Determination of the Corrosion and Biocompatibility Properties of As-Cast TiNi Alloys

Abstract

In this study, a TiNi alloy with a composition of 50 at.% of titanium and 50 at.% of nickel is investigated in terms of its corrosion and biocompatibility behavior for biomedical applications. The corrosion measurements, which include the determination of open-circuit potential and linear polarization resistance measurements, cyclic polarization measurements, and electrochemical impedance spectroscopy in 9 g L⁻¹ NaCl, show that TiNi has satisfactory corrosion stability. According to the SEM and EDS analysis, after cyclic polarization, pitting corrosion occurred, accompanied by the dissolution of the unstable Ti₂Ni inclusions. The analysis also showed that TiNi has good biocompatibility for human osteoblast-like cells, as determined by the mitochondrial activity, which was assessed using a direct contact test following ISO standard 10993-5, via scanning electron microscopy (SEM) and fluorescent microscopy.

1. Introduction

TiNi shape memory alloys (SMAs) are functional biomaterials with a unique ability to change shape and remain plastically deformed at lower temperatures, while exhibiting a reverse feature that recovers the original (pre-deformed, i.e., remembered) shape above critical temperatures [1,2]. The alloy that contains the equiatomic or near-equiatomic composition of titanium and nickel possesses shape memory and exhibits a superelastic effect due to reversible thermoelastic martensitic transformations. Moreover, this composition forms a TiO₂ layer, which significantly reduces the nickel ion release, thus providing sufficient biocompatibility for biomedical applications [3]. As an intelligent material that combines sensing and actuation based on thermal input, the TiNi alloy allows for high work output, thanks to which it is used to make biosensors, tissue engineering devices, and drug delivery systems [4]. Nickel–titanium alloys are used to manufacture coronary and ureteral stents, surgical instruments, flexible micro-needles, endoscopic catheters, endodontic instruments, orthodontic wires, and orthopedic implants, for use in dentistry, etc. [2,5,6,7].

Besides their functional properties, the interaction of TiNi alloys with imaging techniques, particularly magnetic resonance imaging (MRI), is an important aspect of their medical application [8]. TiNi SMA has gained widespread attention in the orthopedic field due to its ability to replicate the unique characteristics of human tissues and bones [9,10]. TiNi alloys exhibit high flexibility values, accommodating up to 8% strain, compared to Co-Cr and stainless steel, which resist only 1% strain [11,12]. TiNi alloys have modulus values of 30 GPa in the martensitic phase, while the modulus of human bone is approximately 20 GPa [13]. Their benefits include ease of manufacturing, simple and rapid application, effective compression osteosynthesis leading to primary bone healing, and relatively low costs; this makes them a promising alternative to traditional bone plates and/or screws [14,15,16,17]. As the TiNi shape memory alloys are characterized by high nickel content, the main concerns for their commercial application are focused on potential nickel release.

In addition to their mechanical and functional properties, the corrosion resistance of TiNi alloys in physiological environments is a key factor at play in their long-term success as biomaterials. Although TiNi is generally considered corrosion resistant due to the formation of a TiO₂-based passive film [18], electrochemical studies in 0.9% NaCl solution (9 g L⁻¹), simulating body fluids, have shown that this alloy is still susceptible to localized corrosion, particularly pitting, under certain conditions [18,19,20]. The presence of nickel, which is essential for the shape memory effect, introduces concerns related to nickel ion release if the passive film is compromised, potentially causing cytotoxicity, allergic reactions, or inflammatory responses [21,22,23]. While electrochemical techniques such as potentiodynamic polarization and electrochemical impedance spectroscopy consistently demonstrate passive behavior and low current densities, discrepancies in the reported results often arise from differences in testing methodologies and surface quality [18,24,25]. Some representative corrosion properties of equimolar TiNi alloy in 9 g L⁻¹ NaCl are as follows. Semin et al. [26] investigated TiNi and Ti⁺-implanted TiNi sample alloys that were electrochemically polished in a mixture of acetic and perchloric acids. In 9 g L⁻¹ NaCl, they obtained open-circuit potentials of −0.111 V after 1 h of exposure. From the polarization curve extrapolation, incorrectly interpreting slopes as Tafel lines, b_c = −210 mV dec⁻¹ and b_a = 170 mV dec⁻¹; with the corrosion potentials, a corrosion current density of 0.13 μA cm⁻² is determined. The alloy shows essentially passive behavior from ~0 V to +0.250 V, with an average passive current density of 3 μA cm⁻². An increase above 0.25 V in the current density indicates the onset of pitting propagation. Ti⁺-implanted TiNi shows slightly better characteristics, with a small decrease in the corrosion current density to 0.083 μA cm⁻². Sun et al. [27] investigated porous and dense NiTi shape memory alloys and obtained a corrosion potential of −0.3 V; the corrosion current density is not reported, while the passive current density is very low, in the range of 0.1 μA cm⁻². Pitting potentials are recorded at 0.35 V. Elshaer et al. [28] compared as-cast TiNi and Ti6Al4V alloys. For the TiNi alloy, a corrosion potential of −0.34 V is observed, with a narrow passive region from −0.3 V to 0.22 V and passive current of ~0.5 μA cm⁻². Above −0.2 V, the current increased quickly until reaching a steady value of about 30 μA cm⁻², probably corresponding to the pseudopassive film. Lethabane et al. [29] investigated TiNi and Ti-Ni-Cu-Nb alloys. The corrosion potential for TiNi was −0.25 V and the corrosion current density was ~110 μA cm⁻², and, interestingly, it did not have a passive region. Generally, ongoing research is focused on manufacturing processes and defining optimal conditions in order to minimize the formation of intermetallic phases, which are considered to be the main cause of localized corrosion and surface chemical instability. This would contribute to the increase in overall biofunctionality. Literature-based evidence regarding the microstructural and chemical analysis of TiNi SMA alloys obtained via vacuum induction melting (VIM) and arc re-melting (VAR) [30,31,32] suggests that both methods meet the requirements prescribed by the ASTM-F 2063-18 standard for medical devices [33]. Still, the obtained ingots exhibit certain disadvantages, mainly concerning the high reactivity of the melt and the tendency to segregation, so re-melting is necessary. However, as the number of re-melting cycles increases, the tendency for oxygen contamination also increases [34,35].

The scattered results and limited knowledge regarding the corrosion and biocompatibility of this alloy during long-term implantation have restricted its use as an implantable material. It should also be noted that corrosion and biocompatibility are also dependent on the manufacturer and manufacturing processes. Thus, the main goal of this study is to investigate the corrosion and biocompatibility behaviors of a TiNi alloy prepared using the vacuum induction melting (VIM) method.

2. Materials and Methods

2.1. Production of the As-Cast TiNi Alloy

A TiNi alloy with a composition of 50 at.% of titanium, and 50 at.% of nickel (Merkur d.d, Celje, Slovenia) was produced using the classical process according to the ASTM F2063-18 Standard [33], i.e., re-melting four times in a vacuum furnace by induction melting (Leybold Hereaeus, Helsinki, Finland), with electro-resistive heating to reduce gas content and remove impurities, before finally casting into ingots.

2.2. Corrosion Testing

Before corrosion testing, the TiNi alloy samples were mechanically ground using SiC papers (1000–4000 grit), polished with 0.05 μm Al₂O₃ on the polishing cloth; this was followed by ultrasonic cleaning in ethanol and deionized water. Electrochemical measurements were conducted using a standard three-compartment glass cell with a volume of 100 cm³, with the TiNi sample (2 cm²) used as the working electrode, a saturated calomel electrode (SCE) as the reference, and, as the counter electrode, the lead wire inserted in the polypropylene bag to avoid hypochlorite formation. According to ISO 10271:2009 [36], the tests were performed in a 9 g L⁻¹ NaCl (Merck KGaA, Darmstadt, Germany) solution at pH = 7.4 adjusted with 0.1 M NaOH (Merck KGaA, Darmstadt, Germany), at room temperature. Open-circuit potential (E_ocp) and linear polarization measurements (±20 vs. E_ocp) were taken in order to determine polarization resistance, R_p; values were simultaneously recorded for 1 h. Then, electrochemical impedance spectroscopy (EIS) was performed at open-circuit potential, E_ocp, with a 5 mV amplitude over a frequency range of 10 kHz to 0.10 Hz. After EIS, potentiodynamic polarization was performed, which was conducted at a scan rate of 1 mV s⁻¹. All measurements were carried out using a Gamry 1010E (Gamry, Warminster, PA, USA) potentiostat/galvanostat. Surface morphology before and after testing was analyzed using a scanning electron microscope (SEM) JEOL JSM-6610LV (JEOL, Tokyo, Japan) at 20 kV, equipped with an energy-dispersive detector (EDS) model: an X-Max Large-Area Analytical Silicon Drift connected to an INCA Energy 350 (USA-Oxford Instruments, Concord, MA, USA). To identify microstructural configurations, the sample was prepared according to the protocol for metallographic preparation: grinding, polishing with a diamond paste suspension of 0.05 μm, and cleaning with deionized water in an ultrasonic bath. The prepared surface was chemically etched with 3 mL HF, 6 mL HNO₃, and 100 mL of water for 45 s. The optical image was taken using a Nikon Epiphot 300 (Nikon, Tokyo, Japan) metallurgical microscope. For the determination of the grain area distribution, ImageJ software (version 1.5t, National Institutes of Health, LOCI, University of Wisconsin, Madison, WI, USA) was used.

2.3. Biocompatibility

2.3.1. Cell Cultivation

Human osteoblast-like cells (MG-63, ATCC^® CRL-1427™) were grown in a complete growth medium: Dulbecco’s modified Eagle’s medium (DMEM) with 4 mM L-glutamine, 10% Fetal bovine serum (FBS), and 1% Antibiotic Antimycotic (ABAM) (all from Sigma-Aldrich, Steinheim, Germany). Cells were cultured in 5% CO₂ at 37 °C in a humidified atmosphere and regularly passaged at 80% confluence.

2.3.2. Cell Seeding on Samples

TiNi discs (number of samples, n = 6, 2r = 9 mm) were compared with control samples of round microscopy cover glass (n = 6, 2r = 9 mm). Before the cell seeding, all samples were sterilized with UV-C light for 1 h and placed in a 12-well culture plate. MG-63 cells (2 × 10⁴ cells/cm²) were seeded directly onto the samples and incubated at 37 °C in a humidified 5% CO₂ atmosphere. For viability assay, the cells were cultured for 24 h, and cell adhesion was evaluated using a fluorescent microscope after 168 h (7 days), and, for SEM, after 24 h and 168 h.

2.3.3. Cell Viability Assay

Mitochondrial activity was assessed using a direct contact test following ISO standard 10993-5:2009 [37]. After 24 h of cell culturing on the samples, metabolic activity was evaluated using the MTT (3-([4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide)) assay (Sigma-Aldrich, St. Louis, MO, USA). To measure cell metabolism, the growth medium was removed from the wells, and a fresh growth medium containing 5 mg/mL MTT was added. The cells were incubated for 4 h in the dark at 37 °C. After incubation, the medium was discarded, and 100 μL of dimethyl sulfoxide, DMSO (Sigma-Aldrich, St. Louis, MO, USA) was added to dissolve the tetrazolium salts from the mitochondria. Absorbance was then measured at 540 nm using an enzyme-linked immunosorbent assay (ELISA) microplate reader (RT-2100c, Rayto, Shenzhen, China).

2.3.4. Preparation of Samples for Fluorescent Microscopy

After washing with phosphate-buffered saline (PBS) to remove non-adherent cells, the samples were stained with Hoechst 33,342 (2 µg mL⁻¹) for 20 min. This dye selectively stains viable cell nuclei, emitting blue fluorescence. Imaging was performed using an Axio Imager Z1 epifluorescence microscope (Carl Zeiss, Oberkochen, Germany), and viable cells were quantified using ImageJ software, IJ 1.46r (Bethesda, Rockville, MD, USA). For each sample type, three wells were analyzed, with cell counts performed over a minimum surface area of 2.5 mm² per well.

2.3.5. Preparation of Samples for SEM Analysis

The morphology of MG-63 cells attached to the samples was analyzed using a JEOL JSM-6500F scanning electron microscope (Tokyo, Japan). Sample preparation, including fixation, contrast enhancement, and dehydration, was performed following previously established protocols [38]. To ensure optimal imaging, the samples were coated with a thin gold layer using a Precision Etching Coating System (682 PECS, Gatan, Pleasanton, CA, USA). The analysis focused on the morphology and adhesion patterns of MG-63 cells on both control samples (cover glass) and TiNi discs.

2.3.6. Inductively Coupled Plasma-Optical Emission (ICP-OES)

A static immersion test was used to provide quantitative data on the nickel and titanium ions released from the specimens’ surfaces into the DMEM solution. Before testing, TiNi discs were prepared according to ISO 10271:2009 for testing materials in dentistry [36]. Specimens were immersed in 5 mL of each test solution in a test tube clogged with a rubber cork using a nylon string. The test durations were 24 h and 48 h at a constant temperature of 37 ± 0.2 °C. In parallel with the samples, the pure solution “zero samples” were treated in the same way. The concentration of nickel and titanium ions (µg L⁻¹, ppb) in the obtained solutions after the static migration test was measured using the analytical technique of Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES). The ICP-OES measurement was performed on a Thermo Scientific iCAP 6500 Duo ICP instrument (Thermo Fisher Scientific, Cambridge, UK) using iTEVA operating software (Issue 9.8, Thermo Fisher Scientific, Cambridge, UK). The sample was introduced into the plasma via the direct aspiration of the liquid sample. Standard solutions for instrument calibration within appropriate concentration ranges were prepared from certified standard solutions: Multi-Element Plasma Standard Solution 4, Specpure^®, 1000 µg/mL, and, for titanium, a standard plasma solution, Specpure^®, Ti 1000 µg/mL (supplied by Alfa Aesar GmbH & Co KG, Karlsruhe, Germany). Quantification was performed at the following emission lines: Ni II 231.604 nm and Ti II 336.121 nm. The correlation coefficients were greater than 0.9999. For each sample, ICP-OES measurements were conducted in triplicate (n = 3). The relative standard deviation (RSD) of repeated measurements was <5%. The limit of detection (LOD) was 0.12 and 0.20 µg L⁻¹, while the limit of quantification (LOQ) was 0.41 and 0.48 µg L⁻¹ for nickel and titanium, respectively. The analytical process quality control (QC) was performed using a certified reference material (CRM): EPA Method 200.7 LPC Solution (ULTRA Scientific, Santa Clara, CA, USA). The achieved agreement with certified values (recovery) for all certified elements was in the range of 98–103%. The concentrations of nickel and titanium ions in the solution are presented in ng mL⁻¹.

3. Results and Discussion

3.1. SEM and EDS Characterization of the TiNi Alloy

Figure 1 shows SEM images of the as-prepared TiNi alloy. In Figure 1a, a relatively homogenous structure can be observed, with some small inclusions of shapeless structures. Similar results for the TiNi alloy have been obtained by many authors [37,38,39,40] who identified inclusions as the Ti₂Ni phase. The results revealed that the TiNi alloy is a hypoeutectic system, consisting of a cubic TiNi matrix and a cubic Ti₂Ni phase. It should be pointed out that inclusions, dispersed in the bulk material, are generally Ti₂NiO_x and eventually form TiC, displaying micrometric sizes that are responsible for pitting or breakdown potential [41].

To determine the composition of the alloy matrix and inclusions, a higher magnification is used, as shown in Figure 1b, and EDS is performed, marked in the figure as spectra 1 and 2.

According to the EDS spectra presented in Figure 2, obtained from the regions marked in Figure 1b, the TiNi matrix (spectrum 1) exhibits a composition of 54 wt.% Ni and 42 wt.% Ti. Converting to at.% composition is Ti₄₉Ni_51, which is practically identical to the TiNi matrix. The composition of the inclusion (spectrum 2) has 60 wt.% of Ti and 37% wt.% of Ni, or 66.5 at.% of Ti and 33.5 at.% of Ni, which corresponds to the cubic Ti₂Ni phase. Based on the phase diagram of TiNi alloys, the cubic Ti₂Ni phase is formed during alloy cooling at temperatures above 800 °C in the Ti-rich melt zones. According to the data in the literature [42], small amounts of this phase ensure the alloy’s ductility. Specifically, during the additional cooling of the melt, the zones of the Ti₂Ni phase represent the centers of nucleation of martensite grains, the carrier of the shape memory effect. Faster cooling of the alloy favors the formation of this phase. However, because this phase has a higher hardness than the TiNi phase that forms the alloy’s matrix, large amounts of Ti₂Ni precipitates can directly reduce the superelastic properties of the alloy. The presence of a small amount of carbon could be attributed to contamination of the surface by air impurities. As well as vacuum induction melting, the most commonly used method of obtaining nickel–titanium alloys is metal melting in arc furnaces (vacuum arc remelting). A graphite crucible furnace is the best choice for melting pure metals to prevent possible melt contamination [43,44,45], but some impurities are possible due to carbon pick-up.

An optical microscopy image of chemically etched TiNi alloys is shown in Figure 3a; it shows a predominant equiaxed structure, with irregular grain sizes and shapes. The area of the grains is in a broad range from ~100 μm² up to ~6500 μm^2, with the main average area of 416 μm². To estimate grain dimensions, in the first approximation, we considered the grains to have an elliptical shape. Knowing that the area of an ellipse is p = a × b × π, where a and b are the radius, we can assume that b = 0.5a, and, by conducting simple calculations for the average area of ~416 μm², the average grain size is determined to be around 32 μm × 16 μm. In addition, the grain area is mostly below 2000 μm², as shown by the inset in Figure 3b, or lower than ~70 μm × 35 μm.

3.2. Corrosion Behavior

Figure 4 shows the dependence of the open circuit potential, E_ocp, and polarization resistance, R_p, obtained from the linear polarization measurements over time. After immersion, open-circuit potentials increase from −0.443 V to −0.428 V, while polarization resistance increases from 62 kΩ cm² to 72 kΩ cm². The observed increase in both values indicates that the naturally formed thin film on the alloy is probably porous, with the pores being progressively passivated over the course of immersion.

Figure 5 presents the cyclic polarization curve of the TiNi electrode starting from a cathodic potential of −0.65 V. The cathodic branch is characterized by a high slope of −193 mV dec⁻¹ that cannot be connected to Tafel behavior. It is most plausible that such a high slope is linked to oxygen reduction under mixed activation–diffusion control and surface film composition change. It is necessary to note that, at E_ocp, there is no change in the cathodic to anodic current densities of the polarization curve, suggesting that this spontaneous potential is the quasi-corrosion potential of the alloy. Similar results were obtained by Gaber et al. [46] for the corrosion of AISI 201 austenitic stainless steel in 1 M H₂SO₄. Namely, the open circuit potential was −0.1 V while the corrosion potential determined from the polarization curves was approximately +0.25 V, but this phenomenon is not explained. Consequently, at E_ocp, it can be suggested that the following reaction occurred. During the immersion period, increases in R_p and E_ocp could be connected with the dissolution of Ti₂Ni as an energetically unstable state that behaves as an anodic place and an oxygen reduction reaction at the TiNi matrix as a more stable phase. The overall reaction is:

Ti₂Ni + 4OH⁻ = TiNi + TiO₂ + 2H₂O + 4e⁻

(1)

It proceeded via the following steps:

Ti₂Ni = TiNi + Ti⁴⁺ + 4e⁻

(2)

Ti⁴⁺ + 4OH⁻ = TiO₂ + 2H₂O

(3)

and:

O₂ + 2H₂O = 4OH⁻ + 4e⁻

(4)

Hence, during the immersion, contact corrosion (between TiNi and Ti₂Ni inclusions) probably takes place, which leads to the passivation of Ti₂Ni, and, after a certain time, this reaction could be neglected.

At the potential of −0.2 V, a change from cathodic to anodic current densities arose that could be connected with the real corrosion potential E_corr, at which point the alloy matrix started to oxidize, followed by passivation. Extrapolating the cathodic line to corrosion potential, we estimate the corrosion current density, j_corr, of around 60 nA cm⁻² or less. The passive behavior occurred in the potential range from −0.1 V to 0.22 V with an average passive current density j_pas of 0.2 μA cm⁻², followed by metastable pitting. At a potential of 0.34 V, the critical pitting potential, E_pit, is observed to be connected with the sharp increase of the current density. A shoulder at 0.36 V is also visible, which could be attributed to the transpassive potential, E_tp, accompanied by a change in the metal oxidation state, most likely from Ni²⁺ to Ni³⁺. In the reverse scan, positive hysteresis is observed, connected to pit growth. The repassivation potential, E_rp, is recorded at ~0 V, followed by an oxygen reduction reaction. In some studies, the numerical values of pitting corrosion resistance, R_pit, are calculated as: R_pit = ΔE = |E_corr − E_pit|. The difference between E_corr and E_pit can be used to evaluate a material’s vulnerability to pitting. A small ΔE indicates a greater susceptibility; therefore, the metal is prone to pitting corrosion even at potentials near to the corrosion potential [47].

In

Table 1. The characteristic corrosion parameters for TiNi in 9 g L⁻¹.

Eocp V	Rp kΩ cm2	jocp μA cm−2	Ecorr V	jcorr nA cm−2	jpas μA cm−2	Epit V	Erp V	Rpit V
−0.430	72	~2	−0.2	60	0.2	0.34	~0	0.540

, due to the better comparisons, characteristic corrosion parameters obtained from the polarization curves and linear polarization measurements are summarized.

In order to examine the corrosion processes, and the plausible oxide nature on the surface, and/or does the repassivation after pitting corrosion occurs, before and after cyclic polarization measurements, the electrochemical impedance is measured. The results, presented as Nyquist and Bode plots, are shown in Figure 6a,b. The interpretation of the obtained impedance spectra is explained based on the above discussion and data from the literature [48,49,50]. The Nyquist plot shows that impedance is practically a straight line, with slopes of 51° before and 57° after polarization, which are higher than 45° (characteristic of pure diffusion control) and lower than 90° (characteristic of pure capacitive behavior). It should be noted that the open-circuit potentials, shown in the inset in Figure 6a, are very close to the potential determined during the linear polarization measurements, as shown in Figure 4. In the Bode plot, in the frequency range from 10,000 to ~200 Hz, the horizontal line that suggests resistance behavior cannot be seen. The slope of the lines of |Z_mod| vs. log f is ~0.67, indicating a complex mixture of capacitive–diffusion behavior, suggesting the usage of the constant phase elements (CPE) for fitting. Decreasing the frequency, the phase reaches the maximum at ~40 Hz with a phase angle of 68°, followed by slow decreases in the phase angle and stabilization around 55° at frequencies lower than ~2 Hz. This behavior could be attributed to surface phenomena, e.g., passivation. The spectra after polarization in the Nyquist plot have a higher slope than before polarization, and the phase peak is slightly shifted to lower frequencies, suggesting that the passive layer is disturbed but undergoes slower recovery.

Based on explanations of the electrochemical impedance behavior, the physical model and the equivalent electrical circuit (EEC) are shown in Figure 7a. The elements of R_f and C_f correspond to film oxide resistance and capacitance. CPE_dl is assigned to the double-layer capacitance, due to the inhomogeneity of the oxide surface. The impedance of a constant phase element (CPE) is described by the formula:

$$Z_{\mathrm{CPE}}={\displaystyle \frac{1}{Y{(j\omega )}^n}}={\displaystyle \frac{1}{Y{(i2\pi f)}^n}}$$

(5)

where Y is the CPE admittance, n is the CPE exponent (a value between 0 and 1), i is the imaginary unit, i² = −1, and ω is the angular frequency (ω = 2πf). R_ct and CPE_w are connected with charge transfer reaction, and diffusion (using the pure Warburg element, the fit was not good), and R_s solution resistance. The presented physical model of the EIS can explain the difference between open circuit and corrosion potential, Figure 5. The inclusion of Ti₂Ni is a Ti-rich phase with a lower local surface potential and probably a different and thinner oxide than the TiNi matrix, which can undergo slow dissolution via reaction 1 and form some kind of oxide. That oxide passivizes the inclusion over time. Thus, the increased slope of the cathodic line could be explained by the sum of passive and oxygen reduction current densities. Due to the relatively small current densities of 2–3 μA cm⁻², at open-circuit potential, the reaction given by Equations (2) and (4) is most probably under diffusion or mixed activation–diffusion control because the concentration of Ti⁴⁺ and dissolved oxygen is very small. During this time, the reactions become rather slow but sufficient to establish E_ocp. Due to the cathodic polarization, Ti₂Ni is additionally covered by an oxide layer, and, only at E_corr, the further oxidation of the TiNi matrix occurs, followed by a passive state.

The fitted parameters of the equivalent electrical circuit are summarized in

Table 2. The fitted parameters of the equivalent electrical circuit, shown in Figure 6a, (1) before and (2) after cyclic polarization. G.F represents the “goodness of fit”.

	Eocp V	Rsl Ω cm2	Rf Ω cm2	Cf μF cm−2	Yo,W μS sn cm−2	n	Rct kΩ cm2	Yo,dl μS sn cm−2	n	G.F 10−3
1.	−0.425	4.4	31.2	15.1	67	0.54	2.7	28.5	0.86	0.097
2.	−0.462	4.2	29.4	3.75	50	0.55	2.0	48	0.83	0.083

. Based on the values of the “goodness of fit” lower than 0.1 × 10⁻³, it can be concluded that the experimental EIS data are successfully interpreted using the chosen equivalent circuit model.

From the obtained fitted data for admittance Y_0,dl, Ω⁻¹cm⁻² sⁿ, and n from CPE_dl, the double-layer capacitance, C_dl, can be estimated according to:

$$C_{\mathrm{dl}}={(Y_{0,\mathrm{dl}}{R}_s^{-(n-1)})}^{{\displaystyle \frac{1}{n}}}$$

(6)

Consequently, before polarization, C_dl is 6.6 μF cm⁻², and, after polarization, it is 4.7 μF cm⁻². The Warburg diffusion element (Z_W) is a constant phase element (CPE), with a constant phase of 45° (phase independent of frequency) or n ≈ 0.5. In an ideal case, it has a magnitude inversely proportional to the square root of the frequency:

$$\left|Z_W\right|={\displaystyle \frac{1}{{\mathsf{{\rm Y}}}_{0.\mathrm{W}}\sqrt{2\pi f}}}+{\displaystyle \frac{1}{{\mathsf{{\rm Y}}}_{0.\mathrm{W}}i\sqrt{2\pi f}}}$$

(7)

The impedance module of Warburg impedance can be presented as:

$$\left|Z_W\right|={\displaystyle \frac{\sqrt{2}}{{\mathsf{{\rm Y}}}_{0.\mathrm{W}}{\left(2\pi f\right)}^n}}$$

(8)

This represents the total opposition to the flow of current in the circuit, considering both resistance and reactance. In Equation (8), Υ_0,W is the Warburg coefficient (or Warburg constant) that represents admittance. To calculate |Z_w|, the frequency must be determined. Because R_ct of ~2 kΩ cm² is much smaller than R_p = 72 kΩ cm², as obtained from linear polarization measurements (Figure 4), it is not sensitive to the reaction control (solution resistance, charge transfer resistance, diffusion). As the R_p is measured ±20 mV around E_ocp with a sweep rate of 1 mV s⁻¹, the total duration is 40 s or ~0.02 Hz. Hence, the |Z_w| is ≈65 kΩ cm², which is already a satisfactory agreement. However, the R_p is the sum of all parameters:

R_p = R_s + R_ct + |Z_w|

(9)

this gives an R_p of 68 kΩ cm²: practically identical values.

If the overpotential, η, is very small (±5 mV) and the electrochemical system is at equilibrium, the classical Stern–Geary equation for the charge-transfer resistance can be represented with the equation:

$$j_{\mathrm{ocp}}={\displaystyle \frac{RT}{nF{R}_{\mathrm{ct}}}}$$

(10)

As shown in Table 2, R_ct is 2700 Ω cm² and 2000 Ω cm²; consequently, the corresponding open-circuit current densities, taking into account 4e⁻ exchange, are 2.3 μA cm⁻² and 3.2 μA cm⁻², respectively, in excellent agreement with the experimentally observed values shown in Figure 4. Moreover, given that [51,52] the thickness, d, of naturally formed oxide on Ti and its alloys, is ~5 nm (5 × 10⁻⁷ cm), it is plausible to calculate the specific resistance of the oxide film, ρ, Ω cm, from the R_f = 31.2 Ω cm², according to:

$$\rho ={\displaystyle \frac{R_{\mathrm{f}}S}{d}}$$

(11)

where S = 1 cm and ρ is in the order of 62 MΩ cm, which corresponds to semiconductors.

After the cyclic polarization, the SEM and EDS are taken, and the results are shown in Figure 8, while EDS analysis from the marked spectra is summarized in

Table 3. Summarized EDS analysis of the spectra marked in Figure 8.

Spectra	Ti wt.%	Ni wt.%	C wt.%	Si wt.%	Al wt.%	Cl wt.%	Ti at.%	Ni at.%	Phase
S1a	42.4	52.9	4.6	0	0	0	50.6	49.4	TiNi
S2a	57.2	38.3	3	1.5	0	0	64.7	35.3	Ti2Ni
S2b	42.5	54.2	3.3	0	0	0	49.0	51.0	TiNi
S1c	41.2	52.7	3.6	0	0	0	48.9	51.1	TiNi
S1c	42.1	54.3	5.4	0	0.3	0	48.7	51.3	TiNi
S1d	37.5	50.1	7.6	0	0.5	0.2	47.8	52.2	TiNi
S2d	42.5	54.0	3.4	0	0	0	49.0	51.0	TiNi

. In Figure 8a, the overall composition of the TiNi matrix is shown, and Table 3 shows that the total Ni content slightly decreases from the spectra taken before polarization. Therefore, Ni may be slightly dissolved, according to the Pourbaix diagram, via the following reactions:

Ni = Ni²⁺ + 2e⁻

(12)

Ni²⁺ + 2OH⁻ = Ni(OH)_2,sln

(13)

This is because of the low Ni²⁺ concentrations and solubility product of 2.0 × 10⁻¹⁵ M⁻², meaning that Ni(OH)₂ is slightly soluble in water. From Figure 6a, it is also possible to see that pitting corrosion occurred, accompanied by micro-galvanic corrosion. The surface does not undergo severe corrosion, and most of the surface is intact (Figure 8b); the existence of Ti₂Ni inclusions is evident (spectrum 1b). It is interesting that, in some places, a hole is formed with intrusions inside the hole (Figure 8c,d). Based on the EDS spectra, the intrusions have the composition of an alloy matrix, e.g., TiNi. This can be explained by the work of Gao et.al. [53]. Namely, the solidified NiTi phase develops swiftly as the temperature of the molten alloy decreases, taking the shape of either columnar or equiaxed grains. As the temperature of the molten pool decreases further to ~980 °C, the Ti-rich molten phase that surrounds the solid NiTi phase experiences a peritectic reaction, transitioning from NiTi + L(Ti) to Ti₂Ni, causing the precipitation of Ti₂Ni around the NiTi grains. Consequently, it seems that the main corrosion mechanism is the contact corrosion of Ti₂Ni as an anodic site, in contact with TiNi intrusions and the surroundings TiNi matrix that represents cathodic sites.

The EDS analysis of wt.% of the SEM spectra shown in Figure 8a–d is summarized in Table 3. In addition to carbon, whose origin is explained in the above text, some small amounts of Si and Al are detected, probably from the SiC grinding paper employed, as well as Al₂O₃ from polishing. Moreover, at spectrum S1d, 0.2 wt.% of Cl is detected, which is contamination from the electrolyte. We can assume that C, Si, Al and Cl, are impurities, so only wt.% of Ti and Ni are recalculated to at.% and the corresponding phase of the TiNi alloy is given.

In the following section, some comparisons of our results and the data for different methods of preparing TiNi alloys available in the literature are given. Sun et al. [26] investigated the corrosion behaviors of porous TiNi and compared them with the dense TiNi alloy prepared in a non-consumable vacuum arc-melting furnace. The corrosion current densities, j_corr, of porous NiTi alloys are markedly higher than the values of the dense ones, and the breakdown potentials of porous NiTi are in the range of 0 V to 0.1 V, much lower than the values of the dense samples at ~0.35 V. The porous TiNi essentially did not show a passive region, while the current density of the dense sample was in the range of 10 nA cm⁻² to ~0.5 μA cm⁻². Recently, laser powder bed fusion (LPBF) has been used to produce TiNi alloys [54]. The breakdown potential in SBF is ~0.45 V, with a passive current density higher than 1 μA cm⁻². Regarding the corrosion behavior in 9 g L⁻¹ of the newly developed method of producing TiNi alloy, using continuous casting [23], the corrosion current density is roughly estimated to be 6 μA cm⁻², with a very limited passive region of ~0.15 V and a breakdown potential at ~0.05 V.

Even the examined TiNi alloy shows similar and, in some cases, better corrosion behavior than the alloys described in the Introduction [26,27,28,29] and the above text [23,54]. The future optimization of the cooling processes should be performed to minimize the formation of Ti₂Ni phases. In addition, the anodization of the TiNi alloy in suitable electrolytes should also be investigated in order to increase corrosion stability.

3.3. Biocompatibility of the TiNi Alloy

3.3.1. Cell Viability and Adhesion

A colorimetric assay, which was used to measure cell viability and proliferation by assessing cellular metabolic activity (MTT), revealed no significant change in cell viability when cultivated on TiNi samples, in comparison to the control samples (98.6 ± 0.5% in control; 97.1 ± 0.3% for TiNi), as shown in Figure 9.

The results of fluorescence microscopy showed that the TiNi sample is not cytotoxic, as the density of attached cells is not statistically significantly lower in the TiNi samples than in the control, as shown in Figure 10.

After 24 h of culturing on control and TiNi specimens, the cell morphology was characterized by polygonal shapes and the presence of cytoplasmic extensions (Figure 11). At higher magnifications, the SEM images clearly show cytoplasmic extensions, indicating the cells’ ability to interact with and spread on the substrate surface. Cells extend their cytoplasm, forming long, thin projections that help them interact with and adhere to the surrounding surface.

At 168 h of exposure in both groups, cells with well-defined cellular and membrane outgrowths are predominant. Notably, the cells spread out and cover the surface of the materials, a sign of cell activity and integration with the environment (Figure 12). However, the cell morphology differed slightly between the control and TiNi specimens. Cells on the control specimens exhibited an elongated shape with multiple filopodia firmly attached to the disc surface. In contrast, cells on the TiNi specimens are more rounded, with a smaller cell footprint but large filopodia. Both specimen types showed numerous cellular interconnections, indicating a high degree of cell–cell contact. No cells with apoptotic or necrotic morphology were observed at 168 h.

3.3.2. Inductively Coupled Plasma Optical Emission Spectroscopy

Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) elemental analysis was conducted to examine the substitution of nickel and titanium ions in the DMEM solution. The results revealed that, during 24 h of immersion, the migration of Ni ions was slightly higher than that of titanium ions (

Table 4. The total concentration of nickel and titanium released into the DMEM solution (in ng/mL).

Solution	Nickel (Ni)	Titanium (Ti)
DMEM	<0.5	<0.5
TiNi 24 h	1.05 ± 0.28	0.84 ± 0.46
TiNi 48 h	<0.5	1.74 ± 0.81

Note: the ng/mL concentration is equivalent to µg/L (ppb) concentration.

). On the contrary, after 48 h, the concentration of titanium ions released exceeded that of nickel ions.

Regarding the analysis of individual ion release (Ni and Ti), the results indicated that nickel ion release is initially higher than that of titanium, which is consistent with previous studies [55,56]. This can be attributed to the dissolution of the TiO₂ surface layer, exposing the underlying layer, which is rich in nickel [57]. Additionally, the lower stability of nickel oxide contributes to its faster degradation compared to titanium oxide. The critical concentration of nickel that is necessary to induce allergies is 600–2500 μg, and the daily dietary intake level is 300–500 μg [58]. The result of the static test performed (TiNi alloy, 24 h, 5 mL) shows that a total of 0.0053 µg of nickel migrated into the solution. This is significantly lower than the daily nickel intake from food, as well as being below the allergenic threshold for nickel. Simultaneously, a total of 0.0042 µg of titanium migrated into the solution in 24 h and 0.0087 µg in 48 h. Previous studies found that the maximum rate of Ni ions release occurred at the beginning of the test, before the rate reduced dramatically due to the thickening of the oxide layer, which serves as a protective barrier. Numerous studies have shown that the cell culture medium facilitates the interaction of Ni and Ti with oxygen, leading to the formation of surface oxides on the nickel–titanium alloy [59]. It should also be pointed out that OH⁻ is one of the toxic reactive oxygen species (ROS) related to pathological cell death. Hydroxyl anions are compounds that perforate the cell membrane and initiate lipid peroxidation, which is a crucial mediator in the mechanism of cell apoptosis and autophagy [60]. Several key damage mechanisms are promoted through the peroxidation of phospholipid bilayers: DNA chain damage, stress of the endoplasmatic reticulum, and mitochondrial apoptosis [60,61].

3.3.3. Cytotoxicity of TiNi

The cytotoxicity of the material’s surface is evaluated through direct contact with human-derived osteoblast-like cells. The findings of this study indicate that mitochondrial activity in cells cultured directly on the tested samples was slightly lower compared to control cells; however, this difference is not statistically significant. Although cell proliferation decreased over the seven-day exposure period, the results did not indicate any subtoxic or lethal effects in standard in vitro conditions ISO 10993-5:2009 [37]. Similar results were obtained in a recent study by Ibrahim et al. [62], with no significant difference in mitochondrial activity when human osteoblasts and cardiac cells were grown on a cast TiNi alloy. Even though cell proliferation is unimpaired, the SEM micrographs, shown in Figure 11, revealed a more rounded morphology of MG-63 cultured on experimental samples on day seven. In the experimental group, cells appear more rounded, likely due to interactions between cells and the material’s surface properties. TiNi alloys form a passive oxide layer in the culture medium over time, which may impair cell adhesion and spreading, leading to a rounded morphology [63]. The passive layer with a duplex structure and porous outer layer may affect the cell’s adhesion and spread through the material surface. The corrosion tests indicated that Cl⁻ rich solutions can promote the formation of hydroxyl anions on the surface of the TiNi alloy, which may lead to adverse clinical effects under in vivo conditions. The observed change in cell morphology, characterized by a more rounded appearance of cells cultured on NiTi samples, as seen under scanning electron microscopy, may indicate increased surface roughness or potential nickel ion release [64].

Accordingly, the study conducted by Ryhänen et al. [57] showed the satisfactory biocompatibility of nitinol alloy composed of 54 wt% nickel and 46 wt% titanium in contact with osteoblasts. When the cells were in direct contact with the alloy produced using the vacuum melting, no inhibitory effect on cell growth was observed. Similar findings were reported by other authors [65] who investigated the biocompatibility of alloys produced using the hot isostatic pressing (HIP) and cold isostatic pressing (CIP) methods. The results showed the good biocompatibility of the Ti-rich alloy, which contains at least 50 wt% titanium. In contrast, the increase in nickel content had a different effect. Alloys with a nickel concentration exceeding 60 wt% caused cells to become detached from the alloy’s surface due to the toxic effects associated with the nickel-rich surface.

Based on all of the above findings, future research should analyze the biological characterization of the alloy under different conditions that mimic the wider human biological environment, such as interactions with other cells, depending on the future purpose of the product.

4. Conclusions

Summarizing the main findings, the following can be concluded:

• As-cast TiNi is a two-phase system. Apart from the primary TiNi phase matrix, the inclusion of an additional Ti₂Ni cubic phase is detected.
• TiNi has acceptable corrosion stability. In addition to its moderately low corrosion potential and pitting corrosion at 0.34 V, the dissolution of the unstable Ti₂Ni inclusions is also responsible for the breakdown potentials.
• Even though the examined TiNi alloy shows good corrosion behavior, the future optimization of the cooling processes should be performed to minimize the formation of Ti₂Ni phases.
• The nickel ions’ release concentration was lower than the daily nickel intake from food and below the allergenic threshold. However, the Ni²⁺ release was higher than that of the Ti ion, indicating the instability of the protective surface oxide.
• The MTT assay and fluorescence microscopy analysis showed that cell viability and adhesion on TiNi samples are not significantly impaired compared to the control sample. Despite the well-preserved cell morphology cultured on the TiNi alloy, subtle changes in the form of a spherical shape were detected, which may suggest greater surface roughness or the possible release of nickel ions.
• Further improvement in the corrosion properties could be achieved by anodization.