.

The EIS measurements were performed to further study the interface information and electrochemical process of the as-received and heat-treated SLM-produced Ti-6Al-4V alloys in the 3.5 wt% NaCl solution. Fig. 6 shows the Bode plots of the tested samples, and the equivalent circuit in Fig. 6a was employed to fit the EIS data. The Nyquist plots (not shown in the paper) only exhibit a huge capacitance loop, while the Bode plots show a wide plateau from high frequency to low frequency. The equivalent circuit with one time constant was used to fit EIS data. In view of the equivalent circuit, R_s represents the solution resistance, R_ct stands for the charge transfer resistance. The constant phase element (CPE) corresponds to the electric double layer. The fitted results of EIS data for the as-received and heat treated SLM-produced Ti-6Al-4V alloys in 3.5 wt% NaCl solution are listed in Table III. Fig. 7 shows the fitted results of R_ct for as-received and heat-treated SLM-produced Ti-6Al-4V alloys. R_ct of the as-received SLM-produced sample also shows a very high value, while the value of R_ct significantly decreases for the heat-treated samples. This reveals that the SLM-produced Ti-6Al-4V alloy subjected to post-heat-treatment exhibits a higher corrosion rate in comparison with their as-received counterparts. The results of the polarization curves strongly support the EIS tests. Additionally, the fitted results of the Y_O (CPE, constant phase element) and n of double layer are also exhibited in Table III. According to the n value closely approaching 1, the electric double layer possesses a very high capacitance characteristic. The double layer capacitances for as-received and heat treated samples also exhibit very close values.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

According to our previous studies,^30,31 the amount of acicular α^' martensite in the microstructure is considered as the major reason for accelerating the corrosion of the SLM-produced Ti-6Al-4V alloy compared to the traditional Grade 5 alloy. It is expected that the elimination of the fine α^' martensitic phase could improve the corrosion resistance of the SLM-produced alloy. Herein, fine α^' martensite is successfully removed through heat-treatment (both below and above the β transus temperature T_β). However, the corrosion resistance of the SLM-produced Ti-6Al-4V alloy is unfortunately not enhanced by heat-treatment, while decreases its corrosion resistance. According to the XRD calculated results, the volume fraction of the β phase increases with the heat-treatment temperature rising, especially at very high temperatures (i.e. 850–1000°C). In our previous studies and in other reports,^30,31,46 it has been proposed that the β phase in the microstructure improves the corrosion resistance of Ti-6Al-4V alloys. On the other hand, the corrosion resistance of SLM-produced Ti-6Al-4V alloys fails to improve in spite of the increase in the content of the β phase after the heat-treatment process. According to the microstructural studies (Fig. 2), the grain size significantly increases with increasing the heat-treatment temperature, despite the slight changes at a low temperature (500°C). When heat treated at 850°C, the width for grain size of plate shaped α phase is about 1.14 ± 0.28 μm, whilst it is 5.71 ± 1.43 μm for width of α phase in lamellar α + β mixture when heat treated beyond T_β at 1000°C. It is worth noting that some authors have shed light on the effects of grain size on the corrosion behavior of the metals in corresponding systems. Huang et al.⁴⁷ investigated the effects of grain refinement on corrosion behavior of Ti–25Nb–3Mo–3Zr–2Sn alloy and proposed that the corrosion resistance of studied alloy considerably increases as the grain size reduces. Ralston et al.⁴⁸ also studied the effect of grain size on corrosion of high purity aluminum in 0.1 M NaCl solution and found that the corrosion rate tends to decrease with decreasing grain size. Therefore, it seems the increase in the grain size could enhance the corrosion resistance of the alloys. In that case, as for the SLM-produced Ti-6Al-4V alloy, the increasing grain size through heat-treatment increases its corrosion rate thereby weakening its passivation behavior. It is believed that the effect of the grain size increasing on corrosion resistance overwhelms the increased amount of the β phase in microstructure.

In order to further investigate the corrosion resistance and passive film stability of the corresponding SLM-produced samples, the surface morphologies of the as-received and heat treated SLM-produced Ti-6Al-4V alloys after electrochemical corrosion beyond passive potential were performed and are shown in Fig. 8. For the as-received sample, some dissolved traces exhibited after polarization beyond the passive potential. Different dissolution property of some parts caused the color difference as can be seen some dark areas. As for the sample heat treated at 500°C (Fig. 8b), some small dark areas are generated and little corrosion pits are observed. As seen from Fig. 8c, more corrosion pits are observed inside the dark regions, suggesting that the passive film formed on the dark regions is weak in the protective ability against the electrochemical polarization. It is more pronounced that more pits are further generated on these dark places for the case of the sample heat treated at 1000°C. All these support the fact that the heat-treatment reduces the corrosion resistance and passive film stability of the SLM-produced samples. Furthermore, it is highly possible that the oxygen could generate during the anodic polarization at above 1 V. As a matter of fact, no occurrence of gas evolution happens during the polarization according to our studied system. The pits forming in the samples could be responsible for the high current through a high potential polarization.

Zoom In Zoom Out Reset image size

Therefore, unlike the enhancement in mechanical properties by heat-treatment, the process seems to not to enhance the corrosion resistance of the SLM-produced Ti-6Al-4V alloy. In contrast, heat-treatment significantly reduces its corrosion resistance and stability of the passive film on the sample surface. With regards to its wide application in many vital fields, considerable work is still needed to improve the corrosion resistance of SLM-produced Ti-6Al-4V alloys.