Effect of the Thickness of TiO2 Films on the Structure and Corrosion Behavior of Mg-Based Alloys

This article discusses the influence of the thickness of TiO2 films deposited onto MgCa2Zn1 and MgCa2Zn1Gd3 alloys on their structure, corrosion behavior, and cytotoxicity. TiO2 layers (about 200 and 400 nm thick) were applied using magnetron sputtering, which provides strong substrate adhesion. Such titanium dioxide films have many attractive properties, such as high corrosion resistance and biocompatibility. These oxide coatings stimulate osteoblast adhesion and proliferation compared to alloys without the protective films. Microscopic observations show that the TiO2 surface morphology is homogeneous, the grains have a spherical shape (with dimensions from 18 to 160 nm). Based on XRD analysis, it can be stated that all the studied TiO2 layers have an anatase structure. The results of electrochemical and immersion studies, performed in Ringer’s solution at 37 °C, show that the corrosion resistance of the studied TiO2 does not always increase proportionally with the thickness of the films. This is a result of grain refinement and differences in the density of the titanium dioxide films applied using the physical vapor deposition (PVD) technique. The results of 24 h immersion tests indicate that the lowest volume of evolved H2 (5.92 mL/cm2) was with the 400 nm thick film deposited onto the MgCa2Zn1Gd3 alloy. This result is in agreement with the good biocompatibility of this TiO2 film, confirmed by cytotoxicity tests.


Introduction
Biodegradable magnesium alloys have many advantages in comparison with other biomaterials used for bone implants. The density, modulus of elasticity (Young's modulus), and tensile strength of magnesium are similar to human bones. Implants based on titanium, cobalt alloys, or stainless steel are used in orthopedic applications at present. They have many advantageous properties and are characterized by good corrosion resistance. However, these materials have to remain in the human body for many years. Magnesium and its alloys are characterized by in vivo biodegradability, and this advantage makes Mg alloys differ from the implants used today [1][2][3][4][5]. Therefore, Mg alloys can be used as temporary implants. An important property is the biocompatibility of magnesium, which means the acceptance of Mg by the surrounding tissues. In the human body, magnesium ions interact with polyphosphate compounds (e.g., adenosine triphosphate (ATP), DNA, and RNA). Moreover, Samples of the studied alloys in the form of cylinders with diameters of 13 mm and heights of 4 mm were the substrate materials for the magnetron sputtering. The process was carried out at a temperature of 100 • C, with an Ar atmosphere using a Kurt J. Lesker PVD 75 device (Kurt J. Lesker Company, Jefferson Hills, USA). A Ti target (99.9% pure) was used for the magnetron sputtering process. Oxygen (99.99% pure) was supplied to the chamber. The process was performed over either 90 or 180 min.
The thickness of the TiO2 films was measured using a Filmetrics F20 reflectometer (KLA Company, San Diego, USA). This device takes the reflection of light from a thin layer, and analyzes this light over a certain range of wavelengths. The reflectometer measured thickness in the range of 15 nm-70 µm.

Structure Study and Phase Analysis
The observations of surface morphology were conducted with a Zeiss SUPRA 35 scanning electron microscope (SEM, Thornwood, New York, USA) (EHT = 3.0 kV; SE mode, in-lens detector), which was equipped with an energy-dispersive (EDS) detector.
Surface topography observations and roughness measurements were carried out using an AFM XE-100 atomic force microscope from Park Systems (Suwon, South Korea). The experiment was run in non-contact mode. The observation area was 25 µm 2 . XEI software (1.8 version) was used for the roughness parameters (e.g., roughness average (Ra) and root mean square (RMS)) calculations.
The phase analysis of the TiO2 films was carried out with a PANalytical X'Pert PRO X-ray diffractometer (PANalytical, Almelo, Netherlands), using Co Kα radiation. The analysis was conducted with the step registration method over a 2θ angular range from 25 • to 80 • . X-ray qualitative analysis was carried out with HighScore Plus software (3.0e version) using a dedicated PAN-ICSD (Inorganic Crystal Structure Database) phase identification card database.

Electrochemical and Immersion Tests
Electrochemical corrosion tests were performed using an Autolab PGSTAT302N Multi BA potentiostat (Metrohm AG, Herisau, Switzerland). A saturated calomel electrode was the reference electrode and a platinum rod was the counter electrode. The experiment was carried out in Ringer's chloride-rich solution at a temperature of 37 • C. The scan rate of the corrosion potential was 1 mV·s −1 . Before the measurements, the samples were immersed in Ringer's solution for 5 min for stabilization. The corrosion parameters (e.g., corrosion potential-E corr , corrosion current density-i corr , and corrosion polarization resistance-R p ) were determined using Tafel's analysis.
Moreover, immersion tests for the estimation of gas corrosion product (volume of H 2 evolution) for the TiO 2 films were performed. The studies were conducted in the Ringer's solution at 37 • C for 24 h.
Cylindrical samples, with a testing area of 1.1 cm 2 for electrochemical studies, and of 1.3 cm 2 for immersion tests, were prepared. The volume of H 2 evolution was measured and calculated, taking into account the frontal area of the samples.

Analysis of Corrosion Products
After the immersion tests, the corroded surfaces of the TiO 2 films were observed using SEM. To remove the corrosion products, the samples were rinsed with distilled water and immersed in CrO 3 solution before the observations. Fourier transform infrared (FTIR) spectroscopy was used to analyze the corrosion products. FTIR spectra were recorded for the MgCa2Zn1 and MgCa2Zn1Gd3 alloys with the TiO 2 films at room temperature using a Nicolet 6700/8700 FTIR spectrometer (Thermo Fisher Scientific, Waltham, USA). For this purpose, the corrosion products were collected from the surface of immersed samples and mixed together with dry KBr. Measurements of samples were conducted in transmission mode in a mid infrared range of 4000-400 cm −1 .

Cytotoxicity Assays
Preliminary biological in vitro investigations were carried out using mouse fibroblast cells. Samples of the MgCa2Zn1 and MgCa2Zn1Gd3 alloys in their as-cast state and with TiO 2 films were used for the experiments. The samples were sterilized in an antibiotic bath. Cell culture was carried out in Medium 199 with 10% fetal bovine serum (FBS) over 24 h. The tests were performed at 37 • C with a CO 2 (5%) atmosphere and a constant 95% humidity. The samples were placed into a dish culture, where the confluence of the cells was up to 90%. The lack of vital functions (necrosis) of the cells was estimated after 24 h. FDA (fluorescein diacetate) was used as the cell viability stain, and propidium iodide (PI) was used to determine the number of dead cells. The relative number of dead cells was measured using a Zeiss AXIO OBSERVER (Zeiss, New York, USA) inverted fluorescence microscope equipped with AxioVision 4.6 software.

Results and Discussion
MgCa2Zn1 and MgCa2Zn1Gd3 alloys were used as substrate materials for the magnetron sputtering process. The titanium dioxide films were applied onto the two Mg-based alloys with 90 and 180 min deposition times.
The thicknesses of the TiO 2 films measured with the reflectometer were 200 nm (for the deposition time of 90 min) and 400 nm (for the 180 min deposition time), respectively. The results of the measurements showed that the thickness of the films increased proportionally to the increase in the deposition time. The 200 nm thick TiO 2 film was selected based on previous studies presented by Kalisz et al. [30]. The authors applied TiO 2 film (200 nm thick) onto biomedical Ti6Al4V alloy using conventional magnetron sputtering process. The thin film had an anatase structure. They stated that the TiO 2 coating showed the best corrosion properties among other tested coatings in 0.5M NaCl solution. Based on these results, the 200 nm thick film was applied onto magnesium alloys to check the corrosion behavior.
The TiO 2 films applied onto the MgCa2Zn1 and MgCa2Zn1Gd3 alloys were similar in surface morphology, the grains of the films being formed with a spherical-like morphology ( Figure 1) [21]. In the case of oxide films deposited onto the MgCa2Zn1 alloy, the grains were slightly different in sizes. The surface of the 200 nm thick TiO 2 was characterized by grains with sizes in the range of 18-120 nm ( Figure 1a). The second film, with a thickness of about 400 nm, was characterized by a larger number of grain aggregations with defined grain boundaries. The grain sizes were from 18 to 160 nm ( Figure 1b). The results of porosity measurements showed that the porosity of the thinner TiO 2 film was equal to 1.85% (±0.7 SD), and of the 400 nm thick film was 3.56% (±1.1 SD). It can be stated that for the thickness of 200 nm and the MgCa2Zn1 alloy as a substrate material, the TiO 2 film was denser and more compact compared to the thicker TiO 2 film. Similar surface morphologies to the 400 nm thick TiO 2 film were noted with the titanium dioxide films deposited onto the MgCa2Zn1Gd3 alloy (Figure 1c,d). It could be observed that the grains of the 200 and 400 nm thick TiO 2 films also had similar sizes to the grains of the 400 nm thick film deposited on the MgCa2Zn1 alloy. The porosity of the thinner film was equal to 3.43 (±1.4 SD), and for the 400 nm thick film was 2.91% (±1.2 SD).
The results of X-ray phase analysis confirmed that particular TiO 2 phases were detected in the studied films (Figures 2 and 3). In the diffraction patterns, there were characteristic peaks at 29.47, 44.17, 56.43, 63.49, and 64.94 degrees of the 2θ angle for both the TiO 2 deposited onto MgCa2Zn1 and that on the MgCa2Zn1Gd3. Based on the phase analysis, it can be stated that the TiO 2 has an anatase structure (space group I41/amd; space group No. 141; lattice parameters: a = 3.784 Å, b = 3.784 Å, c = 9.515 Å). Anatase has a tetragonal crystal structure [24]. This structure is more biologically active compared to other structures, such as rutile or brookite. Moreover, the XRD patterns showed strong diffraction peaks from Mg on the alloy's surface.   The analysis of the topography of the TiO 2 films was performed in an area of 25 µm 2 . The films with a thickness of about 200 and 400 nm were similar in their topography for both the alloys. It could be observed that the studied TiO 2 films had a granular structure (Figure 4). The aggregations of particles for the same TiO 2 thickness for the both Mg alloys had a similar shape. Based on the analysis results for all the samples, it can be stated that the surface irregularities of the TiO 2 films did not exceed 25 nm. During the measurements, roughness parameters (e.g., roughness average, R a , and root mean square, RMS) were determined ( Table 1). The results of the roughness measurements showed that the values of R a and RMS for the 200 and 400 nm thick TiO 2 applied onto the MgCa2Zn1 alloy were similar. The values of the roughness parameters decreased with the increase in the films' thickness. The irregularities occurring on the surface of the thinner films were filled out during the deposition process of the 400 nm thick films, hence the differences in the roughness values. Larger differences in the R a and RMS values occurred for the films deposited on the MgCa2Zn1Gd3 alloy. Electrochemical investigations of the corrosion resistance of the alloys, with and without TiO 2 films, were carried out. The measurements were performed in Ringer's solution, which is enriched in chlorides, at a temperature of 37 • C. Such potentiodynamic curves are known to give some information on the corrosion mechanism of TiO 2 films [5]. The titanium dioxide films deposited on the MgCa2Zn1 and MgCa2Zn1Gd3 alloys were characterized by higher values of corrosion potentials (E corr ), and lower values of corrosion current densities (i corr ) compared to the substrate materials (Figures 5 and 6). This may suggest an improvement in the corrosion resistance of the studied alloys with TiO 2 deposited by the PVD technique [5,17,31]. Li et al. [31] studied the corrosion behavior for titania coatings on Mg-Ca alloy immersed in SBF (simulated body fluid) solution for 48, 168, and 336 h. They stated that the corrosion resistance was improved for the alloy with TiO 2 .  The results of the electrochemical studies showed that the MgCa2Zn1Gd3 alloy with the 400 nm thick TiO 2 film had a better corrosion resistance compared to the 200 nm thick film ( Figure 5). This result was in agreement with the results of roughness measurements, where the RMS and R a values of this film were lower compared to the thinner film. The corrosion parameters for the both the studied alloys and the TiO 2 films applied are listed in Table 2. The corrosion potential of the thicker film shifted to more positive values (E corr for the 200 nm thick TiO 2 film was −1.50 V, and E corr for the 400 nm thick TiO 2 film was −1.43 V). The values of corrosion current densities and polarization resistances (R p ) for the studied films also confirmed an improvement in the corrosion resistance of the MgCa2Zn1Gd3 alloy (i corr for the TiO 2 /200 nm was 130 µA·cm −2 , and R p was 440 Ω·cm 2 ; i corr for the TiO 2 /400 nm was 98 µA·cm −2 , and R p was 510 Ω·cm 2 ). In the case of the oxide films deposited onto the MgCa2Zn1 alloy, differences in the corrosion potential values compared to the films applied on MgCa2Zn1Gd3 were observed ( Figure 6). The E corr for the TiO 2 film with a deposition time of 90 min was −1.40 V, and for the thicker one was −1.43 V. The values of R p were 350.4 and 298.6 Ω·cm 2 , for the 200 and 400 nm thick TiO 2 films, respectively. The corrosion current density for the 200 nm thick TiO 2 film was 78 µA·cm −2 , and for the thicker film it was 96 µA·cm −2 . The results of the electrochemical tests indicated that the increase in TiO 2 thickness did not always decrease the degradation rate. The higher density of the thinner film affected the corrosion behavior by improving the corrosion resistance of the substrate material [17,31].
The results of the roughness studies for the both TiO 2 films show that the RMS and R a values of the 200 nm thick film were slightly higher. In some reports on surface treatments [32], the authors stated that there was no influence of the roughness on the degradation rate of the alloys with deposited coatings. Gawlik et al. [32] indicated that in some cases rougher films are characterized by less pitting corrosion and quite good cell adhesion. However, electrochemical results are the only available information about the initial state of the corrosion behavior [33].
After the potentiodynamic tests, the studied TiO 2 films and uncoated alloys were immersed in Ringer's solution at a temperature of 37 • C, over a time of 24 h. The results of such immersion tests indicate the real long-term corrosion behavior of materials for medical applications [7,33]. After 24 h of immersion, the 400 nm thick film deposited onto MgCa2Zn1 alloy was characterized by a higher H 2 volume of 12.4 mL/cm 2 . The volume of hydrogen evolution for the thinner film was 8.8 mL/cm 2 (Figure 7). The corrosion rates, v corr , calculated from the hydrogen evolution volume after 24 h of immersion were 3.08 and 4.34 mm·y −1 for the 200 and 400 nm thick TiO 2 , respectively. The v corr for the MgCa2Zn1 alloy was 4.9 mm·y −1 . The results of immersion tests confirmed the electrochemical results, where the thinner film was characterized by a slight improvement in the corrosion resistance. This probably resulted from the grain refinement of the 200 nm thick TiO 2 film. Moreover, this film was denser and more complex compared to the thicker one.
A slightly lower volume of evolved hydrogen was noted for the films applied onto MgCa2Zn1Gd3 alloy (Figure 8). It can be observed that the volume of H 2 evolved decreased almost proportionally with respect to the increase in the thickness of the TiO 2 films.
The volume of evolved H 2 was 11.85 and 5.92 mL/cm 2 for the 200 and 400 nm thick films, respectively. The corrosion rate for the thinner film was 4.81 mm·y −1 , and for the thicker one was 2.4 mm·y −1 . Moreover, the 400 nm thick TiO 2 film was characterized by the lowest volume of evolved hydrogen compared to other studied films. This result bodes well for future investigations. The vcorr for the MgCa2Zn1Gd3 alloy was 6.45 mm·y −1 . Amaravathy et al. [23] examined the hydrogen evolution of HA/TiO 2 on Mg alloy. They confirmed the decrease in corrosion rate for the composite coated alloy by the reduction in the evolved hydrogen produced by magnesium corrosion. The hydrogen evolution process was also analyzed by Bakhsheshi-Rad et al. [34]. The H2 evolution rates of Si/TiO 2 , and Si coated Mg-Ca alloys were 1.57 mL/cm 2 /day and 2.22 mL/cm 2 /day, respectively. The uncoated alloy was characterized by a higher hydrogen evolution rate of 5.04 mL/cm 2 per day.  After the immersion tests, the samples of the studied alloys with TiO 2 films with improved corrosion resistance were microscopically observed. The results of the microscopic observations of the 200 nm thick film deposited onto MgCa2Zn1 and of the 400 nm thick film applied on the MgCa2Zn1Gd3 alloy are presented in Figure 9. Microcracks were observed in both the TiO 2 films, due to dehydration during the drying of the samples (Figure 9a,b). Some needle-like and lamellar-shaped corrosion products were visible on the surfaces of the titanium dioxide films (Figure 9c,d). These were probably from magnesium hydroxides [35,36]. Mg(OH) 2 is the main corrosion product of Mg alloys. It forms a loose film and cannot protect the Mg alloy surface from the degradation process [37]. In our study, it was observed that some corrosion products fell off from the surfaces of the films (Figure 9c,d). After the removal of the corrosion products, the grain boundaries were still visible (Figure 9e,f). Small pits were observed on the surfaces of both the oxide films, which may suggest the occurrence of pitting corrosion [33]. It can be stated that the corroded surface of the TiO 2 film deposited onto MgCa2Zn1Gd3 alloy was characterized by lower corrosion damage compared to the second film (Figure 9e). This confirmed an improvement in the corrosion resistance of the film applied onto the Gd-containing alloy. Although this 400 nm TiO 2 film is uniform, it is porous. The chloride ions in Ringer's solution (8.6 g/dm 3 NaCl, 0.3 g/dm 3 KCl, 0.48 g/dm 3 CaCl 2 6H 2 O) penetrate the TiO 2 coating and reach the substrate through the micropores. Increase in immersion time causes that the TiO 2 film begins to degrade [34]. Thus more solution interacts with the Mg alloy. When TiO 2 coated alloy is immersed in Ringer's solution, the MgO in the outer layer starts to react with chloride-containing solution, and converts to Mg(OH) 2 .
The corrosion products of the TiO 2 films were characterized by FTIR spectroscopy. The analysis results, presented in Figure 10 [38]. According to these results, the main corrosion reactions were associated with the formation of hydroxides and carbonates. It can be assumed that the hydroxides were formed first, and then the reaction between them and CO 2 from the air resulted in the formation of carbonates. The presence of Mg(OH) 2 confirmed that the needle-like and lamellar-shaped corrosion products were from magnesium hydroxides. Measurements of the relative number of dead cells after 24 h of culture for the studied alloys and the alloys with TiO 2 films were conducted ( Figure 11). The investigations included only the measurements of the titanium dioxide films with the lowest volume of evolved H 2 from the immersion tests. The hydrogen evolution volume has a strong impact on cell viability and proliferation. The investigation results showed that the MgCa2Zn1 alloy and this alloy with the 200 nm thick TiO 2 film caused a toxic effect on the cell culture compared with the control sample. The number of dead cells was equal to 38% (±4.2 SD) and 32% (±3.6 SD) for the MgCa2Zn1 and the alloy with the 200 nm thick film, respectively. This indicates a cytotoxic effect in accordance with ISO 10993-5 standard [39]. These materials also caused very strong cell lysis (cell disintegration due to the destruction of cell membranes) (Figure 11a). Moreover, changes in cells morphology and proliferation were shown. The cell density was low, most of the cells were rounded, what indicates their degradation. On the other hand, the gadolinium-containing alloy was characterized by less cytotoxicity, the number of dead cells being equal to 28% (±5 SD). The cytotoxicity of the MgCa2Zn1Gd3 was found to be grade 2, which means a mild reactivity (Figure 11b). The achievement of a numerical grade greater than 2 indicates a cytotoxic effect [39]. This material was characterized with slightly lower cell density, the morphology of cells seeded on this substrate was correct. A lower cell necrosis was observed for the 400 nm thick TiO 2 deposited onto MgCa2Zn1Gd3 alloy, the number of dead cells being equal to 18% (±3.7 SD) in comparison with the control sample (Figure 11c).
The titanium dioxide film was characterized by a slight reactivity, achieving a grade of 1 [39]. This result was in agreement with the immersion test results, where the volume of evolved H 2 for the 400 nm thick TiO 2 film was 5.92 mL/cm 2 . The density of cells cultured on this material was compared to control. The cells had proper, characteristic to fibroblasts, spindle-shaped morphology.
The results of the cytotoxicity tests for the MgCa2Zn1Gd3 alloy with 400 nm thick titanium dioxide film are promising for future research.
Cell culture studies were also performed by Amaravathy et al. [40]. They stated that a TiO 2 coated AZ31 alloy showed higher cell attachment and proliferation compared to the uncoated alloy.

Conclusions
The analysis of the investigation's results showed that the surface morphology of the studied TiO 2 films was homogeneous, with grains of a spherical shape and dimensions from 18 to 160 nm. Slight differences in the morphology were observed for the 200 nm thick TiO 2 film applied onto MgCa2Zn1 alloy. This film was denser and more complex compared to the other studied films. Based on the XRD analysis, it can be stated that all the titanium dioxide films had an anatase structure, which is known to have antibacterial properties. The results of the Tafel's analysis showed that the oxide films slightly improved the corrosion resistance of the studied MgCa2Zn1 and MgCa2Zn1Gd3 alloys. The corrosion resistance of the TiO 2 films applied onto MgCa2Zn1Gd3 was slightly improved by increasing the thickness of the films. In the case of the oxide films deposited on the MgCa2Zn1 alloy, the results of electrochemical and immersion tests differed. The 200 nm thick titanium dioxide film was characterized by a higher corrosion resistance compared to the thicker film. This was because of the grain refinement of the thinner TiO 2 film. The films deposited onto the MgCa2Zn1Gd3 alloy showed a lower volume of hydrogen evolution in comparison to the TiO 2 applied on the MgCa2Zn1 alloy. After 24 h of immersion, the volume of evolved H 2 for the 400 nm thick TiO 2 film deposited onto MgCa2Zn1Gd3 was equal to 5.92 mL·cm −2 , and for the thinner film was 11.85 mL·cm −2 . The hydrogen evolution volumes for the oxide films applied onto the MgCa2Zn1 alloy were 12.4 and 8.8 mL·cm −2 for the 400 and 200 nm thick TiO 2 , respectively. The results of the cytotoxicity tests confirmed that the titanium dioxide film (400 nm thick) deposited on the magnesium alloy with gadolinium addition fostered cell proliferation. The number of dead cells calculated using cytotoxicity tests was equal to 18%. This result may suggest a good biocompatibility of this TiO 2 film.
Author Contributions: A.K. conceived and designed the experiments; P.N., A.N.-C., A.R., and A.K. performed the experiments; R.B. and A.K. analyzed the data and wrote this article; A.K. and R.B. participated in the revision of the manuscript. All the authors have commented on the manuscript. All authors have read and agreed to the published version of the manuscript.