Atomic Layer Deposition of aTiO2 Layer on Nitinol and Its Corrosion Resistance in a Simulated Body Fluid

Nitinol is a group of nearly equiatomic alloys composed of nickel and titanium, which was developed in the 1970s. Its properties, such as superelasticity and Shape Memory Effect, have enabled its use, especially for biomedical purposes. Due to the fact that Nitinol exhibits good corrosion resistance in a chloride environment, an unusual combination of strength and ductility, a high tendency for self-passivation, high fatigue strength, low Young’s modulus and excellent biocompatibility, its use is still increasing. In this research, Atomic Layer Deposition (ALD) experiments were performed on a continuous vertical cast (CVC) NiTi rod (made in-house) and on commercial Nitinol as the control material, which was already in the rolled state. The ALD deposition of the TiO2 layer was accomplished in a Beneq TFS 200 system at 250 °C. The pulsing times for TiCl4 and H2O were 250 ms and 180 ms, followed by appropriate purge cycles with nitrogen (3 s after the TiCl4 and 2 s after the H2O pulses). After 1100 repeated cycles of ALD depositing, the average thickness of the TiO2 layer for the CVC NiTi rod was 52.2 nm and for the commercial Nitinol, it was 51.7 nm, which was confirmed by X-ray Photoelectron Spectroscopy (XPS) and Scanning Electron Microscope (SEM) using Energy-dispersive X-ray (EDX) spectroscopy. The behaviour of the CVC NiTi and commercial Nitinol with and without the TiO2 layer was investigated in a simulated body fluid at body temperature (37 °C) to explain their corrosion resistance. Potentiodynamic polarisation measurements showed that the lowest corrosion current density (0.16 μA/cm2) and the wider passive region were achieved by the commercial NiTi with TiO2. Electrochemical Impedance Spectroscopy measurements revealed that the CVC NiTi rod and the commercial Nitinol have, for the first 48 h of immersion, only resistance through the oxide layer, as a consequence of the thin and compact layer. On the other hand, the TiO2/CVC NiTi rod and TiO2/commercial Nitinol had resistances through the oxide and porous layers the entire immersion time since the TiO2 layer was formatted on the surfaces.


Introduction
Nitinol is a group of alloys that are in the equiatomic composition range of nickel and titanium. It shows unique properties, such as superelasticity and a Shape-Memory Effect. It also exhibits good corrosion resistance in a chloride environment, an unusual combination of strength and ductility, a high tendency for self-passivation, high fatigue strength, low Young's modulus and excellent biocompatibility. Its properties have enabled its use, especially for biomedical purposes (orthodontic treatments, cardiovascular surgery for stents and guide wires, orthopaedic surgery for various staples and rods, maxillofacial and reconstructive surgery). In addition, Nitinol has been used in the Aerospace, Automotive, Marine and Chemical industries and Civil and Structural Engineering [1][2][3][4][5][6][7].
NiTi + O 2 → Ni 3 Ti + TiO 2 → Ni 4 Ti + TiO 2 →Ni + TiO 2 (1) Due to the four-times lower Gibbs free energy of formation of titanium oxide than that of nickel oxide, titanium oxide growth is preferred on the Nitinol surface. Observations are consistent with a model of oxygen absorption on the NiTi surface that reacts with outward diffusing Ti to form TiO 2 . During the early stages of oxidation, the growth of the TiO 2 layer is the only contributor to thickness, and, therefore, oxide formation is relatively rapid. However, the preferential oxidation of Ti creates a Ti-depleted (Ni-rich) zone at the NiTi/TiO 2 interface. The formation of the Ni-rich layer increases the effective diffusion distance with an associated decrease in overall oxidation kinetics. Therefore, continued oxide growth involves the simultaneous nucleation and growth of titanium oxides and Ni-rich phases. Ultimately, these processes lead to the formation of a protective oxide scale, which prevents further oxidation of the base material [20,21]. Naturally grown titanium oxide on a Nitinol surface is approximately 5 nm thick, and thicker titanium oxide should be formed to achieve better properties [22].
The most common methods employed for layer formation are anodisation, plasma spraying, Atomic Layer Deposition (ALD), etc. ALD is an interesting technique for producing TiO 2 thin films due to its simplicity, reproducibility, high conformity of thin films and excellent control of the layer thickness at the angstrom level. The thin film is formed as a result of repeated deposition cycles. At least two subsequent self-limited surface reactions are used to form a new layer [23][24][25][26]. In the literature, there are results for a commercial flat-annealed NiTi foil and NiTi wires (φ = 0.2 mm) [27], and for Al 2 O 3 and Pt, ALD coatings on NiTi thin films [28]. None of the studies, however, dealt with the study of ALD deposition on the surface of a continuous vertical cast (CVC) NiTi rod (φ = 11 mm), which was vacuum remelted before casting. Namely, the production of Nitinol is still very complex, and, therefore, new faster processes resulting in the shape of a rod with the smallest diameter possible are being sought.
The ALD technique for producing TiO 2 on a CVC NiTi rod made in-house [29] was tested in this study. For comparison, a commercial Nitinol was used to evaluate how the ALD coating works. CVC NiTi rod-testing production is explained in our previous works [30,31], while the commercial Nitinol was already in the rolled state. With the ALD technique, the thin layer of titanium oxide was deposited on both types of specimens. The formation of TiO 2 was investigated by X-ray Photoelectron Spectroscopy (XPS) and Scanning Electron Microscopy (SEM) using Energy-dispersive X-ray (EDX) spectroscopy. TiO 2 -protected specimens were tested on corrosion behaviour in simulated body fluid to clarify their corrosion resistance and confirm the positive effect of the formatted TiO 2 layer in order to reduce the corrosion rate.

Atomic Layer Deposition of a TiO 2 Layer
The process of depositing the layer onto the surface with ALD is cyclic. Each cycle is set up from 4 steps (see Figure 1). In the first step, the first precursor (TiCl 4 ) is pulsed into the reactor. TiCl 4 is adsorbed to the specimen's surface and reacts there with the reactive sites (initially, these are organic impurities). In the second step, nitrogen is pulsed into the reactor to purge the excess of the reactant and by-products. In the third step, the second precursor (H 2 O) is pulsed into the reactor. H 2 O is adsorbed to the specimen s surface and reacts there with the reactive sites. In the last step, nitrogen is again pulsed into the reactor to purge the excess of the reactant and by-products. 1100 cycles were made in this study. The simplified equation for this reaction is: Metals 2021, 11, x FOR PEER REVIEW of ALD deposition on the surface of a continuous vertical cast (CVC) NiTi rod (ϕ mm), which was vacuum remelted before casting. Namely, the production of Nitin still very complex, and, therefore, new faster processes resulting in the shape of a rod the smallest diameter possible are being sought. The ALD technique for producing TiO2 on a CVC NiTi rod made in-house [29 tested in this study. For comparison, a commercial Nitinol was used to evaluate how ALD coating works. CVC NiTi rod-testing production is explained in our previous w [30,31], while the commercial Nitinol was already in the rolled state. With the ALD nique, the thin layer of titanium oxide was deposited on both types of specimens formation of TiO2 was investigated by X-Ray Photoelectron Spectroscopy (XPS) and ning Electron Microscopy (SEM) using Energy-dispersive X-ray (EDX) spectros TiO2-protected specimens were tested on corrosion behaviour in simulated body flu clarify their corrosion resistance and confirm the positive effect of the formatted TiO2 in order to reduce the corrosion rate.

Atomic Layer Deposition of a TiO2 Layer
The process of depositing the layer onto the surface with ALD is cyclic. Each cy set up from 4 steps (see Figure 1). In the first step, the first precursor (TiCl4) is pulsed the reactor. TiCl4 is adsorbed to the specimen's surface and reacts there with the rea sites (initially, these are organic impurities). In the second step, nitrogen is pulsed int reactor to purge the excess of the reactant and by-products. In the third step, the se precursor (H2O) is pulsed into the reactor. H2O is adsorbed to the specimen`s surface reacts there with the reactive sites. In the last step, nitrogen is again pulsed into the re to purge the excess of the reactant and by-products. 1100 cycles were made in this s The simplified equation for this reaction is: TiCl4 + 2 H2O → TiO2 + 4 HCl The deposition of a TiO2 layer on the surface of a continuous vertical cast (CVC) rod and commercial Nitinol was accomplished in a Beneq TFS 200 system at 250 °C pulsing times for TiCl4 and H2O were 250 ms and 180 ms, followed by appropriate p cycles with nitrogen (3 s after the TiCl4 and 2 s after the H2O pulses). The chemical positions of specimens were as follows: Commercial Nitinol (45 wt.% Ti, 55 wt.% Ni the CVC NiTi rod made in-house (composition determined by XRF analysis in-hou 38.9 wt.%, 59.8 wt.% Ni). The detailed microstructure investigation revealed that the NiTi rod containing over 50 at. % Ni, consisted of Ti2Ni and cubic NiTi, with corresp ing EDX spectra in the field of investigation-see Figure 2. The chemical compositi The deposition of a TiO 2 layer on the surface of a continuous vertical cast (CVC) NiTi rod and commercial Nitinol was accomplished in a Beneq TFS 200 system at 250 • C. The pulsing times for TiCl 4 and H 2 O were 250 ms and 180 ms, followed by appropriate purge cycles with nitrogen (3 s after the TiCl 4 and 2 s after the H 2 O pulses). The chemical compositions of specimens were as follows: Commercial Nitinol (45 wt.% Ti, 55 wt.% Ni) and the CVC NiTi rod made in-house (composition determined by XRF analysis in-house: Ti 38.9 wt.%, 59.8 wt.% Ni). The detailed microstructure investigation revealed that the CVC NiTi rod containing over 50 at. % Ni, consisted of Ti 2 Ni and cubic NiTi, with corresponding EDX spectra in the field of investigation-see Figure 2. The chemical composition of the CVC NiTi rod varied through the cross and longitudinal sections because the drawing process was not optimal; the manufacturing problem is described in more detail in our previous study [29].
Metals 2021, 11, x FOR PEER REVIEW 4 the CVC NiTi rod varied through the cross and longitudinal sections because the draw process was not optimal; the manufacturing problem is described in more detail in previous study [29]. The X-ray Photoelectron Spectroscopy (XPS) analyses were performed in orde identify the oxidation states of the elements on the surface of the Ti-oxide films and culate their surface composition. XPS analyses were carried out on the PHI-TFA XPS sp trometer (Physical Electronics Inc., MN, USA), equipped with a monochromatic Al sou The analysed area was 0.4 mm in diameter, and the analysed depth was about 3-5 High energy resolution XPS spectra were taken with a pass energy of 29 eV, energy re lution of 0.6 eV and energy step of 0.1 eV. Quantification of the surface composition performed from the XPS peak intensities, taking into account the relative sensitivity tors provided by the instrument manufacturer [32]. Two places on every specimen w analysed, and the average composition was calculated. Chemical bonding of the eleme was deduced from the high-energy resolution XPS spectra using reference XPS1. The M tipak software package (version 9.9, ULVAC-PHI Inc., Japan) was used for XPS spe processing. The error in the binding energy of the measured spectra is ±0.3 eV.
Additionally, a Jeol JSM-7800F Field Emission SEM, equipped with EDX spect copy X-Max N 80, Oxford Instruments (Oxford Instruments, Abingdon, UK), was used identification of the EDX line chemical composition of the formatted TiO2 analysis on surface of the CVC NiTi rod and commercial Nitinol. For this purpose, the CVC NiTi The X-ray Photoelectron Spectroscopy (XPS) analyses were performed in order to identify the oxidation states of the elements on the surface of the Ti-oxide films and calculate their surface composition. XPS analyses were carried out on the PHI-TFA XPS spectrometer (Physical Electronics Inc., MN, USA), equipped with a monochromatic Al source. The analysed area was 0.4 mm in diameter, and the analysed depth was about 3-5 nm. High energy resolution XPS spectra were taken with a pass energy of 29 eV, energy resolution of 0.6 eV and energy step of 0.1 eV. Quantification of the surface composition was performed from the XPS peak intensities, taking into account the relative sensitivity factors provided by the instrument manufacturer [32]. Two places on every specimen were analysed, and the average composition was calculated. Chemical bonding of the elements was deduced from the high-energy resolution XPS spectra using reference XPS1. The Multipak software package (version 9.9, ULVAC-PHI Inc., Japan) was used for XPS spectra processing. The error in the binding energy of the measured spectra is ±0.3 eV.
Additionally, a Jeol JSM-7800F Field Emission SEM, equipped with EDX spectroscopy X-Max N 80, Oxford Instruments (Oxford Instruments, Abingdon, UK), was used for identification of the EDX line chemical composition of the formatted TiO 2 analysis on the surface of the CVC NiTi rod and commercial Nitinol. For this purpose, the CVC NiTi rod was cut into a cylinder shape with φ = 11 mm and h = 1 cm, using an Accutom 50 (IMT, Ljubljana, Slovenia) electronic saw for precision cutting. The commercial Nitinol, which was in the rolled state in the form of sheets with width of 1.5 mm, was cut to circles with φ = 10 mm and a width of 1.5 mm using a water jet cutter (Faculty of Mechanical Engineering, Ljubljana, Slovenia) (Omax, Kent, WA, USA). For easier polishing, the specimens were then hot-pressed into Bakelite. Mechanical polishing was performed on Struers Abramin apparatus (IMT, Ljubljana, Slovenia) (Struers, Copenhagen, Denmark). The grinding was performed with 320-grit SiC abrasive paper, mechanical polishing with MD-Largo (Struers, Cleveland, OH 44145, USA) discs with 9 µm diamond suspension and with peroxide grains in a chemically aggressive suspension-OP-S (colloidal silica). At the end, specimens were cleaned with detergent, washed well with water and put in an ultrasound bath in alcohol. Surface images were recorded at 20,000× magnification, where the SEM working parameters were 0.7 kV voltage and 2 mm working distance. Cross-section images were recorded at 200,000× magnification and SEM working parameters of 10 kV voltage and 4 mm working distance.

Corrosion Tests
The corrosion tests were performed on the CVC NiTi rod with and without a formatted TiO 2 layer and on commercial Nitinol with and without a formatted TiO 2 layer, respectively. All the measurements were held at body temperature (37 • C). Potentiodynamic polarisation measurements and Electrochemical Impedance Spectrometry (EIS) (Biologic, Seyssinet-Pariset, France) were used to study the electrochemical behaviour of the specimens. All the measurements were recorded by a BioLogic Modular Research Grade Potentiostat/Galvanostat/FRA Model SP-300 (Seyssinet-Pariset, France) with an EC-Lab Software (V 11.27, Biologic, Seyssinet-Pariset, France) and three-electrode cell. In this cell, the specimen, with an exposed area of 1 cm 2 , was a working electrode, and a saturated calomel electrode (SCE, 0.242 V vs. SHE) was used as a reference electrode, and the Counter Electrode (CE) was a platinum net. The experiment was held in simulated physiological Hank's solution, containing 8 g/L NaCl, 0.40 g/L KCl, 0.35 g/L NaHCO 3 , 0.25 g/L NaH 2 PO 4 × 2H 2 O, 0.06 g/L Na 2 HPO 4 × 2H 2 O, 0.19 g/L CaCl 2 × 2H 2 O, 0.41 g/L MgCl 2 × 6H 2 O, 0.06 g/L MgSO 4 × 7H 2 O and 1 g/L glucose, at pH = 7.8. All the chemicals were from Merck, Darmstadt, Germany. The potentiodynamic curves were recorded after 1 h specimen stabilisation at the Open-Circuit Potential (OCP), starting the measurement at 250 mV vs. SCE more negative than the OCP. The potential was then increased, using a scan rate of 1 mV s −1 until the transpassive region was reached. Long-term open-circuit potentiostatic electrochemical impedance spectra were obtained for the investigated specimens. The impedance was measured at the OCP, with sinus amplitude of 5 mV peak-to-peak and a frequency range of 65 kHz to 1 mHz, in the sequence of 0 h, 1 h, 2 h, 6 h, 12 h, 24 h, 48 h, 72 h, 96 h, 120 h, 144 h, 168 h and 192 h. The impedance data are presented in terms of Nyquist plots. Zview v3.4d Scribner Associates software (V 3.4d, Southern Pines, NC, USA) was used for the fitting process. All the experiments were repeated 3 times.

ALD TiO 2 Thin Layer Deposition
SEM investigation revealed that after 1100 repeated ALD cycles, the average thickness of the TiO 2 layer for the CVC NiTi rod was 52.2 nm, and for the commercial Nitinol 51.7 nm, as visible in Figure 3. As seen in Figure 4, the deposited TiO2 layer is quite similar on both surfaces (CVC NiTi rod and commercial Nitinol). The particles are quite different in size (from 50 nm to 1 μm), while their shape is fairly tetrahedral with quite pointed edges, and the grain boundaries are clearly visible.   The O/Ti ratio for both specimens was the same, i.e., 2.5. The presence of carbon atoms and part of the oxygen atoms are probably related to surface contamination and/or specimen preparation, taking into account that part of the oxygen is related to surface contamination. The Ti/O ratio of 2.5 indicated the TiO2 composition of the deposited films on both specimens. XPS data on the surface composition show that the TiO2/CVC NiTi rod had a higher concentration of carbon, which may be related to the higher degree of surface contamination. As seen in Figure 4, the deposited TiO 2 layer is quite similar on both surfaces (CVC NiTi rod and commercial Nitinol). The particles are quite different in size (from 50 nm to 1 µm), while their shape is fairly tetrahedral with quite pointed edges, and the grain boundaries are clearly visible. As seen in Figure 4, the deposited TiO2 layer is quite similar on both surfaces (CV NiTi rod and commercial Nitinol). The particles are quite different in size (from 50 nm 1 μm), while their shape is fairly tetrahedral with quite pointed edges, and the gra boundaries are clearly visible.   The O/Ti ratio for both specimens was the same, i.e., 2.5. The presen of carbon atoms and part of the oxygen atoms are probably related to surface contamin tion and/or specimen preparation, taking into account that part of the oxygen is related surface contamination. The Ti/O ratio of 2.5 indicated the TiO2 composition of the depo ited films on both specimens. XPS data on the surface composition show that th TiO2/CVC NiTi rod had a higher concentration of carbon, which may be related to th higher degree of surface contamination.  The presence of carbon atoms and part of the oxygen atoms are probably related to surface contamination and/or specimen preparation, taking into account that part of the oxygen is related to surface contamination. The Ti/O ratio of 2.5 indicated the TiO 2 composition of the deposited films on both specimens. XPS data on the surface composition show that the TiO 2 /CVC NiTi rod had a higher concentration of carbon, which may be related to the higher degree of surface contamination. In order to get an insight into the surface chemistry, high-energy resolution XPS spectra of Ti 2p, O 1s and C 1s were acquired on both specimens. The high-energy resolution XPS spectra Ti 2p, O 1s and C 1s from TiO2/com.NiTi were deconvoluted into different chemical components and are shown in Figure 6. The high-energy resolution XPS spectra Ti 2p, O 1s and C 1s from the TiO2/CVC NiTi rod were deconvoluted into different chemical components and are shown in Figure 7. It is visible that the high-energy resolution XPS spectra from both specimens are very similar. On the surface of both specimens, the Ti 2p3/2 peak is at 458.6 eV, and the Ti 2p1/2 peak is at 464.5 eV. This binding energy is related to the Ti(4+) oxidation state, which shows the presence of the TiO2 compound (Figures 6a and 7a). No presence of Ti was identified in the lower oxidation states, at least inside the sensitivity of the XPS method. The Ti 2p3/2 peak was very narrow, indicating an ordered TiO2 film. The Full Width at Half Maximum (FWHM) of the Ti 2p3/2 peak for both specimens was 1.07 eV. The oxygen spectra O 1s were composed of three peaks, the O1 peak at 530.0 eV, the O2 peak at 531.3 eV and a small peak O3 at 532.6 eV. The O1 peak is related to the O 2− anions in the TiO2 lattice. The O2 peak at 531.3 eV may be related to the presence of surface OH-groups or O-vacancies in the oxide. The O3 peak may be related to H2O and/or C-O species, mainly due to surface contamination. The carbon C 1s spectra from both specimens contained four peaks, which were related with different chemical bonds of the carbon atoms: the peak at 284.8 eV (C-C/C-H bonds), the peak at 286. In order to get an insight into the surface chemistry, high-energy resolution XPS spectra of Ti 2p, O 1s and C 1s were acquired on both specimens. The high-energy resolution XPS spectra Ti 2p, O 1s and C 1s from TiO 2 /com.NiTi were deconvoluted into different chemical components and are shown in Figure 6. The high-energy resolution XPS spectra Ti 2p, O 1s and C 1s from the TiO 2 /CVC NiTi rod were deconvoluted into different chemical components and are shown in Figure 7. It is visible that the high-energy resolution XPS spectra from both specimens are very similar. On the surface of both specimens, the Ti 2p 3/2 peak is at 458.6 eV, and the Ti 2p 1/2 peak is at 464.5 eV. This binding energy is related to the Ti(4+) oxidation state, which shows the presence of the TiO 2 compound (Figures 6a and 7a). No presence of Ti was identified in the lower oxidation states, at least inside the sensitivity of the XPS method. The Ti 2p 3/2 peak was very narrow, indicating an ordered TiO 2 film. The Full Width at Half Maximum (FWHM) of the Ti 2p 3/2 peak for both specimens was 1.07 eV. The oxygen spectra O 1s were composed of three peaks, the O1 peak at 530.0 eV, the O2 peak at 531.3 eV and a small peak O3 at 532.6 eV. The O1 peak is related to the O 2− anions in the TiO 2 lattice. The O2 peak at 531.3 eV may be related to the presence of surface OH-groups or O-vacancies in the oxide. The O3 peak may be related to H 2 O and/or C-O species, mainly due to surface contamination. The carbon C 1s spectra from both specimens contained four peaks, which were related with different chemical bonds of the carbon atoms: the peak at 284.8 eV (C-C/C-H bonds), the peak at 286.  In accordance with the XPS analyses, the results of the performed line EDX analysis on the formatted TiO 2 layer confirmed the increased contents of Ti and O (Figure 8) in both specimens, exactly where the TiO 2 layer formed. The TiO 2 layer is seen as a bright area. The increased concentration profiles for Ti and O are appropriate, as well as the width of the range corresponding to the resulting thickness of the TiO 2 layer, as seen with SEM ( Figure 3). Due to the extremely thin TiO 2 layer, only line EDX analysis could be used in this case, as the usual point/planar EDX analysis would capture a significantly larger volume, and in this way, the signal from the base alloy would be obtained, and the results would be irrelevant. In accordance with the XPS analyses, the results of the performed line EDX analysis on the formatted TiO2 layer confirmed the increased contents of Ti and O (Figure 8) in both specimens, exactly where the TiO2 layer formed. The TiO2 layer is seen as a bright area. The increased concentration profiles for Ti and O are appropriate, as well as the width of the range corresponding to the resulting thickness of the TiO2 layer, as seen with SEM ( Figure 3). Due to the extremely thin TiO2 layer, only line EDX analysis could be used in this case, as the usual point/planar EDX analysis would capture a significantly larger volume, and in this way, the signal from the base alloy would be obtained, and the results would be irrelevant.   Figure 9 shows the potentiodynamic curves for different specimens: CVC NiTi rod, commercial Nitinol (com. NiTi), CVC NiTi rod with a TiO 2 layer (TiO 2 /CVC NiTi rod) and commercial Nitinol with a TiO 2 layer (TiO 2 /com NiTi). All the measurements were carried out in Hank's solution at body temperature (37 • C). The electrochemical parameters of the potentiodynamic test are shown in Table 1. Corrosion potentials (E corr ) and corrosion current densities (i corr ) were obtained from the Tafel region. Following that region, the specimens exhibited a passive region, which was limited by the breakdown potential (E bd ), corresponding to the transpassive oxidation of metal species. The corrosion current density was the lowest for TiO 2 /com NiTi (0.16 µA/cm 2 ), which means that the passive layer on these specimens was the most stable and resistant to external influences. This is followed by specimens of com. NiTi (0.30 µA/cm 2 ), TiO 2 /CVC NiTi rod (0.34 µA/cm 2 ) and CVC NiTi rod (0.44 µA/cm 2 ). The formation of a passive layer is characterised by the width of the passive region. The wider the passive region, the more corrosion-resistant the material is. The CVC NiTi rod had the smallest passive range has a (from −100 mV to 330 mV, which is 430 mV), followed by TiO 2 /CVC NiTi rod (650 mV), com. NiTi (730 mV) and TiO 2 /com. NiTi. The CVC NiTi rod had the lowest breakdown potential (329 mV), followed by com. NiTi (634 mV) and TiO 2 /CVC NiTi rod (643 mV), while TiO 2 /com. NiTi did not reach the breakdown potential in the measurement range. As the name already suggests, this passive layer breaks down at this potential, which enables the formation of pits on the surface. This is also accompanied by a rapid increase in anode current due to the passivity breakdown. The corrosion rate is, thus, the smallest for TiO 2 /com. NiTi; the TiO 2 /CVC NiTi rod and com. NiTi are far away but very close together, while the CVC NiTi rod shows the highest corrosion rate or the lowest corrosion resistance by the potentiodynamic test.  Figure 9 shows the potentiodynamic curves for different specimens: CVC NiTi rod, commercial Nitinol (com. NiTi), CVC NiTi rod with a TiO2 layer (TiO2/CVC NiTi rod) and commercial Nitinol with a TiO2 layer (TiO2/com NiTi). All the measurements were carried out in Hank's solution at body temperature (37 °C). The electrochemical parameters of the potentiodynamic test are shown in Table 1. Corrosion potentials (Ecorr) and corrosion current densities (icorr) were obtained from the Tafel region. Following that region, the specimens exhibited a passive region, which was limited by the breakdown potential (Ebd), corresponding to the transpassive oxidation of metal species. The corrosion current density was the lowest for TiO2/com NiTi (0.16 μA/cm 2 ), which means that the passive layer on these specimens was the most stable and resistant to external influences. This is followed by specimens of com. NiTi (0.30 μA/cm 2 ), TiO2/CVC NiTi rod (0.34 μA/cm 2 ) and CVC NiTi rod (0.44 μA/cm 2 ). The formation of a passive layer is characterised by the width of the passive region. The wider the passive region, the more corrosion-resistant the material is. The CVC NiTi rod had the smallest passive range has a (from −100 mV to 330 mV, which is 430 mV), followed by TiO2/CVC NiTi rod (650 mV), com. NiTi (730 mV) and TiO2/com. NiTi. The CVC NiTi rod had the lowest breakdown potential (329 mV), followed by com. NiTi (634 mV) and TiO2/CVC NiTi rod (643 mV), while TiO2/com. NiTi did not reach the breakdown potential in the measurement range. As the name already suggests, this passive layer breaks down at this potential, which enables the formation of pits on the surface. This is also accompanied by a rapid increase in anode current due to the passivity breakdown. The corrosion rate is, thus, the smallest for TiO2/com. NiTi; the TiO2/CVC NiTi rod and com. NiTi are far away but very close together, while the CVC NiTi rod shows the highest corrosion rate or the lowest corrosion resistance by the potentiodynamic test.

Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy (EIS) measurements were performed at Open-Circuit Potential conditions in simulated body fluid for 8 days. Figure 10 shows (for 4 selected examples: 12 h, 96 h, 168 h and 192 h) the Nyquist impedance diagrams for the CVC NiTi rod, commercial Nitinol, and also for both with a deposited TiO2 layer at different times of immersion. The system response, shown through the Nyqvist plots, shows a typical depressed semicircle shape, and the response was increasing with the immersion time for all specimens.

Electrochemical Impedance Spectroscopy
Electrochemical Impedance Spectroscopy (EIS) measurements were performed at Open-Circuit Potential conditions in simulated body fluid for 8 days. Figure 10 shows (for 4 selected examples: 12 h, 96 h, 168 h and 192 h) the Nyquist impedance diagrams for the CVC NiTi rod, commercial Nitinol, and also for both with a deposited TiO 2 layer at different times of immersion. The system response, shown through the Nyqvist plots, shows a typical depressed semicircle shape, and the response was increasing with the immersion time for all specimens.
The analysed data for the Nyquist plots predicted the equivalent circuits shown in Figure 11. For inhomogeneous layers, a similar equivalent circuit was applied by Izquierdo J. et al. [33] and Figueira N. et al. [34]. R 1 represents the resistance through the porous external oxide layer, while R 2 represents the resistance through the inner compact oxide layer. R S is the resistance of the solution, while CPE 1 and CPE 2 are constant phase elements corresponding to R 1 and R 2 . The use of a Constant Phase Element (CPE) was required to confirm the non-ideal capacitive response observed as a depressed semicircle in the corresponding Nyquist diagrams. The CPE originates from the surface roughness and inhomogeneities present in the TiO 2 layers at the microscopic level. The equivalent circuit in Figure 9a has only resistance through the oxide layer and was used only for the specimens of CVC NiTi rod and commercial Nitinol in the first 48 h of immersion. After this exposure time, the equivalent circuit in Figure 9b was applied, and this equivalent circuit was also valid for all the immersion times for the TiO 2 /CVC NiTi rod and TiO 2 /commercial Nitinol. The difference can be explained by not having an oxide layer at the beginning on the surface of the CVC NiTi rod and the commercial Nitinol. Therefore, for the first 48 h of immersion, the resulting layer is still very thin and compact, which can be represented by only 1 resistor. After this time, the oxide layer became thicker and inhomogeneous, so another resistance was added, which represents the resistance through the porous oxide layer. On the surface of the TiO 2 /CVC NiTi rod and the TiO 2 /commercial Nitinol specimens, a nanosized TiO 2 layer was deposited previously with ALD, so an equivalent circuit with two resistances was used from the beginning of the immersion.  The analysed data for the Nyquist plots predicted the equivalent circuits shown in Figure 11. For inhomogeneous layers, a similar equivalent circuit was applied by Izquierdo J. et al. [33] and Figueira N. et al. [34]. R1 represents the resistance through the porous external oxide layer, while R2 represents the resistance through the inner compact oxide layer. RS is the resistance of the solution, while CPE1 and CPE2 are constant phase elements corresponding to R1 and R2. The use of a Constant Phase Element (CPE) was required to confirm the non-ideal capacitive response observed as a depressed semicircle in the corresponding Nyquist diagrams. The CPE originates from the surface roughness and inhomogeneities present in the TiO2 layers at the microscopic level. The equivalent first 48 h of immersion, the resulting layer is still very thin and compact, which can be represented by only 1 resistor. After this time, the oxide layer became thicker and inhomogeneous, so another resistance was added, which represents the resistance through the porous oxide layer. On the surface of the TiO2/CVC NiTi rod and the TiO2/commercial Nitinol specimens, a nanosized TiO2 layer was deposited previously with ALD, so an equivalent circuit with two resistances was used from the beginning of the immersion.   Table 2 shows the resistance values of the porous layer (R1) and the oxide layer (R2) for all the specimens. The error obtained when fitting the EIS experimental data was below 0.02% for all the specimens. The CVC NiTi rod and the commercial Nitinol, for the first 48 h of immersion, had only resistance through the oxide layer as a consequence of the thin and compact layer (R2). After this time, resistance through the porous layer was also considered. Both the TiO2/CVC NiTi rod and TiO2/commercial Nitinol had resistances through both layers for the total immersion time (R1 + R2), since the TiO2 layer was previously deposited on the surface of the specimens. The resistance R1 was considerably lower for all the specimens, which was not surprising since the solution will penetrate through the porous layer much more easily than through the compact layer. While the resistance through the oxide layer was increasing over the immersion time for all the specimens, the layer was getting thicker, so the resistance through the porous layer was quite constant.
. Porous corrosion resistance R1 and oxide corrosion resistance R2 of different specimens at certain times of immersion.

h)
R1com × 10 5 (Ω)   Table 2 shows the resistance values of the porous layer (R 1 ) and the oxide layer (R 2 ) for all the specimens. The error obtained when fitting the EIS experimental data was below 0.02% for all the specimens. The CVC NiTi rod and the commercial Nitinol, for the first 48 h of immersion, had only resistance through the oxide layer as a consequence of the thin and compact layer (R 2 ). After this time, resistance through the porous layer was also considered. Both the TiO 2 /CVC NiTi rod and TiO 2 /commercial Nitinol had resistances through both layers for the total immersion time (R 1 + R 2 ), since the TiO 2 layer was previously deposited on the surface of the specimens. The resistance R 1 was considerably lower for all the specimens, which was not surprising since the solution will penetrate through the porous layer much more easily than through the compact layer. While the resistance through the oxide layer was increasing over the immersion time for all the specimens, the layer was getting thicker, so the resistance through the porous layer was quite constant.  Figure 12 represents the polarisation or the total corrosion resistance R p as a function of time. R p can be calculated according to equation:  Figure 12 represents the polarisation or the total corrosion resistance Rp as a function of time. Rp can be calculated according to equation: The diagram of total corrosion resistance ( Figure 12) shows clearly that the corrosion resistance increased with time for all 4 specimens. This was a consequence of the formation of a protective oxide layer. In the case when the TiO2 layer was deposited on the specimen, it is noticeable that the corrosion resistance was increasing in the initial time, while for the CVC NiTi rod and commercial Nitinol, in the initial 48 h, there was no significant change in resistance as a consequence of the thin oxide layer. The highest corrosion resistance was for the TiO2/commercial Nitinol, followed by the commercial Nitinol and the TiO2/CVC NiTi rod, and the poorest corrosion resistance was for the CVC NiTi rod. The cause of the poorer corrosion properties of the CVC NiTi rod was inhomogeneity in the chemical composition and higher nickel content than in the commercial NiTi.

Conclusions
The following conclusions can be drawn from the current research work: 1. The thickness of the formatted TiO2 layers on the CVC NiTi rod was 52.2 nm, and on the commercial Nitinol it was 51.7 nm. 2. The formatted TiO2 layers were confirmed by XPS and SEM/EDX analyses. 3. The high-energy resolution XPS spectra for TiO2 from both specimens were very similar. The Ti 2p3/2 peak at 458.6 eV and the Ti 2p1/2 peak at 464.5 eV were observed on the surface of both specimens. This corresponds to the binding energy, which is related with the Ti(4+) oxidation state. This shows the presence of the TiO2 compound on the surfaces of both investigated ALD TiO2-covered specimens. The diagram of total corrosion resistance ( Figure 12) shows clearly that the corrosion resistance increased with time for all 4 specimens. This was a consequence of the formation of a protective oxide layer. In the case when the TiO 2 layer was deposited on the specimen, it is noticeable that the corrosion resistance was increasing in the initial time, while for the CVC NiTi rod and commercial Nitinol, in the initial 48 h, there was no significant change in resistance as a consequence of the thin oxide layer. The highest corrosion resistance was for the TiO 2 /commercial Nitinol, followed by the commercial Nitinol and the TiO 2 /CVC NiTi rod, and the poorest corrosion resistance was for the CVC NiTi rod. The cause of the poorer corrosion properties of the CVC NiTi rod was inhomogeneity in the chemical composition and higher nickel content than in the commercial NiTi.

Conclusions
The following conclusions can be drawn from the current research work: 1.
The thickness of the formatted TiO 2 layers on the CVC NiTi rod was 52.2 nm, and on the commercial Nitinol it was 51.7 nm. 2.
The formatted TiO 2 layers were confirmed by XPS and SEM/EDX analyses. 3.
The high-energy resolution XPS spectra for TiO 2 from both specimens were very similar. The Ti 2p 3/2 peak at 458.6 eV and the Ti 2p 1/2 peak at 464.5 eV were observed on the surface of both specimens. This corresponds to the binding energy, which is related with the Ti(4+) oxidation state. This shows the presence of the TiO 2 compound on the surfaces of both investigated ALD TiO 2 -covered specimens.

4.
The potentiodynamic test showed that the passive layer on the TiO 2 /commercial Nitinol was the most stable and resistant to external corrosion influences. The corrosion stability fell from the commercial Nitinol and the CVC NiTi rod with and without the TiO 2 layer. 5.
The corrosion rate was the smallest for TiO 2 /commercial Nitinol; the TiO 2 /CVC NiTi rod and commercial Nitinol were far away but very close together, while the CVC NiTi rod showed the highest corrosion rate, or the lowest corrosion resistance, by the potentiodynamic test.

6.
Electrochemical Impedance Spectroscopy is interpreted using the Nyquist impedance diagrams, where a typical depressed semicircle shape and the response increasing with the immersion time were shown for all specimens. 7.
With the help of using the resistance through the porous external oxide layer (R 1 ) and the resistance through the inner compact oxide layer (R 2 ), it was determined that the CVC NiTi rod and the commercial Nitinol had, for the first 48 h of immersion, only resistance through the oxide layer as a consequence of the thin and compact layer (R 2 ). On the other hand, the TiO 2 /CVC NiTi rod and TiO 2 /commercial Nitinol had resistances through both layers for the total immersion time (R 1 + R 2 ). The resistance R 1 was considerably lower for all the specimens, which was not surprising since the solution will penetrate through the porous layer much more easily than through the compact layer. 8.
It was proven that adding a TiO 2 layer on the Nitinol surface was significant for improving the corrosion resistance, but a decisive role can be still attributed to the chemical composition and microstructure of the substrate that the ALD coating is applied.