Influence of Surface Nanocrystallization on Microstructure and Properties of Anodic Oxide Film on Pure Titanium

Abstract

The surface gradient nanocrystalline structure was fabricated by the surface mechanical attrition treatment (SMAT), and the influence of surface nanocrystallization on the surface microstructure and characteristic of passive film on pure titanium was studied. The anodic oxidation was conducted in 0.5 mol/L H₂SO₄ solution under the potentiostatic mode at 30 V for different durations to investigate the surface characteristic and corrosion resistance of the passive films by optical microscope, X-ray diffraction, Raman spectrometer, X-ray photoelectron spectroscopy, and electrochemical test. The results show that the gradient nanocrystalline structure can increase the film thickness, promote the crystallization, and improve the corrosion resistance of passive film. The influence mechanism of surface nanocrystallization on the corrosion resistance can be explained by the cation-anion-vacancy condensation.

Keywords

surface mechanical attrition treatment; gradient nanocrystalline structure; titanium; anodic oxidation; ceramic layer; corrosion resistance

Science Press

Titanium alloys are widely used in aerospace, biological engineering, chemical engineering, and marine engineering due to their good biocompatibility, high specific strength, and excellent corrosion resistance^[1] caused by the stable native oxide film^[2,3]. However, the native oxide film has an amorphous structure with the thickness of 4~6 nm^[4], which leads to inferior mechanical properties and potential damage at low sheer stress^[5,6]. Furthermore, the native oxide film contains multivalent Ti^[7] which is less thermodynamically stable than Ti⁴⁺ and can be destroyed easily under severe corrosion environment, resulting in the crevice corrosion^[8].

Various surface modification methods have been used to enhance the corrosion resistance and wear resistance of titanium alloys^[9,10]. The anodic oxidation is an important treatment method due to its low cost and easy operation^[11]. In general, the growth and crystallization of anodic titanium oxide films can be promoted by raising the applied voltage^[12], prolonging the anodizing time^[13], or enhancing the solution temperature^[14]. However, it is found that the anodic oxide film prepared by anodic oxidation treatment is partially crystallized and has lots of “flower-like” or “nodule-like” structures on the surface^[15].

The substrates also have great influence on the formation and crystallizing process of passive film on titanium alloys. Capek et al^[16] found that the anodic oxide films formed on the polished titanium surface are smoother and more compact than those on unpolished titanium substrates. The surface mechanical attrition treatment (SMAT) is an effective way to fabricate surface nanocrystallization on metallic materials. Tong et al^[17] found that SMAT can reduce the nitriding temperature of Fe. Lei et al^[18] fabricated a ceramic coating of 10 μm in thickness by micro arc oxidation (MAO) process after SMAT on 2024 Al alloy, and found that SMAT-MAO treatment has a positive influence on corrosion resistance of films. Ou et al^[19] fabricated the superhydrophobic film on NiTi alloy surface by SMAT and anodization, and found that SMAT increases the thickness of the passive film and improves the wetting uniformity of the NiTi surface.

In this research, the pure titanium was selected as the substrate. The influence of surface nanocrystallization on the passive films on pure titanium was investigated. The crystallization process, film thickness, chemical composition, corrosion resistance, and the breakdown process of passive films were discussed.

1 Experiment

The chemical composition of the pure titanium (TA2) is listed in Table 1. During SMAT, the specimen surface was strongly impacted in a short time using a high frequency system (50 kHz). Then the TA2 and SMATed TA2 specimens were cut into the size of 50 mm×10 mm×1 mm, and cleaned with acetone, alcohol, and deionized water under ultrasonic condition. The anodic oxidation was conducted in a two-electrode electrochemical cell. Each specimen was used as the anode and the graphite plate was used as the cathode. The anode and cathode were positioned face-to-face with 10 mm in dis-tance. During the anodic oxidation treatment in 0.5 mol/L H₂SO₄ solution, the potential was kept at 30 V for 3600 s. The TA2 and SMATed TA2 specimens after the final anodic oxidation were named as O-Ti₃₀ and SMAT-Ti₃₀ specimens, respectively.

Table 1  Chemical composition of TA2 (wt%)

C	Fe	O	N	H	Ti
0.02	0.10	0.15	0.02	0.0011	Bal.

The surface morphology was observed using scanning electron microscopy (SEM, LEO 1530Vp, Carl Zeiss AG, Germany) at an acceleration voltage of 15 kV. A Raman spectrometer (LabRAM HR Evolution, Horiba Group, Japan) equipped with an optical microscope (OM) was used to detect the crystallization of anodic oxide films with the excitation wavelength of 532 nm and the incident power of 10.4 mW. For the thickness measurement of anodic oxide films, X-ray photoelectron spectroscope (XPS, PHI5000 Versaprobe-II, Shimadzu Co., Ltd, Japan) analysis was performed using Al-Kα radiation. The surface was sputtered over an area of 4 mm×4 mm using the Ar⁺ ion beam of 3000 eV.

The corrosion resistance was evaluated via the electro-chemical test on the electrochemical workstation (CHI 760E, Shanghai Chenghua Instrument Co., Ltd, China) in 10wt% NaCl solution at 70 °C. A saturated calomel electrode (SCE) was used as the reference electrode, a platinum coil was used as the counter electrode, and the specimens were used as the working electrodes. The potentiodynamic polarization tests were initiated after immersion of the specimens for 1 h until a stable open circuit potential was obtained, and the scanning rate was 3 mV/s from -3000 mV to 4000 mV. The corrosion potential (E_corr) and corrosion current density (j_corr) were determined from the polarization curves by the Tafel extrapolation method. The anodic oxide films on the O-Ti₃₀ and SMAT-Ti₃₀ specimens were characterized via the Mott-Schottky analysis, during which the potential was scanned along the anode-to-cathode direction at a scanning interval of 50 mV. The alternating current (AC) signal with a frequency of 1000 Hz and a peak-to-peak magnitude of 10 mV was superimposed on the scanning potential, and the capacitance was measured at each applied potential.

2 Results and Discussion

2.1 Surface microstructure

Fig.1 shows OM images of original TA2 (Fig.1a) and cross-section microstructure of SMATed TA2 specimens (Fig.1b). Before processing, the TA2 specimen was annealed in vacuum at 873 K for 6 h. The microstructure of TA2 shows an equiaxed microstructure with the average grain size of 50 μm. During SMAT process, the severe plastic deformation region is formed in the surface, and there is no sharp interface between the deformation zone and the matrix. The deformation layer consists of two parts: the severely deformed surface layer of 100 µm in thickness and the subsurface layer with many deformation bands and twins. The closer the subsurface layer to the surface layer, the denser the microstructure.

Fig.1  OM images of TA2 (a) and cross-sectional SMATed TA2 (b) specimens; XRD patterns of TA2 and SMATed TA2 specimens (c)

Because TA2 specimen has a small amount of active slip at room temperature, the twinning plays an important role in deformation under dynamic loading^[20]. The similar phenom-ena can also be observed in the near-α titanium alloy^[21] and α-titanium^[22] after severe plastic deformation treatment.

Fig.1c shows XRD patterns of the original TA2 and SMATed TA2 specimens, indicating that the full width at half maximum (FWHM) of the diffraction peaks of SMATed TA2 specimen is larger than that of the original TA2 specimen, because of the grain refinement and the increasing micro-strain at atomic-level^[23]. The average crystallite size of SMATed TA2 specimens is about 35 nm, according to the calculation based on FWHM of Bragg diffraction peaks of (002), (101), (102), and (103) planes via Scherrer-Wilson equation^[18]. The Scherrer-Wilson equation is as follows^[24]:

D=kλ/βcosθ

(1)

where λ is the X-ray wavelength, β is the FWHM of diffraction peak, θ is the Bragg angle, D is the average crystallite size, and the constant k≈1. This result demonstrates that the grain size decreases to nano-scale through SMAT.

2.2 Chemical composition of anodic oxide films

As shown in Fig.2, the wide-range XPS spectra of the O-Ti₃₀ and SMAT-Ti₃₀ specimens have the peaks of Ti 2p, O 1s, and C 1s. Fig.3 shows the variation of atomic percentages of O and Ti with sputtering depth in the passive films formed on different specimens. Vesel et al^[25] found that the depth where the oxygen content is 50% of its maximum value indicates the thickness of the passive film. As shown in Fig.3, the thickness of the passive film of O-Ti₃₀ and SMAT-Ti₃₀ specimens is about 114 and 1160 nm, respectively.

Fig.2  XPS spectra of anodic oxide films of O-Ti₃₀ (a) and SMAT-Ti₃₀ (b) specimens

Fig.3  Relationship between element contents and sputtering depth in anodic oxide films of O-Ti₃₀ (a) and SMAT-Ti₃₀ (b) specimens

Fig.4 and Fig.5 show the high-resolution XPS spectra of Ti 2p_3/2 and O 1s at different depths of the passive films of O-Ti₃₀ and SMAT-Ti₃₀ specimens, respectively. It can be inferred that the Ti, Ti²⁺, Ti³⁺, Ti⁴⁺, O^2-, OH^-, and H₂O may exist in the passive films according to the chemical state of Ti and O elements, as shown in Table 2^[26,27].

Fig.4  High-resolution XPS spectra of Ti 2p_3/2 (a~e) and O 1s (f~j) at different depths of passive oxide films of O-Ti₃₀ specimens:

(a, f) 0 nm, (b, g) 52 nm, (c, h) 75 nm, (d, i) 143 nm, and (e, j) 196 nm

Fig.5  High-resolution XPS spectra of Ti 2p_3/2 (a~e) and O 1s (f~j) at different depths of passive oxide films of SMAT-Ti₃₀ specimens:

(a, f) 0 nm, (b, g) 252 nm, (c, h) 508 nm, (d, i) 1010 nm, and (e, j) 1252 nm

Table 2  Standard energies of different XPS peaks (eV)^[26,27]

Ti 2p3/2					O 1s
Ti	Ti2+	Ti3+	Ti4+		O2-	OH-	H2O
453.6	455.9	457.3	530.2		530.2	532.5	533.4

After deconvolution by Gaussian-Lorentzian functions, the Ti 2p_3/2 narrow-scan spectra reveal a major peak at 458.01 eV corresponding to Ti⁴⁺, as shown in Fig.4a. The O 1s narrow-scan spectra (Fig.4b) consist of three peaks corresponding to O^2-, OH^-, and adsorbed water. The O^2- may originate from TiO₂, and the OH^- and the adsorbed water may come from the hydrated titanium oxides^[6]. These peaks suggest that the pas-sive film surface of O-Ti₃₀ specimen consists of TiO₂, Ti(OH)₄, and TiO₂·nH₂O. The Ti 2p spectra at the depth of 52 nm result from the Ti⁴⁺ and Ti²⁺, and the O 1s spectra at the depth of 52 nm result from O^2- and OH^-. XPS analysis indicates that the anodic oxide film mainly consists of TiO₂, TiO, and Ti(OH)₄ in the depth range of 0~52 nm. At the depth of 75 nm, the Ti⁴⁺ peaks weaken whereas the Ti²⁺ peaks increase. In the depth from 75 nm to 143 nm, the anodic oxide film is a mixture of Ti₂O₃, TiO, and metallic Ti. At the depth of 196 nm, both Ti₂O₃ and TiO disappear and only the metallic Ti can be observed.

As shown in Fig.5, the Ti element mainly appears in the forms of TiO₂, Ti(OH)₄, and TiO₂·nH₂O in the outmost surface (0 nm). In the depth from 0 nm to 252 nm, the components in the oxide film are a mixture of TiO₂, TiO, and Ti(OH)₄. In the depth range of 252~508 nm, the components in the oxide film are a mixture of TiO₂ and Ti(OH)₄ with a small amount of metallic Ti. In the depth of 508~1010 nm, the components mainly consist of TiO₂, Ti₂O₃, and metallic Ti. At the depth of 1252 nm, both Ti-oxide and Ti-hydroxide disappear and only the metallic Ti can be observed.

The main components of anodic oxide film of both O-Ti₃₀ and SMAT-Ti₃₀ specimens are oxides at high valence and hydroxide at the outer part of oxide film, and metallic Ti and oxides at low valence at the inner part of oxide film. The nanocrystalline surface prepared by SMAT greatly increases the thickness of the oxide film from 114 nm to 1160 nm, and the range of oxides at high valence also increases. During the anodic oxidation, the lattice diffusion is not significant owing to its considerably small temperature coefficient, and the grain boundary diffusion plays a significant role^[5]. Due to the numerous grain boundaries in the severe plastic deformation region, the adsorption and diffusion of oxygen are faster for the SMATed TA2 specimen than the original TA2 specimen.

2.3 Crystallization

The OM images and Raman spectra of the passive films of O-Ti₃₀ and SMAT-Ti₃₀ specimens are shown in Fig.6. The anodized film on original TA2 (Fig.6a) has two different regions: a bright region and a dark region, indicating that the crystallization behavior of the anodized film is not uniform. On the contrary, the microstructure of SMAT-Ti₃₀ specimen shows a relatively homogeneous feature (Fig.6b), indicating a relatively uniform crystallization.

Fig.6  OM images of passive films of O-Ti₃₀ (a) and SMAT-Ti₃₀ (b) specimens; Raman spectra of anodic oxide films of O-Ti₃₀ and SMAT-Ti₃₀ specimens (c)

In Fig.6c, the Raman spectra consist of four peaks at around 144, 405, 516, and 639 cm^-1, which are related to the long-range order and short-range order of anatase phase^[28]. In detail, the Raman band at around 144 cm^-1 indicates the long-range order of anatase phase, and Raman bands at about 405, 516, and 639 cm^-1 indicate the short-range order of anatase phase^[29]. When the Raman micro-laser beam is focused on the dark region of the passive film of O-Ti₃₀ specimen, a weak peak can be observed at around 144 cm^-1. However, there is no peak when the beam is centered on the light region.

Xing et al^[13] suggested that the intensity of Raman bands is directly proportional to the crystallinity of the passive film. Therefore, the peak intensity of SMAT-T₃₀ specimen is generally stronger than that of O-Ti₃₀ specimen, indicating that the crystallization of passive films is promoted by the surface nanocrystallization.

2.4 Electrochemical tests

The potentiodynamic polarization curves of O-Ti₃₀ and SMAT-Ti₃₀ specimens in 10wt% NaCl solution at 70 °C are shown in Fig.7a. The anodic part of the polarization curves indicates that both the O-Ti₃₀ and SMAT-Ti₃₀ specimens show the passivation characteristics. In the passive region of the SMAT-Ti₃₀ specimen, the current density barely changes. However, the current density of the O-Ti₃₀ specimen is increased with increasing the applied corrosion potential, which may be attributed to the dissolution of the oxide film in 10wt% NaCl solution at 70 °C.

Fig.7  Potentiodynamic polarization curves (a) and Mott-Schottky plots (b) of O-Ti₃₀ and SMAT-Ti₃₀ specimens in 10wt% NaCl solution at 70 °C

Based on the potentiodynamic polarization curves, the corrosion potential E_corr and corrosion current density j_corr can be obtained. The j_corr can be obtained by extrapolating the cathodic Tafel line to the E_corr point^[30,31]. Thus, the value of E_corr is -1.12 and -0.68 V vs. SCE and corresponding j_corr is 5.01×10^-5 and 1.25×10^-5 A·cm^-2 for the O-Ti₃₀ and SMAT-Ti₃₀ specimens, respectively. The SMAT-T₃₀ specimen has higher E_corr and lower j_corr, compared with the O-Ti₃₀ specimen, which indicates the less tendency to corrosion and slower corrosion rate. Thus, the SMAT-Ti₃₀ specimen exhibits a better corrosion resistance^[32].

It is well known that a relationship exists between the semiconducting properties and corrosion resistance of an anodic oxide film. These properties can be determined by the Mott-Schottky plot with capacitance as a function of electrode potential, which reflects the charge distribution in the passive film. According to Mott-Schottky analysis, two capacitances should be considered: the capacitance of space charge layer and the capacitance of Helmholtz layer^[33]. These capacitances are in series and the total capacitance C is expressed as follows:

$$\frac{\mathrm{1}}{C^{\mathrm{2}}}=\frac{\mathrm{1}}{C_{\mathrm{S}\mathrm{C}}^{\mathrm{2}}}+\frac{\mathrm{1}}{C_{\mathrm{H}}^{\mathrm{2}}}$$

(2)

where C_SC is the capacitance of the space charge layer and C_H is the capacitance of the Helmholtz layer. Under high frequency (>1 kHz), the capacitance of space charge layer is significantly smaller than that of Helmholtz layer. Therefore, the contribution of the capacitance of Helmholtz layer to the total capacitance can be neglected, and the capacitance of space charge layer can be considered as the total capacitance^[34]. The charge distribution at the semiconductor/electrolyte interface can be typically determined from the Mott-Schottky relations which describe the dependence of C_SC on electrode potential E^[35], as follows:

$$\frac{\mathrm{1}}{C_{\mathrm{S}\mathrm{C}}^{\mathrm{2}}}=\frac{\mathrm{2}}{\epsilon {\epsilon }_{\mathrm{0}}e{N}_{\mathrm{A}}}(E-E_{\mathrm{f}\mathrm{b}}-\frac{kT}{e})$$ for n-type semiconductor

(3)

$$\frac{\mathrm{1}}{C_{\mathrm{S}\mathrm{C}}^{\mathrm{2}}}=-\frac{\mathrm{2}}{\epsilon {\epsilon }_{\mathrm{0}}e{N}_{\mathrm{D}}}(E-E_{\mathrm{f}\mathrm{b}}-\frac{kT}{e})$$ for p-type semiconductor

(4)

where E is the applied potential; E_fb is the flat band potential; ε₀ is the permittivity of free space; ε is the relative dielectric constant; N_A is the donor density for n-type semiconductors; N_D is the acceptor density for p-type semiconductors; k is the Boltzmann constant; T is the absolute temperature; e is the charge of the electron. N_D/N_A and E_fb can be determined from the slope of the experimentally obtained 1/C²-E plots and extrapolation result of the linear portion to 1/C²=0, respectively. The shape of the Mott-Schottky plot can infer the conductivity type of the semiconductor, i.e., the negative and positive slopes correspond to the p-type and n-type semiconductors, respectively. Furthermore, the concentration of free charge carriers can be determined by the slope of the 1/C_SC²-E plot; E_fb can be determined by the intercept with the potential axis (1/C_SC²=0) of the 1/C_SC²-E plot^[36].

In order to understand the semiconductive properties of the anodic oxide films of O-Ti₃₀ and SMAT-Ti₃₀ specimens, Fig.7b shows the Mott-Schottky plots of specimens passivated under the open circuit potential. A significantly lower donor density N_D (3.5×10¹⁹ cm^-3) can be obtained for the SMAT-Ti₃₀ specimen, compared with that of the O-Ti₃₀ specimen (5.6×10²¹ cm^-3). Since N_D represents the affinity of aggressive anions to the anodic oxide film, the O-Ti₃₀ specimens with higher N_D present a larger sensitivity to corrosion with an easier incorporation of aggressive anions to the anodic oxide film. This result is coincident with the lower breakdown potential in the potentiodynamic curves for O-Ti₃₀ specimens.

Point defects transfer the electrical current through the anodic oxide film and then the defect reactions occur at the metal|film (M|f) and film|solution (f|s) interfaces, which ultimately causes the breakdown of the anodic oxide film^[37]. The schematic diagram of point defect model (PDM) and related reactions used to describe the breakdown process of oxide films on Ti surface is shown in Fig.8.

Fig.8  Schematic diagram of breakdown process of anodic oxide film based on point defect model

The equations are expressed in Kröger-Vink notation^[38], where the subscripts O and Ti denote the structural oxygen site and titanium site, respectively; the superscript “·” or “′” denotes one positive or one negative charge, respectively; V represents a vacancy. The electrolyte solution consists of Ti⁴⁺ (aq) hydrated metal cations.

According to PDM^[39], the defects are generated and disappear at the Ti|f and f|s interfaces. The reactions Ti+$V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$→Ti_Ti and Ti→Ti_Ti+2$V_{\mathrm{O}}^{··}$ occur at the Ti|f interface, which expresses the submergence of cation vacancies ($V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$) into the Ti lattice and the generation of anion vacancies, respectively. The Ti_Ti is formed in the oxide lattice and a vacancy $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ is generated in the Ti lattice. Then the atom Ti reacts with $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ in the Ti lattice to generate oxygen vacancies ($V_{\mathrm{O}}^{··}$) and Ti_Ti in the oxide lattice.

The reactions Ti_Ti+($V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$$V_{\mathrm{O}}^{··}$)′′→$V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$+(Ti_Ti2$V_{\mathrm{O}}^{··}$) and O_O←$V_{\mathrm{O}}^{··}$+H₂O→2H⁺_aq occur at the f|s interface. The former reaction leads to the formation of cation vacancy $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ and hydrated titanium complex Ti(H₂O)_n⁴⁺. Then the Ti(H₂O)_n⁴⁺ ions leave the oxide and therefore metal cations Ti_Ti from the oxide lattice are transferred into the solution. The latter reaction indicates the absorption of O^2- from H₂O into $V_{\mathrm{O}}^{··}$, resulting in the oxygen ions O_O into the oxide lattice. When the specimen is immersed in the solution, there was a potential difference due to the difference between the metal phase and the solution phase. The potential difference is distributed across the interphases of M|f (ϕ_M|f) and f|s (ϕ_f|s)^[38]. $V_{\mathrm{O}}^{··}$ and $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ are diffused through the oxide film by electric field and concentration gradient. The O_O ions are transferred through $V_{\mathrm{O}}^{··}$ towards the Ti|f and f|s interfaces simultaneously. $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ is transferred towards the Ti|f interface and it disappears at the Ti side. The cation-vacancy disappearance leads to $V_{\mathrm{O}}^{··}$/$V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ interactions, which are driven by the electrostatic attraction between opposite-charged defects of high contents^[40].

During the reaction of Cl^-·nH₂O→nH₂O+Cl^-, Cl^- is formed via the diffusion and migration of Cl^- through the f|s interface. Some Cl^- anions may be absorbed by oxygen vacancies, leading to the formation of Cl defects ($\mathrm{C}{\mathrm{l}}_{\mathrm{O}}^{·}$) which reacts with Ti_Ti, thereby resulting in the generation of Ti_Ti$\mathrm{C}{\mathrm{l}}_{\mathrm{O}}^{·}$. Subsequently, Ti_Ti$\mathrm{C}{\mathrm{l}}_{\mathrm{O}}^{·}$ extraction occurs at f|s interface, resulting in the generation of Cl^-, Ti⁴⁺, and a cation vacancy/oxygen vacancy pair ($V_{\mathrm{O}}^{··}$$V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$), which corresponds to the passive film dissolution. Moreover, the reaction between Cl^- and $V_{\mathrm{O}}^{··}$ reduces the concentration of $V_{\mathrm{O}}^{··}$, which further leads to the accumulation of $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$.

As the excess $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ is accumulated at the Ti|f interface, the $V_{\mathrm{T}\mathrm{i}}^{\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}\mathrm{\text{'}}}$ condensate is formed, resulting in the separation of the oxide layer from the metal substrate. The breakdown may occur when the average areal concentration of vacancies exceeds the critical value (7.17×10¹⁵ cm^-2)^[41]. Therefore, N_D is increased with increasing the affinity of Cl^-. In general, SMAT reduces the donor density of the surface layer, and therefore SMAT-Ti₃₀ specimen possesses better corrosion resistance. The breakdown of the oxide film is dominated by the cation-anion-vacancy condensation mechanism.

3 Conclusions

1) The surface nanocrystallization has a positive effect on the crystallization of the anodic oxide film, which also increases the thickness of the anodic oxide film. The higher corrosion potential and smaller corrosion current density can be achieved with increasing the anodic potential, therefore obtaining the passive film with better corrosion resistance. A significantly smaller donor density for n-type semiconductors of 3.5×10¹⁹ cm^-3 is achieved for the Ti alloys after surface nanocrystallization with passive film.

2) The surface mechanical attrition treatment can reduce the donor density of the anodic oxide film.

3) The breakdown of the oxide film is dominated by the cation-anion-vacancy condensation mechanism.