Electrochemical corrosion characteristics and biocompatibility of nanostructured titanium for implants

Applied Surface Science 434 (2018) 63–72

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Full Length Article

Electrochemical corrosion characteristics and biocompatibility of nanostructured titanium for implants Jinwen Lu a , Yong Zhang b , Wangtu Huo a , Wei Zhang a , Yongqing Zhao a , Yusheng Zhang a,∗ a

Advanced Materials Research Center, Northwest Institute for Nonferrous Metal Research, Xi’an 710016, China Herbert Gleiter Institute of Nanoscience, School of Materials Science and Engineering, Nanjing University of Science and Technology, Nanjing 210094, China b

a r t i c l e

i n f o

Article history: Received 7 April 2017 Received in revised form 21 October 2017 Accepted 24 October 2017 Keywords: Titanium Nanocrystalline Corrosion behavior Biocompatibility Implant materials

a b s t r a c t In the present study, a nano-grained (NG) surface layer on a commercial pure (Grade-2) titanium sheet was achieved by means of sliding friction treatment. The surface characteristics, in vitro corrosion behavior and biocompatibility of NG Ti were investigated, compared with those of the conventional coarse-grained (CG) substrate. The protective passive film on NG Ti surface is thicker than that on CG Ti, leading to its enhanced biological corrosion resistance in simulated body fluid (SBF) solution. In addition, NG Ti shows a much higher hydrophilicity and nano-roughness, which is related to its significantly improved cell attachment, spreading, proliferation and maturation relative to CG Ti. The enhanced biological anti-corrosion properties and biocompatibility render NG Ti a promising biomaterial for implants. © 2017 Elsevier B.V. All rights reserved.

1. Introduction Titanium and titanium alloys are employed extensively for orthopaedic and dental implant fabrication [1–3]. However, concerns about possible biotoxicity of alloying elements in the long-term performance, such as Al and V in the Ti-6Al–4 V alloy commercially applied, have been driving the development of purity titanium as an alternative to alloys because of its excellent biocompatibility [4,5]. Commercially pure Ti would take a prominent place among its alloys if the loss of strength due to lack of alloying elements are compensated [3,5,6]. Grain refinement is one feasible way. Excitingly, a number of reports revealed that the combination performances (mechanical property, corrosion resistance and biocompatibility) of Ti can be tailored through grain refinement by severe plastic deformation (SPD) techniques such as equal channel angular pressing [5,7], high pressure torsion [8] and groove pressing [9]. The corrosion behavior of Ti alloy is mainly dependent on the passive film properties (such as thickness and layer mode) and the microstructure of the substrate (grain size and dislocation density) [3,10,11]. Numerous results suggest that nanocrystallization does not change the single structure of the passive film but improves its

∗ Corresponding author at: Advanced Materials Research Center, Northwest Institute for Nonferrous Metal Research, Xi’an 710016, China. E-mail address: [email protected] (Y. Zhang). https://doi.org/10.1016/j.apsusc.2017.10.168 0169-4332/© 2017 Elsevier B.V. All rights reserved.

electrochemical stability [6,9,12], which was mainly attributed to a Ti and O ion rich layer at the passive film/metal interface [13]. Moreover, some studies [6,10] have reported the nanocrystalline structure could induce a quickly formed passive layer, which obviously enhanced the corrosion resistance and reduced ion release in corrosion medium. While further work is still needed to deeply understand the formation process of passive film with respect to nanocrystallization effect in Ti alloys. The biocompatibility of nanostructured metal is dominated by both the surface architecture and microstructure, including grain size distribution, grain boundary misorientation, and so on, which are related to manufacturing processes. The work by Truong [2] and Estrin [14] showed that the enhanced attachment and spreading of human bone marrow-derived mesenchymal stem cells (hMSCs) in the initial stages of culture was caused by increased surface roughness on nanoscale. Huang [15] and Wen’s [16] revealed that the significantly improvement of osteoblast adhesion and proliferation on nanocrystalline Ti fabricated by surface mechanical attrition treatment can be attributed to decreased grain size and increased hydrophilicity. Shri et al. [17] demonstrated that the increased protein adsorption and cell growth was remarkably improved by deformation-induced martensites. Sliding friction treatment (SFT) is a new developed SPD method that can generate nano-grains on metal surface [18,19]. The objective of the present study is to evaluate the suitability of a NG surface layer on Ti prepared by SFT for implant application with respect to in vitro corrosion behavior and biocompatibility, and CG Ti is also

64

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

investigated as control. An excellent combination of in vitro biological anti-corrosion and biocompatibility properties was achieved by introducing a NG surface layer on pure Ti and the beneficial effect of the NG layer are also discussed. 2. Experimental 2.1. Material preparation and microstructure characterization A coarse-grained (CG) commercially pure (ASTM Grade-2) Ti sheet with 3 mm thickness was used in this study. The schematic illustration of SFT process can be seen in Ref. [18]. In this work, the processing parameters for SFT were selected as follows: 500 N in load, 0.2 m/s in sliding velocity, 50 mm in amplitude, and 100 in cycle. In order to eliminate the influence of surface pollution during SFT, the top surface layer of 1–3 ␮m was removed by polishing carefully. The microstructure of NG Ti and CG Ti was characterized by a transmission electron microscope (TEM, JEOL, JEM-2100) operated at a voltage of 200 kV and an optical microscopy (OM, Leica MPS 30), respectively. For TEM observation, cross-sectional thin foil samples were cut from the samples and thinned by ion thinning at low temperatures. The grain size was calculated and estimated by a combination of linear intercept method and TEM or OM observation. 2.2. Electrochemical corrosion measurement All the electrochemical experiments were conducted using a IM6 Zahner-electrik Gmbh (Zenniom, Germany) electrochemical workstation in SBF solution at 37 ± 1 ◦ C controlled by thermostatic water bath. A traditional three-electrode system was adopted using a saturated calomel electrode (SCE) as reference electrode and a platinum wire as counter electrode, and the measured sample with a certain exposed area (10mm × 10 mm) was used as working electrode. The detailed decomposition of SBF solution can be found in Ref. [20]. Prior to all electrochemical measurements, the specimens were initially reduced potentiostatically at −1.0 V for 300 s to remove air-formed oxides on the surface and then kept in solution until a stable corrosion potential was reached. The open circuit potential (OCP) measurement was carried out for 30 min starting from the electrode immersing into the electrolyte. The electrochemical impedance spectroscopy (EIS) measurements were carried out under potentiostatic condition at OCP with 10 mV amplitude AC voltage signal, and the applied frequency range was from 105 Hz down to 10−2 Hz. The impedance date was analyzed using ZSimpWin 3.0 software. The potentiodynamic polarization (PDP) curves were obtained after 30 min immersion in SBF solution, in the range from −1 V to 2 V using a scan rate of 0.5 mV/s. The corrosion potential (Ecorr ) and corrosion current density (jcorr ) were determined by Tafel slope extrapolation, and the passive current density (jpp ) was obtained from the passive zone where the corrosion current remained approximately constant. Mott-Schottky experiments were performed in the potential range of −0.5–1 V (vs. SCE) with the applied frequency of 1 kHz. For electrochemical experiments, the SBF solution of 500 ml was changed for each tested sample, and all electrochemical tests were repeated three times as per the proposed ASTM standard to ensure reproducibility and statistically analyzed to gain the standard deviations. After electrochemical measurement, the surface morphologies were observed by a JSM-6460 scanning electron microscope (SEM), and the chemical composition of the surface oxide film on NG and CG Ti was analyzed by X-ray photoelectron spectrometer (XPS, ESCAKAB-250Xi, USA).

2.3. Surface characterization Before biocompatibility experiments, surface topography and roughness were observed by atomic force microscopy (AFM, Dimension, Icon) with the contact mode at a rate of 1 Hz. In order to obtain further information of the surface wettability of the NG and CG Ti, the contact angles of deionized water droplet (0.5 ␮l) were measured using an automatic contact angle meter (KRUSS Gmbh, DSA10-Mk2) at room temperature. Three samples from each group were measured and two measurements were performed on each sample to evaluate the average contact angle. The surface energies of the NG and CG Ti were also calculated based on contact angle measurements, and the method for surface energy calculation can be seen in Ref. [21]. 2.4. Investigations of biocompatibility 2.4.1. Protein adsorption assay For the protein adsorption assay, bovine serum albumin (FBS, Thermo Science, USA) was used as a standardized model protein, and 1 ml droplet of Dulbecco’s modified Eagle/Ham’s F121:1 (DMEM) medium (Thermo Scientific, USA) containing 10% fetal bovine serum albumin was pipeted onto the samples surface placed in 24-well plate. After incubation for 1, 4, and 24 h at 37 ◦ C, the samples were transferred to new 24-well plates and washed three times with phosphate buffer saline (PBS, Sigma, USA). Afterwards, 500 ␮l of 1% sodium dodecyl sulfate (SDS) solution was added to these wells and shaken for 15 min to detach proteins from the samples surface. The total protein concentrations in the collected SDS solutions were determined using NanoDrop 2000C device (Thermo Scientific, USA) at wavelength of 280 nm. 2.4.2. Cell culture Cell culture experiments were performed using human fetal osteoblast cell line (hFOB1.19) provided by Institute of Biochemistry and Cell Biology of Chinese Academy of Sciences (Shanghai, China). The sheet samples were placed in 24-well culture plates and incubated into DMEM medium supplemented with 10% FBS, 0.3 mg/ml Geneticine418 (Sigma, USA), 0.5 mM sodium-pyruvate (Sigma, USA) and 1.2 g/l Na2 CO3 , and incubated in a humidified atmosphere incubator with 5% CO2 at 37 ◦ C, and the complete medium was refreshed every 2 days. 2.4.3. Cell cytotoxicity Cytotoxicity was evaluated on cells cultured in medium conditioned by the presence of implants. The lactate dehydrogenase (LDH) activity in the culture media was used as an index of cytotoxicity. The LDH activity was determined spectrophotometrically according to the manufacturer’s instructions. 2.4.4. Cell adhesion and proliferation assessment The cells (hFOB1.19) were seeded on the sample substrates with a density of 8 × 104 cells/ml in 24-well tissue culture plate. The adhesion of cells was assessed after 1, 4 and 24 h of incubation in 24well plates. At the end of each time period, the complete medium was removed from each well, and the samples were washed three times with PBS then transferred to new 24-well plates. The cells adhered on the samples were subsequently digested out of the samples with 0.3 ml 0.25% trypsin (Sigma, USA) for 5 min, then 0.7 ml complete medium was added to stop digestion. The released cells were counted with a hemocytometer on a Nikon Eclipse inverted fluorescence microscope. In order to study the proliferation of the hFOB1.19 on the samples, an initial density of 104 cells/ml was seeded on the surface of each substrate and incubated for 3, 7 and 14 days.

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

2.4.5. Cell morphology observation Cell density and morphological characteristics of hFOB1.19 on the sample substrates were studies using a JSM-6460 SEM. The cells were cultured the substrates for 3 days, washed three times with PBS and fixed with 2.5% glutaraldehyde for 1 h at 4 ◦ C. The fixed cell samples were dehydrated with graded series of ethanol, and then vacuum-dried. After gold-palladium coating, the samples were observed under a FESEM (JEOL JSM-6700F, Japan) to characterize the cell morphology. 2.4.6. Alkaline phosphatase (ALP) activity To assay for ALP activity, cells were cultured on sheets for 3, 7 and 14 days at an initial seeding density of 5 × 104 cells well. After the end of each time period, the cells seeded on the surface were washed three times with PBS, and transferred to a new cell culture plates, then lysed in 0.1vol.% Triton X-100 using five standard freeze-thaw cycles, and then shaken for 10 min. The intracellular ALP activity was determined by using commercially available human ELISA kits (R&D, USA). The optical absorbance at 450 nm was recorded spectrophotometrically. 2.5. Statistical analysis Statistical analysis was performed using the independent T-test to evaluate the differences between groups. The significance level of p < 0.05 was considered to be statistically significant difference. 3. Results 3.1. Microstructure and surface characterization TEM bright-field micrograph combined with the selected area electron diffraction pattern in Fig. 1(d) reveal that a NG Ti surface layer with the average grain size of about 90 nm has been success-

65

fully obtained by SFT, while the average grain size of CG Ti is around 42 ␮m (Fig. 1(a)). NG Ti exhibits significantly higher nano-roughness than CG Ti (Fig. 1(b, c, e and f)), and the corresponding roughness analyses are presented in Table 1. Three conventional parameters (the average roughness (Ra ), the root mean square roughness (Rq ) and the maximum roughness (Rmax )) are used to evaluate topography characterization. The NG Ti surface exhibits average Rq and Ra values of approximately 6.15 nm and 4.23 nm, while those values for CG Ti are 3.80 nm and 2.16 nm, respectively. Moreover, the NG and CG Ti samples differ significantly in Rmax (166.00 nm for NG Ti and 94.30 nm for CG Ti). The measured contact angles and the calculated surface free energy for NG and CG Ti are also summarized in Table 1, which indicates that NG possesses a higher hydrophobicity compared with CG Ti.

3.2. Electrochemical corrosion 3.2.1. Open circuit potential measurements Fig. 2(a) presents OCP curves of NG and CG Ti with an increasing immersion time in SBF solution at 37 ◦ C. The shift of OCP in the positive direction indicates the formation of a passive film, and the shift in the negative direction may be the consequence of local breakdown or dissolution of the film or no film formation [22]. It can be seen that both curves go up sharply at the beginning, which is associated with the formation of the protective passive film [23]. Then the curves become gradually stable after a certain period, indicating the stabilization of the passive film. Many research findings revealed that this stable OCP value was related to the transformation from TiO or Ti2 O3 to TiO2 on the electrode/electrolyte interface [24,25]. However, the Eocp of NG Ti increases more rapidly with the time at the first 100 s, which implies a more quickly formed passive film, and then keeps a relatively higher stable value.

Fig. 1. Metallographic observation of CG Ti (a) and a bright-field TEM image of NG Ti (d), AFM topographies and the corresponding height profiles along the horizontal scanning lines of NG Ti (b, c) and CG Ti (e, f). The CG Ti has an average grain size of 42 ± 5 ␮m, and the size of NG Ti is about 90 ± 5 nm.

66

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

Table 1 Comparative evaluationof grain size, surface roughness, contact angle and surface free energy for the NG and CG Ti (n = 3). Samples

Average grain size

Rq /nm

Ra /nm

Rmax /nm

Contact angle/degrees

Surface free energy/mJ/m2

NG Ti CG Ti

90 ± 5 nm 42 ± 5 ␮m

6.15 ± 0.08 3.80 ± 0.12

4.23 ± 0.10 2.61 ± 0.05

166.00 ± 1.26 94.30 ± 0.62

64.65 ± 0.87 71.52 ± 1.18

31.17 ± 0.65 23.08 ± 0.42

Fig. 2. Open-circuit potential (EOCP ) as a function of time (a), potentiodynamic polarization curves (b) for NG and CG Ti in SBF solutions at 37 ◦ C. Typical EIS experimental data and simulated curves by the ZsimpWin software: Bode diagram (c), Nyquist diagram and equivalent electrical circuit model (inset) (d) for NG and CG Ti.

3.2.2. Potentiodynamic polarization curves Fig. 2(b) shows that both NG and CG samples display similar passive behavior and typical active-passive transition. The two samples present a rapidly increasing current density with the corrosion potential before entering a passive region. The passive region is comparable for both samples (0.5–2 VSCE for NG Ti and 0.2–2 VSCE for CG Ti), while NG Ti possess a much higher corrosion potential (-0.56 V) than its CG form (-0.73 V). This indicates passivation characteristics of the spontaneously developed oxide film upon immersion in the test electrolyte [10,22]. For better understanding the polarization behavior of NG Ti, the corrosion parameters (including Ecorr , jcorr , and passive current density (jpp )) are calculated from the PDP curves, and the results are shown in Table 2. The corrosion potentials (Ecorr ) of NG and CG Ti are −0.5648 V and −0.7337 V, respectively. The results show that NG Ti has a notably lower jcorr (0.548 ␮A/cm2 ) than CG Ti (1.160 ␮A/cm2 ). When the current density reaches 50.32 ␮A/cm2 (jpp ), NG sample enters the passivation region with an almost constant current density during relatively wider potential range from 0.70 to 1.3 V. As the potential

is increased to about 1.50 V, a slight decrease in the current density is observed, which may be related to the local breakdown or dissolution of the oxide film. 3.2.3. Electrochemical impedance spectroscopy tests The EIS plots and simulated curves of NG and CG Ti are presented in Fig. 2(c, d) with the form of Bode and Nyquist plots. NG Ti exhibits much higher module of impedance (Z) and phase angle () than CG Ti, indicating its nobler electrochemical performance. As shown as in Fig. 2(c), Z in the high frequency region is almost frequency independent due to  nearly approaching 0◦ , which indicates a representative response of solution resistance. In the intermediate frequency region, the relationship of the modulus and the frequency presents a linear relation with a slope close to −1. This is a typical capacitive performance of the electrode/electrolyte interface. In the low frequency region, the phase angles for both samples shift to a value between −75◦ and −55◦ , displaying a decreased capacitive influence in the electrochemical behavior of the electrode [22]. In addition, the Nyquist plots (Fig. 2(c)) of both samples

Table 2 Corrosion parameters of potentiodynamic polarization of the NG and CG Ti in SBF solutions at 37 ◦ C. Samples

ba /V/dec

bc /V/dec

Ecorr /V

jcorr /␮A/cm2

Rp /k·cm2

NG Ti CG Ti

−0.234 ± 0.008 −0.167 ± 0.012

0.441 ± 0.014 0.5370 ± 0.019

−0.565 ± 0.015 −0.734 ± 0.023

0.548 ± 0.084 1.160 ± 0.125

121.580 ± 3.870 47.680 ± 1.560

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

67

Table 3 Electrochemical impedance parameters of the NG and CG Ti in SBF solution at 37 ◦ C. Samples

Rs /·cm2

Rb /k·cm2

Qb /(␮F/cm2 )

n

Chi-square

NG Ti CG Ti

33.05 ± 0.05 29.47 ± 0.07

394.40 ± 0.36 138.50 ± 0.12

26.29 ± 0.28 28.21 ± 0.06

0.88 ± 0.08 0.81 ± 0.05

0.00025 0.00028

Fig. 3. Current vs. time plots (a) and Mott-schottky plots (b) for NG and CG Ti in SBF solutions at 37 ◦ C, and SEM micrographs showing the surface morphologies of NG (c) and CG (d) Ti after potentiodynamic polarization.

exhibit only one capacitance loop with linear-like rather than arclike behavior over the frequency range, which suggests that the growing up of passive film delays the charge transfer processes to lower frequencies [26]. In order to obtain detailed information for the passive films, a single time constant model as shown in the inset of Fig. 2(d) is used to fit the EIS results, and the specific values are presented in Table 3. Accordingly, the meanings of circuit elements are as follows: Rs represents the electrolyte resistance; Rb and CPE represent the resistance and constant phase element of passive film. CPE replaced the capacitor to account for the non-ideal behavior of the capacitive elements due to different physical phenomena such as surface heterogeneity which results from surface roughness, impurities, dislocations or grain boundaries [27]. Besides Z and , the NG Ti also exhibits lower Qb than CG sample, which indicates the formation of a highly stable passive film in SBF solution. The barrier resistance Rb is strongly dependent on the passive film characteristic and a higher Rb value for NG Ti implies its good resistance. Moreover, Rb is much higher than Rs by a factor of around 5 × 103 , indicating that the protection for NG and CG Ti is predominantly provided by the surface passive film [28]. As shown above, it is evident that the absolute values of Z at low frequency, the max phase angle, as well as Rb of NG sample are all higher than CG sample,

which implies again that the corrosion resistance of NG surfaces is significantly enhanced than that of CG surface. 3.2.4. Potentiostatic polarization test The relationship of current density (I) with time (t) can be expressed as the follow formula, I = 10−(A+klgt) [29], where A is the constant and k is the slope of the double-log plot for potentiostatic polarization (PP) (Fig. 3(a)). Each curve for NG and CG samples is fitted by two lines with the slopes of −0.78, −1.07 and −0.68, −0.91, respectively. It is well known that k → −1 indicates the formation of a compact passive film, while k → −0.5 means the presence of a porous film as a result of a dissolution and precipitation process [10]. Therefore, it can be drawn that a compact and well protective passive film is quickly formed on NG Ti surface. But for CG sample, a porous passive film is formed firstly, and then become dense with the immersion time. It is also well accordant with the EIS results in Fig. 2(c). 3.2.5. Mott-Schottky analysis The passive film formed on the titanium alloy surface can be considered as a typical semiconductor [10,30]. The Mott-Schottky (MS) plots are employed to analyze the structural changes of the passive film, as shown in Fig. 3(b). It is found that the potential E and C−2 (C is the capacitance of the space charge layer) have a

68

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

clear linear relationship with a positive slope in the range 0.4-1.0 V. This demonstrates that both passive films formed on NG and CG surface belong to n-type semiconductors, in which the dominant crystallographic defects are donors (oxygen vacancies and titanium interstitials) [11,31]. The donor density is calculated from the slopes in the MS plots. From the slope of the linearly fitted plot in Fig. 3(b), the donor density ND can be calculated as 1.13 × 1020 cm−3 and 1.39 × 1020 cm−3 for NG and CG Ti, respectively. The donor concentrations are directly connected with ionic and electronic conductivity characteristics of the passive layer. A lower ND value for NG Ti implies the formation of less defective and highly protective passive film and a higher ND value for CG Ti is a strong indication of a nonstiochiometric or highly disordered passive film formed in electrolyte. For NG Ti, it is considered that the presence of abundant nanograin boundaries can generate more oxygen vacancies for the injection of oxygen ion, and thus increasing the net donor density. Moreover, the thickness of the space charge layer can be estimated by the equation available in the Ref. [29] for n-type semiconductor to be 14.56 nm and 8.17 nm for NG and CG Ti respectively, which is directly proportional to that of the passive film. The pronounced enhancement in thickness of the space-charge layer for NG implies its lower chances of passive layer breakdown and pitting initiation. Therefore, the passive film formed on NG Ti has a more stable structure with a stronger protection against corrosion, in compared with CG Ti.

suggested to be covered by the compact passive layer without obvious corrosion pitting, which agrees well with the k values measured for NG and CG Ti (Fig. 3(a)). Fig. 4 shows the XPS spectra of Ti 2p and O 1 s obtained from the surface passive film on NG and CG Ti after polarization measurement at different depths. The outermost layer is identified as Ti4+ state with the corresponding Ti 2P3/2 peak at 459.1 eV, being ideal oxides of TiO2 for both NG and CG Ti. It is reasonable because TiO2 has the highest stability due to thermodynamic reasons (The free energy of formation G (298 K) for TiO and TiO2 is −495 kJ/mol and −889.5 kJ/mol, respectively [32]). With an increasing sputtering time, the peak shifts toward the low binding energy and the peak width of Ti 2p is broadened. It is indicated the emergence of abundant suboxides beneath the outer TiO2 layer, which are Ti2 O3 and TiO as the oxidation state of Ti decreased gradually from Ti4+ through Ti3+ and Ti2+ . When the sputtering time is increased to 60 s, some spectra of Ti 2p for NG Ti are still retained, while only some suboxides of Ti are detected for CG Ti. At sputtering time exceeding 200 s, the spectra of suboxides are detected for NG Ti, while CG Ti only exhibits peaks of pure Ti. Therefore, the main composition of the passive layers for NG and CG Ti samples is TiO2 and the suboxides of Ti, and the thickness of passive layer on the former is much thicker than that on the latter, consistent with electrochemical results.

3.2.6. Surface characterization of the passive film Fig. 3(c, d) shows the typical SEM surface morphologies of NG and CG samples after polarization measurements. Both samples are

3.3.1. Protein adsorption on the surfaces The amounts of total protein adsorbed on NG and CG Ti surfaces from DMEM medium containing serum after 1, 4 and 24 h

3.3. Evaluation of biological compatibility

Fig. 4. Typical and enlarged XPS spectra of NG and CG Ti after polarization measurement with different sputtering depths: (a) and (c) Ti 2p, (b) and (d) O 1s.

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

incubation are displayed in Fig. 5(a). It is found that the adsorption amounts of total protein on NG and CG Ti surface is increased with an increasing incubation time. Surface nanocrystallization significantly promotes the adsorption of proteins, especially at long-time incubation, where a sharply enhancement in adsorbability in NG Ti. 3.3.2. Cytotoxicity assessment The cytotoxicity test is applied to evaluate the toxic effect of a metallic ions released from matrix to cells. In this study, the cytotoxicity is qualitatively assayed using the LDH activity in the supernatants after culturing for 1, 4, and 24 h, and the results are presented in Fig. 5(b). The NG Ti shows similar and even slightly lower cytotoxicity compared to CG Ti. 3.3.3. Cell adhesion and proliferation assessment Fig. 5(c) shows the amount of attached hFOB1.19 cells on both NG and CG Ti obviously increases from 1 to 24 h, and at each time point more cells are adsorbed on NG Ti than CG Ti. Similarly, the cell proliferation after 3, 7 and 14 days of culture is also higher for NG Ti relative to CG Ti (Fig. 5(d)). Fig. 5(e-h) demonstrates the cell morphology after 1 and 3 days of culture on NG and CG Ti surfaces. The greater attachment already occurs on NG Ti within the first day of culture (Fig. 5(g)). After 3 days culturing, obvious cell proliferation on NG Ti is observed, since the cell density on NG Ti surface was significantly higher than that on the CG Ti surface (Fig. 5(f) and (h)). Additionally, cells on NG Ti present extensive filopodia and excellent spreading in multi-directions, indicating strong cellular adhesion and growth. These features are less pronounced on the CG surface, further verifying the cytocompatibility and bioactivity of Ti are enhanced by SFT processing. 3.3.4. ALP activity The result (insets of Fig. 5(d)) shows the intracellular ALP activity of the cells cultured on NG and CG Ti surfaces for 3, 7 and 14 days. ALP activity increases greatly with the culture time and the cells on NG Ti surface have significantly higher ALP activity than those on the CG Ti at each incubation time point. Moreover, the difference for ALP activity between both samples has been enlarged with an increasing culture time. 4. Discussion 4.1. Effect of nanocrystalline structure on corrosion behavior According to the PP and MS analysis, there are two stages involved in the mechanism underlying the growth of passive film on NG and CG Ti. The schematic diagrams for the passivation process on NG and CG Ti are illustrated in Fig. 6. At the initial stages, the surface of NG Ti can rapidly form a continuous oxide film (Fig. 6(a)), which is consistent with the OCP and PDP curves (Figs. 2 (a) and 3 (a)). However, the passive film on CG Ti surface represents a much different morphology, being discontinuity with a large number of hollows (Fig. 6(c)). Passivation processing usually first starts on the surface with crystalline lattice defects for its much higher electrochemical activity. Therefore, the NG Ti surface has a higher density of nucleation sites for passivation, which leads to the quick formation of a passive layer. With elongated immersion duration, both NG and CG Ti share an identical compact morphology in passive films, in well agreement with the electrochemical results in Fig. 2 while the maximal difference between two samples is in depths. A significantly thickened passive film is formed on NG Ti (Fig. 6(b)), which is related to the nanocrystalline structure, in that the more abundant grain boundaries provide a greater number of oxidation channels and active

69

sites due to the increasing of specific surface area and surface free energy [15]. With the thickening of the outermost oxides layer, the diffusion of oxygen from the electrolyte to the reactions interface is suppressed. Then anoxia environment caused by the blocking effect of the protective passive film resulted in the formation of many scattered suboxides at the film/metal interface [10], which is verified in the XPS results in Fig. 4, improving the stability of the outermost passive film. NG Ti is preponderant in thickness and growth rate of passive films compared with CG Ti, which render it more effective in preventing electrochemical reactions occurring continually on metals surface. Moreover, the uniform distribution grain boundaries can effectively decrease the difference of Gibbs free energy between grain boundary and intracrystal part, beneficial for relieving localized corrosion. Consequently, NG Ti exhibits better electrochemical corrosion resistance than CG Ti in SBF (Figs. 3 and 4). The superior anti-corrosion property also helps to create a more stable microenvironment that benefits the cellular functions at the cell-substrate interface as well [33].

4.2. Role of nanocrystal structure on cellular response Nanostructured Ti alloys possess an enormous development potential for biomedical material application due to their good cytocompatibility to simultaneously enhance the bone implant osseointegration [34]. Excitingly, our result shows that the NG Ti surface strikingly enhanced protein adsorption, osteoblast cells adhesion and proliferation (Fig. 5). This agrees well with a large number of reports [33–35], which demonstrated that the ultrafine or NG Ti could significantly improve protein adsorption and cell spreading. It is well known that the protein adsorption on the biomaterial surface is the initial key step determining the cellular interaction with biomaterial, because the cells sense the foreign surface through the adsorbed protein layer. Moreover, it has been generally accepted that the protein adsorption and cellular interaction with the implants can be altered by the physicochemical properties of the material surface and the structure of the macromolecules [35]. Given the observed similarities in the phase composition and crystal structure between NG and CG Ti, it can be confirmed that the differences in the cell adhesion and spreading of hFOB1.19 on NG and CG Ti are attributed to surface nanotopography and degree of hydrophilicity. The numerous and densely spaced peaks in AFM observations indicate the considerably complex surface architecture for NG Ti surface, while the CG surface is more smooth (Fig. 1(b, c, e and f)). It is reported that significant improvement for cell responses could be induced by increased nano-topography on Zr and Ta surfaces [34,36]. Herein, the demonstrated obvious nano-roughness on NG Ti should be responsible for the increment in cell adhesion and spreading. On the other hand, NG Ti surface demonstrated a lower water contact angle and higher surface energy compared to the CG Ti surface (Table 1), indicating improved hydrophobicity. The hydrophilic surface with higher surface energy can effectively enhance the adsorption of the anchoring proteins such as fibronectin and vinculin, which promotes a favorable osteogenic microenvironment [37]. Thus, the increased surface wettability of the NG Ti surface should also contribute to the notably improved protein adsorption and subsequently significantly pronounced cell responses including adhesion (Fig. 5 (c)), proliferation (Fig. 5 (d)), as well as early differentiation (Fig. 5(e–h)), which is also in agreement with other studies reporting improved cell response on ultrafine grained Ti substrate [36].

70

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

Fig. 5. Adsorption of total protein (a), cell cytotoxicity assay (b), cell attachment (c), cell proliferation and ALP activity (insets) (d) after different incubation times on NG and CG Ti surface. SEM micrographs of hFOB cells cultured on NG (e and f) and CG Ti (g and h) for different culture times, and the insets are corresponding enlarged images of the white squares. The magnified images (insets of Fig. 5 (e–f)) show that the extent of cell spreading is remarkably greater on the NG Ti surface compared to the CG Ti surface.

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

71

Fig. 6. Schematic diagrams of the passivation process for NG Ti (a and b) and CG Ti (c and d). The better corrosion resistance for NG Ti is attributed to the formation of a thicker passive film, which delays the electron migration and electrochemical reactions between nanocrystalline surface and ion transfer.

5. Conclusions In the present study, we have demonstrated that the nanocrystalline surface layer of Pure Ti produced by SFT exhibited an excellent combination of in vitro biological anti-corrosion and biocompatibility properties. The better corrosion resistance for NG Ti is attributed to the formation of a thicker passive film, which delays the electron migration and electrochemical reactions between nanocrystalline surface and ion transfer. The improved osteoblastic cells response on NG Ti is due to the higher wettability, surface energy and nano-roughness caused by the presence of nanocrystal structures. This study may promote the foundation of a new branch of nanostructured materials for biomedical applications.

Acknowledgments This work was performed under the support of the Nation Natural Science Foundation of China (Grant Nos. 51641107 and 51471136), Innovation team in key areas of Shaanxi Province (2016KCT-30) and the Nation Natural Science Foundation of Shaanxi Province (2017JQ5084).

References [1] L. Rojo, B. Gharibi, R. Mclister, B.J. Meenan, S. Deb, Self-assembled monolayers of alendronate on Ti6Al4V alloy surfaces enhance osteogenesis in mesenchymal stem cells, Sci. Rep. 6 (2016) 30548. [2] V.K. Truong, R. Lapovok, Y.S. Estrin, S. Rundell, J.Y. Wang, C.J. Fluke, R.J. Crawford, E.P. Ivanova, The influence of nano-scale surface roughness on bacterial adhesion to ultrafine-grained titanium, Biomaterials 31 (2010) 3674–3683. [3] R. Chelariu, G. Bolat, J. Izquierdo, D. Mareci, D.M. Gordin, T. Gloriant, R.M. Souto, Metastable beta Ti-Nb-Mo alloys with improved corrosion resistance in saline solution, Electrochim. Acta 137 (2014) 280–289. [4] J.W. Park, Y.J. Kim, C.H. Park, D.H. Lee, Y.G. Ko, J.H. Jang, C.S. Lee, Enhanced osteoblast response to an equal channel angular pressing-processed pure titanium substrate with microrough surface topography, Acta Biomater. 5 (2009) 3272–3280. [5] X. Xue, J. Wang, Y. Zhu, Q. Tu, N. Huang, Biocompatibility of pure titanium modified by human endothelial cell-derived extracellular matrix, Appl. Surf. Sci. 256 (2010) 3866–3873. [6] H. Carbacz, M. Pisarek, K.J. kurzydlowski, Corrosion resistance of nanostructured titanium, Biomol. Eng. 24 (2007) 559–563.

[7] T.N. Kim, A. Balakrishnan, B.C. Lee, W.S. Kim, K. Smetana, J.K. Park, B.B. Panigrahi, In vitro biocompatibility of equal channel angular processed (ECAP) titanium, Biomed. Mater. 2 (2007) S117–S120. [8] I. Dimic, I. Alagic, B. Volker, A. Hohenwarter, R. Pippan, D. Veljovic, M. Rakin, B. Bugarski, Microstructure and metallic ion release of pure titanium and Ti-13Nb-13Zr alloy processed by high pressure torsion, Mater. Des. 91 (2016) 340–347. [9] A. Thirugnanam, T.S.S. Kumar, U. Chakkingal, Tailoring the bioactivity of commercially pure titanium by grain refinement using groove pressing, Mater. Sci. Eng. C 30 (2010) 203–208. [10] J. Li, S.J. Li, Y.L. Hao, H.H. Huang, Y. Bai, Y.Q. Hao, Z. Guo, J.Q. Xue, R. Yang, Electrochemical and surface analyses of nanostructured Ti-24Nb-4Zr-8Sn alloys in simulated body solution, Acta Biomater. 10 (2014) 2866–2875. [11] J. Li, S.J. Li, Y.L. Hao, R. Yang, Electrochemical characterization of nanostructured Ti-24Nb-4Zr-8Sn alloy in 3.5% NaCl solution, Inter. J. hydrogen energ. 39 (2014) 17452–17459. [12] A. Balyanov, J. Kutnyakova, N.A. Amirkhanova, V.V. Stolyarov, R.Z. Valiev, X.Z. Liao, Y.H. Zhao, Y.B. Jiang, H.F. Xu, T.C. Lowe, Y.T. Zhu, Corrosion resistance of ultra fine-grained Ti, Scr. Mater. 51 (2004) 225–229. [13] S. Mathur, R. Vyas, K. Sachdev, S.K. Sharma, XPS characterization of corrosion films formed on the crystalline, amorphous and nanocrystalline states of the alloy Ti60Ni40, J. Non-Cryst. Solids 357 (2011) 1632–1635. [14] Y. Estrin, E.P. Ivanova, A. Michalska, V.K. Truong, R. Lapovok, R. Boyd, Accelerated stem cell attachment to ultrafine grained titanium, Acta Biomater. 7 (2011) 900–906. [15] R. Huang, H. Zhuang, Y. Han, Second-phase-dependent grain refinement in Ti-25Nb-3Mo-3Zr-2Sn alloy and its enhanced osteoblast response, Mater. Sci. Eng. C 35 (2014) 144–152. [16] M. Wen, C. Wen, P. Hodgson, Y. Li, Improvement of the biomedical properties of titanium using SMAT and thermal oxidation, Colloid Surf. B 116 (2014) 658–665. [17] D.N.A. Shri, K. Tsuchiya, A. Yamamoto, Effect of high-pressure torsion deformation on surface properties and biocompatibility of Ti-50.9mol.% Ni alloys, Biointerphases 9 (2014) 029007. [18] Y.S. Zhang, L.C. Zhang, H.Z. Niu, X.F. Bai, S. Yu, X.Q. Ma, Z.T. Yu, Deformation twinning and localized amorphization in nanocrystalline tantalum induced by sliding friction, Mater. Lett. 127 (2014) 4–7. [19] Y.S. Zhang, Q.M. Wei, H.Z. Niu, Y.S. Li, C. Chen, Z.T. Yu, X.F. Bai, P.X. Zhang, Formation of nanocrystalline structure in tantalum by sliding friction treatment, Int. J. Refract. Met. Hard Mater. 45 (2014) 71–75. [20] T. Kokubo, H. Takadama, How useful is SBF in predicting in vivo bone bioactivity, Biomaterials 27 (2006) 2907–2915. [21] D. Khang, J. Lu, C. Yao, T.J. Webster, The role of nanometer and sub-micron surface features on vascular and bone cell adhesion on titanium, Biomaterials 29 (2008) 970–983. [22] I. Miloseva, G. Zerjava, J.M.C. Morenob, M. Popa, Electrochemical properties, chemical composition and thickness of passive film formed on novel Ti-20Nb-10Zr-5Ta alloy, Electrochim. Acta 99 (2013) 176–189. [23] Y. Bai, S.J. Li, F. Prima, Y.L. Hao, R. Yang, Electrochemical corrosion behavior of Ti–24Nb–4Zr–8Sn alloy in a simulated physiological environment, Appl. Surf. Sci. 258 (2012) 4035–4040.

72

J. Lu et al. / Applied Surface Science 434 (2018) 63–72

[24] M.V. Popa, I. Demetrescu, E. Vasilescu, P. Drob, A.S. Lopez, J. Mirza-Rosca, C. Vasilescu, D. Ionita, Corrosion susceptibility of implant materials Ti-5Al-4V and Ti-6Al-4Fe in artificial extra-cellular fluids, Electrochim. Acta 49 (2004) 2113–2121. [25] C.E.B. Marino, E.M. Oliveira, R.C. Rocha-Fliho, S.R. Biaggio, On the stability of thin-anodic-oxide films of the titanium in acid phosphoric media, Corros. Sci. 43 (2001) 1465–1476. [26] J. Lu, Y. Zhao, H. Niu, Y. Zhang, Y. Du, W. Zhang, W. Huo, Electrochemical corrosion behavior and elasticity properties of Ti-6Al-xFe alloys for biomedical applications, Mater. Sci. Eng. C 62 (2016) 36–44. ˜ [27] J. Laboulais, A. Mata, V. Borrás, A. Munoz, Electrochemical characterization and passivation behavior of new beta-titanium alloys (Ti35Nb10Ta-xFe), Electrochim. Acta 227 (2017) 410–418. [28] A.M. Fekry, The influence of chloride and sulphate ions on the corrosion behavior of Ti and Ti-6Al-4V alloy in oxalic acid, Electrochim. Acta 54 (2009) 3480–3489. [29] V.A. Alves, C.M.A. Brett, Characterisation of passive films formed on mild steels in bicarbonate solution by EIS, Electrochim. Acta 47 (2002) 2081–2091. [30] H. Krawiec, V. Vignal, E. Schwarzenboeck, J. Banas, Role of plastic deformation and microstructure in the micro-electrochemical behavior of Ti-6Al-4V in sodium chloride solution, Electrochim. Acta 104 (2013) 400–406. [31] B. Munirathinam, R. Narayanan, L. Neelakantan, Electrochemical and semiconducting properties of thin passive film formed on titanium in chloride medium at various pH conditions, Thin Solid Films 598 (2016) 260–270.

[32] D.G. Li, J.D. Wang, D.R. Chen, P. Liang, Influence of passive potential on the electronic property of the passive film formed on Ti in 0.1M HCl solution during ultrasonic cavitation, Ultrason. Sonochem. 29 (2016) 48–54. [33] J. Jayaraj, A.R. Shankar, U.K. Mudali, Electrochemical and passive characterization of a beta type Ti45Zr38Al17 cast rod in nitric acid medium, Electrochim. Acta 85 (2012) 210–219. [34] L. Saldana, A. Mendez-Vilas, L. Jiang, M. Multigner, J.L. Gonzalez-Carrasco, M.T. Perez-Prado, M.L. Gonzalez-Martin, L. Munuera, N. Vilaboa, In vitro biocompatibility of an ultrafine grained zirconium, Biomaterials 28 (2007) 4343–4354. [35] Z.Q. Yao, Y. Lvanisenko, T. Diemant, A. Caron, A. Chuvilin, J.Z. Jiang, R.Z. Valiev, M. Qi, H.J. Fecht, Synthesis and properties of hydroxyapatite-containing porous titania coating on ultrafine-grained titanium by micro-arc oxidation, Acta Biomater. 6 (2010) 2816–2825. [36] R.Z. Valiev, I.P. Semenova, V.V. Latysh, H. Rack, T.C. Lowe, J. Petruzelka, Nanostructured titanium for biomedical applications, Adv. Eng. Mater. 8 (2008) 15–18. [37] D. Khang, S.Y. Kim, P. Liu-Snyder, G.T.R. Palmore, S.M. Durbin, T.J. Webster, Enhanced fibronectin adsorption on carbon nanotube/poly (carbonate) urethane: independent role of surface nano-roughness and associated surface energy, Biomaterials 28 (2007) 4756–4768.