Study on Corrosion Behavior of Ultrafine-Grained Ti-6Al-7Nb Fabricated by Equal Channel Angular Pressing

The aim of this study was to investigate the corrosion resistance of ultrafine-grained (UFG) Ti-6Al-7Nb fabricated by equal channel angular pressing (ECAP) and coarse-grained (CG) Ti6Al7Nb. The microstructure of each specimen was investigated by the electron backscattered diffraction (EBSD) method. The corrosion behavior of each specimen was determined by electrochemical measurement in Ringer’s solution. The surface corroded morphologies and oxide film formed on Ti-6Al-7Nb alloy after electrochemical measurement were investigated by scanning electron microscope (SEM) and X-ray photoelectron spectroscopy (XPS). EBSD investigation shows that the grain size of UFG Ti-6Al-7Nb decreased to ~0.4 μm, accompanied by low angle grain boundaries (LAGBs) accounting for 39%. Potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) results indicated that UFG Ti-6Al-7Nb alloy possessed a better corrosion resistance. The surface corroded morphologies revealed many small and shallow corrosion pits, which can be attributed to the good compactness of the oxide film and a rapid selfrepairing ability of the UFG Ti-6Al-7Nb alloy.


Introduction
Titanium and its alloys have been widely used in aerospace engineering, petro-and chemical engineering, as well as biomedical engineering due to their high specific strength, high corrosion resistance, low elastic modulus, and good biocompatibility [1][2][3][4][5][6]. Ti-6Al-4V alloy was first introduced as an engineering material becoming the most used titanium alloy worldwide. It has been widely applied in aircraft turbines [4,5], chemical equipment [7], and bio-implants [8] etc. Ti-6Al-7Nb alloy, a modified type of Ti-6Al-4V by replacing the toxic V by bio-inert Nb and maintaining the good mechanical properties, is considered as a better material for biomedical application [9,10]. Recently, ultrafine-grained (UFG) and nanocrystalline (NC) materials processed by severe plastic deformation (SPD) have been widely investigated and found to have unique physical and chemical properties [11][12][13]. Enhanced mechanical behavior and biocompatibility of UFG/NC Ti-6Al-7Nb alloy were acquired after the SPD process [14][15][16][17][18][19]. For instance, Polyakova et al. [9] revealed that the tensile strength of the UFG Ti-6Al-7Nb alloy with grain size of 330 nm increased to 1210 MPa after equal channel angular pressing (ECAP), which was 20% higher than observed in coarse-grained (CG) counterparts. Ashida et al. [15] Metals 2020, 10, 950 2 of 11 obtained a Ti-6Al-7Nb grain size of~100 nm by high pressure and torsion (HPT) at room temperature. Results indicated that Vickers microhardness increases from 325 HV to 385 HV after HPT. Furthermore, the sample with refined grain size exhibited an excellent superplastic elongation of 930% at an initial strain rate of 2 × 10 −3 s −1 . Oliveira et al. [16] reported that Ti-6Al-7Nb alloy processed by ECAP with an average grain size~200 nm had higher fatigue properties, which make it suitable for practical application. At the same time, a highly bioactive behavior of UFG Ti-6Al-7Nb alloy after phosphoric acid etching, regardless of the usage of alkaline treatment, was reported by Oliveira et al. [17].
Implanted biomedical materials have been immersed in human body fluids for several years, hence they are inevitably subject to attack by a corrosive environment. However, the corrosion behavior of UFG Ti-6Al-7Nb has not received enough attention and the effect of the refined grain size on corrosion resistance of titanium alloys is in dispute [20][21][22][23][24][25][26][27]. For instance, Legostaeva et al. [23] investigated the electrochemical behavior of CG and NC titanium and found that the dissolution rate of NC titanium increases in a Ringer-Locke solution in comparison with the CG state. However, Fattah-Alhosseini et al. [24] investigated the passivation and electrochemical response of NC titanium in Ringer's physiological solution and revealed that the passivation response of the NC sample improved compared to the CG state due to the formation of a thicker and less defective oxide film.
Therefore, the corrosion behavior of UFG Ti-6Al-7Nb alloy fabricated by ECAP was investigated by the methods of potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) in Ringer's solution. The present study aimed to clarify the effect of the grain size on the corrosion resistance by considering the impact of other microstructural parameters and the formation of surface oxide film.

Materials
Grain refinement of Ti-6Al-7Nb alloy used in this study was conducted by the ECAP method (Ufa State Aviation Technical University, Ufa, Russia). Thermal treatment was used to obtain a CG sample after ECAP, which was annealed at 995 • C for 1 h, then annealed at 550 • C for 4 h followed by air cooling to form a "duplex" structure.

Microstructure Observation
Microstructure characterization of Ti-6Al-7Nb samples was performed by electron backscattered diffraction (EBSD) embedded in the Field Emission Scanning Electron Microscope (SEM, JSM-6700F, Jeol, Tokyo, Japan). Before the EBSD observation, the samples were mechanically polished to a mirror like surface by SiC paper followed by electro-polishing (solution: HClO 4 :C 2 H 5 OH = 19:1 (volume ratio); voltage: 40 V; polishing time: 90 s). Automated EBSD scans were applied by Flamenco data acquisition software (Aztec 3.2, Oxford Instruments, Abingdon, United Kingdom). The average grain size and grain boundary orientation were processed with HKL Technology Channel 5 software (Oxford Instruments, Abingdon, United Kingdom).

Electrochemical Measurements
Corrosion behavior of Ti-6Al-7Nb samples was investigated by electrochemical testing in Ringer's solution (9 g/L NaCl, 0.42 g/L KCl, and 0.25 g/L CaCl 2 ) at pH 7.4. The experiment used a three-electrode system with the sample set as working electrode, the saturated Ag/AgCl electrode set as reference electrode, and a platinum piece set as auxiliary electrode. The surface of both tested samples and reference electrode were mechanically polished to a mirror like surface by SiC paper before each test. All tests were carried out by using electrochemical station (CHI660E, CH Instruments, Austin, TX, USA)) in a 37 • C water bath. The Open Circuit Potential was measured for 10 min. Electrochemical impedance spectroscopy (EIS) was performed at a frequency range from 100 kHz to 10 mHz with an AC voltage amplitude of ±10 mV. Potentiodynamic polarization data were acquired between -650 to 3 of 11 850 mV vs Ag/AgCl with 2 mV/s sweeping rate after the EIS test. Corroded morphology was observed by Scanning Electron Microscopy (SEM, JSM-6700F, Jeol, Tokyo, Japan) after the electrochemical tests.

Passivation Film Characterization
The samples' oxide film character was investigated by X-ray Photoelectron Spectroscopy (XPS), with an electron spectrometer (K-Alpha+, Thermo Scientific, Waltham, MA, USA). Al Ka was used as the X-ray source, the calibration of the photoelectron binding energy was performed using a C 1s signal with a binding energy of 284.8 eV. The surface chemical composition of the passivation film formed on the samples was fitted and analyzed by Avantage software (Avantage 5.9918, Thermo Scientific, Waltham, MA, USA). Figure 1 illustrated the microstructure and grain size distribution of different Ti-6Al-7Nb samples. It can be observed that fairly equiaxed and homogeneous grains exsist in the microstructure of the CG sample (Figure 1a), and the average grain size achieved 13.8 µm (Figure 1c). On the other hand, the microstructure of the UFG sample displays a remarkable grain refinement (Figure 1b), which can be attributed to the large volume of accumulative strain during the ECAP process. Although several non-uniform deformation areas were presented in the UFG Ti-6Al-7Nb sample, equiaxed and homogeneous microstructure were still revealed in the most part of the microstructure. The average grain size was reduced to 0.4 µm with volume fraction (~55%) located between 200-400 nm (Figure 1d). after the electrochemical tests.

Passivation Film Characterization
The samples' oxide film character was investigated by X-ray Photoelectron Spectroscopy (XPS), with an electron spectrometer (K-Alpha+, Thermo Scientific, Waltham, MA, USA). Al Ka was used as the X-ray source, the calibration of the photoelectron binding energy was performed using a C 1s signal with a binding energy of 284.8 eV. The surface chemical composition of the passivation film formed on the samples was fitted and analyzed by Avantage software (Avantage 5.9918, Thermo Scientific, Waltham, MA, USA). Figure 1 illustrated the microstructure and grain size distribution of different Ti-6Al-7Nb samples. It can be observed that fairly equiaxed and homogeneous grains exsist in the microstructure of the CG sample (Figure 1a), and the average grain size achieved 13.8 µm ( Figure  1c). On the other hand, the microstructure of the UFG sample displays a remarkable grain refinement (Figure 1b), which can be attributed to the large volume of accumulative strain during the ECAP process. Although several non-uniform deformation areas were presented in the UFG Ti-6Al-7Nb sample, equiaxed and homogeneous microstructure were still revealed in the most part of the microstructure. The average grain size was reduced to 0.4 µm with volume fraction (~55%) located between 200-400 nm ( Figure 1d).

Microstructural Evolution
Grain boundary and disorientated distribution of CG and UFG Ti-6Al-7Nb alloy are displayed in Figure 2. The black lines represent the high angle grain boundaries (HAGBs, θ > 15°), while the red lines represent the low angle grain boundaries (LAGBs, 2° < θ < 15°). It can be seen that the grain boundary character changed obviously accompanied by an increase of grain size (Figure 2a,b) after heat treatment. The grain boundary character distribution indicates that the fraction of HAGBs could achieve 61% for the UFG state ( Figure 2d) and increased to 77.6% (Figure 2c) for the CG counterparts. Furthermore, the fraction of HAGBs equal to 70° achieved 40% for the CG sample, which is different from the UFG state.  Grain boundary and disorientated distribution of CG and UFG Ti-6Al-7Nb alloy are displayed in Figure 2. The black lines represent the high angle grain boundaries (HAGBs, θ > 15 • ), while the red lines represent the low angle grain boundaries (LAGBs, 2 • < θ < 15 • ). It can be seen that the grain boundary character changed obviously accompanied by an increase of grain size (Figure 2a,b) after heat treatment. The grain boundary character distribution indicates that the fraction of HAGBs could achieve 61% for the UFG state ( Figure 2d) and increased to 77.6% (Figure 2c) for the CG counterparts. Furthermore, the fraction of HAGBs equal to 70 • achieved 40% for the CG sample, which is different from the UFG state.

Corrosion Behavior
It can be seen from Figure 3 that the open circuit potential (OCP) value of the sample reached stability within 10 min. Figure 4 represents a typical potentiodynamic polarization curves of the samples in Ringer's solution. The Ti-6Al-7Nb alloys with different grain size exhibit a similar polarization behavior. It could be found that the pitting corrosion took place on both samples. As the potential increased to ~0.4 V vs Ag/AgCl, both the UFG and CG sample reached their stable passivation current densities. The anodic zone clearly revealed a similar current slope before passivation, which indicates the formation of passive film on the sample surface. After that, with the potential continually increasing, the current density changed significantly which indicated the occurrence of pitting corrosion. Table 1 illustrates the results of electrochemical parameters calculated from potentiodynamic polarization curves in Ringer's solution, the average value of these electrochemical parameters was obtained based on three valid data. Ecorr is the corrosion potential while icorr is the corrosion current density. Origin software was used to manually fit Ecorr and icorr by the Tafel extrapolation method. In general, corrosion rate R is proportional to icorr, which can be calculated by Equation (1) = 3.27 × 10 × Where A represents the relative atomic mass, n and ρ represent the chemical valence state and density respectively.

Corrosion Behavior
It can be seen from Figure 3 that the open circuit potential (OCP) value of the sample reached stability within 10 min. Figure 4 represents a typical potentiodynamic polarization curves of the samples in Ringer's solution. The Ti-6Al-7Nb alloys with different grain size exhibit a similar polarization behavior. It could be found that the pitting corrosion took place on both samples. As the potential increased to~0.4 V vs Ag/AgCl, both the UFG and CG sample reached their stable passivation current densities. The anodic zone clearly revealed a similar current slope before passivation, which indicates the formation of passive film on the sample surface. After that, with the potential continually increasing, the current density changed significantly which indicated the occurrence of pitting corrosion. Table 1 illustrates the results of electrochemical parameters calculated from potentiodynamic polarization curves in Ringer's solution, the average value of these electrochemical parameters was obtained based on three valid data. E corr is the corrosion potential while i corr is the corrosion current density. Origin software was used to manually fit E corr and i corr by the Tafel extrapolation method. In general, corrosion rate R is proportional to i corr , which can be calculated by Equation (1) where A represents the relative atomic mass, n and ρ represent the chemical valence state and density respectively. can be calculated by Equation (1) = 3.27 × 10 × Where A represents the relative atomic mass, n and ρ represent the chemical valence state and density respectively. As can be seen from Table 1, the ipass of the UFG Ti-6Al-7Nb is lower than the CG state, which implies that the passivation film of UFG Ti-6Al-7Nb possesses a better self-repairing ability. The Epit of the UFG Ti-6Al-7Nb is higher than that of the CG state, which indicates that the pitting corrosion resistance of UFG sample has been improved. Generally, icorr and R are used to characterize the corrosion resistance of materials intuitively [21], the values of icorr and R of the UFG Ti-6Al-7Nb are much lower than the CG state, which indicates that the UFG Ti-6Al-7Nb sample possesses a better corrosion resistance.   Typical Nyquist diagrams for CG and UFG Ti-6Al-7Nb in Ringer's solution are presented in Figure 5. At the same time, Figure 5 shows an electrical equivalent circuit used to fit the experimental EIS data and evaluate the corrosion resistance of different Ti-6Al-7Nb samples. As can be seen in Figure 5, R1 and CPE1 represent the resistance and constant phase element of the passivation film, R2 corresponds to the resistance of the charge-transfer reactions while CPE2 represents the constant phase element of the electric double layer. The overall resistance of the material should be the sum of R1 and R2, which is the polarization resistance value (Rp). It is clearly shown in Figure 5 that the Nyquist diagram of the two tested samples demonstrates a similar semicircle, which implies that the same corrosion mechanism occurred. In general, the larger arc radius implies a better corrosion resistance of the sample [28][29][30]. The diameter of the semicircle for UFG Ti-6Al-7Nb is larger than the CG one, considering the exhibited higher polarization resistance (Rp) value of UFG Ti-6Al-7Nb alloy (Table 2), which means that the UFG Ti-6Al-7Nb alloy possesses a better corrosion resistance. As can be seen from Table 1, the i pass of the UFG Ti-6Al-7Nb is lower than the CG state, which implies that the passivation film of UFG Ti-6Al-7Nb possesses a better self-repairing ability. The E pit of the UFG Ti-6Al-7Nb is higher than that of the CG state, which indicates that the pitting corrosion resistance of UFG sample has been improved. Generally, i corr and R are used to characterize the corrosion resistance of materials intuitively [21], the values of i corr and R of the UFG Ti-6Al-7Nb are much lower than the CG state, which indicates that the UFG Ti-6Al-7Nb sample possesses a better corrosion resistance.

Epit
Typical Nyquist diagrams for CG and UFG Ti-6Al-7Nb in Ringer's solution are presented in Figure 5. At the same time, Figure 5 shows an electrical equivalent circuit used to fit the experimental EIS data and evaluate the corrosion resistance of different Ti-6Al-7Nb samples. As can be seen in Figure 5, R 1 and CPE 1 represent the resistance and constant phase element of the passivation film, R 2 corresponds to the resistance of the charge-transfer reactions while CPE 2 represents the constant phase element of the electric double layer. The overall resistance of the material should be the sum of R 1 and R 2 , which is the polarization resistance value (R p ). It is clearly shown in Figure 5 that the Nyquist diagram of the two tested samples demonstrates a similar semicircle, which implies that the same corrosion mechanism occurred. In general, the larger arc radius implies a better corrosion resistance of the sample [28][29][30]. The diameter of the semicircle for UFG Ti-6Al-7Nb is larger than the CG one, considering the exhibited higher polarization resistance (R p ) value of UFG Ti-6Al-7Nb alloy ( Table 2), which means that the UFG Ti-6Al-7Nb alloy possesses a better corrosion resistance. Metals 2020, 10, x FOR PEER REVIEW 6 of 10

Discussion
Based on the experimental results above, it can be concluded that corrosion resistance of UFG Ti-6Al-7Nb is much better than that for CG. The surface corroded morphologies of CG and UFG Ti-6Al-7Nb after electrochemical experiments are shown in Figure 6. A few corrosion pits were generated on the surface of all test samples. In particular, the corrosion pits of CG Ti-6Al-7Nb are much bigger and deeper than the UFG ones, reaching as far as ~4 µm. For CG Ti-6Al-7Nb, some corrosion points show a brighter picture due to the oxide formation surrounding the point as oxides holds a higher electron capture ability under SEM. It seems that the corrosion pits occur at the grain boundaries and then extend, as indicated in Figure 6. The corrosion is more likely to occur and extend at the HAGBs due to high interfacial energy and impurity segregation [31,32], while the presence of LAGBs could block intergranular corrosion from the perspective of grain boundary engineering [33,34]. Meanwhile, it is worth pointing out that the grain boundary misorientation angle in UFG Ti-6Al-7Nb distributes randomly, while the CG state concentrates at 70° (as shown in Figure 2), and corrosion pits can be found on the boundary.
Titanium and its alloys have good corrosion resistance, since an oxide film forms easily on the surface which can isolate the substrate of the metals from corrosive environments. Hence, the character of the oxide film plays an important role in the corrosion behavior of titanium alloys. As mentioned before, UFG Ti-6Al-7Nb alloy possesses a better ability on the oxide film of self-repairing, which can restrain the development of pitting when it occurs. It is well known that a large density of grain boundaries and dislocations can be generated with grain refinement in UFG Ti-6Al-7Nb during the SPD process. These defects could provide more nucleation sites for the oxide film and lead to a more rapid formation and a higher self-repairing ability of a passivation layer on the UFG Ti-6Al-7Nb alloy than for the CG [24,35].

Discussion
Based on the experimental results above, it can be concluded that corrosion resistance of UFG Ti-6Al-7Nb is much better than that for CG. The surface corroded morphologies of CG and UFG Ti-6Al-7Nb after electrochemical experiments are shown in Figure 6. A few corrosion pits were generated on the surface of all test samples. In particular, the corrosion pits of CG Ti-6Al-7Nb are much bigger and deeper than the UFG ones, reaching as far as~4 µm. For CG Ti-6Al-7Nb, some corrosion points show a brighter picture due to the oxide formation surrounding the point as oxides holds a higher electron capture ability under SEM. It seems that the corrosion pits occur at the grain boundaries and then extend, as indicated in Figure 6. The corrosion is more likely to occur and extend at the HAGBs due to high interfacial energy and impurity segregation [31,32], while the presence of LAGBs could block intergranular corrosion from the perspective of grain boundary engineering [33,34]. Meanwhile, it is worth pointing out that the grain boundary misorientation angle in UFG Ti-6Al-7Nb distributes randomly, while the CG state concentrates at 70 • (as shown in Figure 2), and corrosion pits can be found on the boundary.  Titanium and its alloys have good corrosion resistance, since an oxide film forms easily on the surface which can isolate the substrate of the metals from corrosive environments. Hence, the character of the oxide film plays an important role in the corrosion behavior of titanium alloys. As mentioned before, UFG Ti-6Al-7Nb alloy possesses a better ability on the oxide film of self-repairing, which can restrain the development of pitting when it occurs. It is well known that a large density of grain boundaries and dislocations can be generated with grain refinement in UFG Ti-6Al-7Nb during the SPD process. These defects could provide more nucleation sites for the oxide film and lead to a more rapid formation and a higher self-repairing ability of a passivation layer on the UFG Ti-6Al-7Nb alloy than for the CG [24,35].
The contents of titanium, niobium, aluminum, and their oxides were calculated and are listed in Table 3. As is well known, TiO 2 plays a more important role in corrosion resistance than other titanium oxides [36]. Obviously, the content of TiO 2 in the passivation film formed on UFG Ti-6Al-7Nb was higher than for CG. Also, the content of Nb2O5 in the UFG sample's passivation film was higher than for CG. The corrosion resistance and stability of Ti-6Al-7Nb alloy is much better than Ti-6Al-4V alloy, which is ascribed to the incorporation of niobium oxides, mainly Nb 2 O 5 [37,38]. Furthermore, the existence of Nb 5+ on the passivation film could lead to the excellent passivating properties of the anodically formed Ti(IV)-based surface oxide film and high corrosion resistance since Nb 5+ can eliminate stoichiometric defects (anion vacancies) caused by the presence of titanium suboxides [39]. Besides that, titanium alloy with dense Ti and Nb oxide in the passivation film has better corrosion performance than Ti-6Al-4V, and its passivation range is larger [40]. It can be seen from Table 3 that the content of Al 2 O 3 in the passivation film formed on UFG Ti-6Al-7Nb was lower than for CG. Al and its oxide Al 2 O 3 in the passivation film could be attacked easily by the corrosive environment which may cause surface pitting corrosion. It is worth mentioning that the oxides in the passivation film have different solubilities in different simulated body fluid environments. Milošev [41] reported that Al 2 O 3 has a higher solubility in artificial saliva containing fluoride ions in comparison to TiO 2 and Nb 2 O 5 , which indicate a worse corrosion resistance of Al 2 O 3 than TiO 2 and Nb 2 O 5 . Finally, it can be concluded that good compactness of the passivation film could be formed on UFG Ti-6Al-7Nb which led to a better corrosion resistance.

Conclusions
The corrosion behavior of Ti-6Al-7Nb alloy fabricated by ECAP was investigated by the methods of potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) in Ringer's solution in this study. The following conclusions can be drawn: (1) The average grain size of UFG Ti-6Al-7Nb processed by ECAP reached 0.4 µm. After heat treatment at 995 • C/1 h + 550 • C/4 h, the grain size increased to 13.8 µm and the volume of LAGBs decreased from 39% to 22.4%; (2) Electrochemical experiments indicated that UFG Ti-6Al-7Nb exhibits a lower corrosion current density and passivation current density than the CG state. The surface corroded morphologies of UFG Ti-6Al-7Nb revealed many small and shallow corrosion pits due to rapid self-repairing of the oxide film; (3) Pitting corrosion can be found on all samples while the pitting points of the CG state sample are mainly located on the grain boundaries and further extended. The UFG sample shows a relative lower grain boundary corrosion due to a higher LAGB volume; (4) Higher content of TiO 2 and Nb 2 O 5 formed in the oxide film of UFG Ti-6Al-7Nb led to a denser passive film and thus ensured a better corrosion resistance.