Experimental Investigation of the Biofunctional Properties of Nickel–Titanium Alloys Depending on the Type of Production

Nickel–titanium alloys used in dentistry have a variety of mechanical, chemical, and biofunctional properties that are dependent on the manufacturing process. The aim of this study was to compare the mechanical and biofunctional performances of a nickel–titanium alloy produced by the continuous casting method (NiTi-2) with commercial nitinol (NiTi-1) manufactured by the classical process, i.e., from remelting in a vacuum furnace with electro-resistive heating and final casting into ingots. The chemical composition of the tested samples was analyzed using an energy dispersive X-ray analysis (EDX) and X-ray fluorescence (XRF). Electron backscatter diffraction (EBSD) quantitative microstructural analysis was performed to determine phase distribution in the samples. As part of the mechanical properties, the hardness on the surface of samples was measured with the static Vickers method. The release of metal ions (Ni, Ti) in artificial saliva (pH 6.5) and lactic acid (pH 2.3) was measured using a static immersion test. Finally, the resulting corrosion layer was revealed by means of a scanning electron microscope (SEM), which allows the detection and direct measurement of the formatted oxide layer thickness. To assess the biocompatibility of the tested nickel–titanium alloy samples, an MTT test of fibroblast cellular proliferation on direct contact with the samples was performed. The obtained data were analyzed with the IBM SPSS Statistics v22 software. EDX and XRF analyses showed a higher presence of Ni in the NiTi-2 sample. The EBSD analysis detected an additional NiTi2-cubic phase in the NiTi-2 microstructure. Additionally, in the NiTi-2 higher hardness was measured. An immersion test performed in artificial saliva after 7 days did not induce significant ion release in either group of samples (NiTi-1 and NiTi-2). The acidic environment significantly increased the release of toxic ions in both types of samples. However, Ni ion release was two times lower, and Ti ion release was three times lower from NiTi-2 than from NiTi-1. Comparison of the cells’ mitochondrial activity between the NiTi-1 and NiTi-2 groups did not show a statistically significant difference. In conclusion, we obtained an alloy of small diameter with an appropriate microstructure and better response compared to classic NiTi material. Thus, it appears from the present study that the continuous cast technology offers new possibilities for the production of NiTi material for usage in dentistry.


Introduction
Shape memory alloys (SMAs), thanks to their unique thermomechanical function, belong to the group of functional materials that is manifested through the effects of shape memory and the superelasticity. Based on the primary alloying elements, shape memory alloys can be classified into three groups: Cu-based, Fe-based, and nickel-titanium (commercial name nitinol) alloys. From these, nitinol is particularly popular in dental devices

Sample Preparation
The materials used in the present study were nickel-titanium alloys, divided into two groups of samples. The first group, named NiTi-1 samples, consisted of commercially available market samples purchased from Merkur d.d. (Celje, Slovenia). Alloys were manufactured by the classical process, i.e., from remelting in a vacuum furnace with electro-resistive heating and a final casting into various ingots (as written on the supplier's statement).
The second group of samples, NiTi-2, were produced by continuous casting in the form of rods, according to the briefly described methodology [17]. The experimental device for obtaining NiTi-2 alloys consists of a vacuum induction melting (VIM) furnace and a vertical continuous caster. The appropriate amount of solid alloy (about 20 kg) was put into a crucible and melted in a vacuum induction furnace (10-2 mbar). VIM was performed at a medium frequency (8000 Hz). The temperature of casting was around 50 • C above the temperature of melting for NiTi alloys, which is 1350 • C. The melt continuously flowed under an Ar atmosphere into the mold cooled by water where it solidified.
All samples were cut by electro-erosion (EDM): (i) the NiTi-1 was cut into square shapes (with slide length 10 mm, and thickness 1.2 mm)and (ii) the NiTi-2 into shaped discs (ø = 11 mm, thickness 1.6 mm) suitable for further metallographic preparation and testing. The metallographic preparation included mounting in a hot-mounting mass and grounding with abrasive paper in grades of 180-4000 on the grinding/polishing machine BUEHLER Automet 250, and EcoMet 250, (Lake Bluff, IL, USA). Samples were polished on the same devices with a Naples cloth and 1µm polishing suspension. After polishing, samples were cleaned with acetone, alcohol, and deionized water in ultrasound. For microstructure observations all samples were etched according to the protocol (ASTM Standard E 407, number 192 "Kroll's") as written below: NiTi-1-3 mL HF, 6 mL HNO 3 , 100 mL-etching time 30 s; NiTi-2-3 mL HF, 6 mL HNO 3 , 100 mL-etching time 30 s.
A metallographic characterization was obtained to determine the NiTi grain size.

The Determination of Chemical Composition and Phase Distribution by Scanning Electron Microscopy (SEM)
Qualitative and quantitative analyses of the samples' composition of NiTi-1 and NiTi-2 were performed with X-ray fluorescence (XRF) analysis, using Thermo Scientific Niton XL3t GOLDD equipment.
A scanning electron microscope (SEM), Sirion 400NC (FEI, Hillsboro, OR, USA), with an energy-dispersive X-ray spectroscope (EDX), INCA 350 (Oxford Instruments, Oxford, UK), was used for the detailed microchemical analyses. The SEM/EDX measurements were as follows: (i) For NiTi-1: In the central part of the sample, two segments were chosen. In the first segment, the measurement was performed at five measuring points, and in the second segment at four points. In the marginal part of the sample, measurement was performed at four points of the measuring segment. (ii) For NiTi-2: In the central part of the sample three measuring segments with three measuring points were chosen. In the marginal part there were two measuring segments, one with four and one with two measuring points.
Detailed microstructure investigation (the visibility of the different phases of NiTi-1 and NiTi-2 samples) was performed using SEM thermal field emission (SEM JEOL JSM-6500F), equipped with electron backscatter diffraction (EBSD) analytical techniques (IMT, Ljubljana, Slovenia). Secondary-electron images and backscattered electron images were recorded at different magnifications and SEM working parameters of a 15 kV voltage, 7 nA probe current and 10 mm working distance.

Hardness
In order to determine the mechanical properties, the hardness of the samples was measured by the static Vickers method. Indentation hardness was measured due to the procedure corresponding to the ISO 6507-1:2018 Standard. The hardness, HV50, was measured on each specimen from each experimental group on a WPN HPO 250 machine that applied the nominal value of a 49.03 N test force load for 15 s. In this method, a diamond tip in the form of a regular four-sided pyramid was used as an indenter, in which the opposite sides overlap the angle of (136 + 1) • . An overview of the hardness imprints was examined on a Nikon EPIPHOT 300 microscope (Mellvile, NY, USA).

Immersion Testing and ICP-MS Analysis
A static immersion test was conducted to investigate the potential release of metal ions from both NiTi-1 and NiTi-2 samples' surfaces. Two different mediums with different pH values were used: a solution of artificial saliva (with a composition of 1.5 gL −1 KCL; 1.5 gL −1 NaHCO 3 ; 0.5 gL −1 NaH 2 SO 4 × H 2 O; 0.5 gL −1 KSCN; and 0.9 gL −1 lactic acid) with a 6.5 pH value was used, corresponding to the oral environment. In order to simulate a corrosive environment, a solution containing 5.85 gL −1 NaCl + 10 g L −1 lactic acid was also used, with the pH adjusted to 2.3 with 0.1 M NaOH, according to ISO 10271 for testing materials in dentistry. Both groups of samples, NiTi-1 and NiTi-2,were immersed in 5 mL of each test solution in a test tube clogged with a rubber cork using nylon string. The test duration was 168 h at a constant temperature of 37 ± 0.2 • C. Parallel with the samples, the pure solution "zero samples" were treated in the same way. To evaluate the migration Molecules 2022, 27, 1960 5 of 14 of ions from the samples into the solution, the chemical content of the suspensions was overseen by ICP-MS.

FIB Cross-Section of the Immersed Samples
After the immersion test was performed, the formed corrosion layers on all samples were characterized with the focus ion beam (FIB) technique. A Quanta 200 3D electron microscope with anion beam gun was used for measuring the depth of the formatted corrosion layer. A focused beam of primary gallium ions enabled the etching of the surface of the samples and the cutting of a cross-section without contamination, thus gaining a direct insight into the resulting corrosion layer. The thickness of the formatted surface oxide layer of the NiTi-1 and NiTi-2 samples was measured in 4 segments, and 5 measurement points were chosen within each of the respective segments, followed by statistical processing of the results at the end of the process.

In Vitro Determination of Biocompatibility
An MTT test was performed to evaluate the biofunctionality and biocompatibility testing of the NiTi-1 and NiTi-2 samples. Human gingival tissues were obtained with written consent from healthy donors, and the gingival tissue was minced into 1 mm 3 fragments and subjected to the outgrowth method. The minced tissue was placed in 25 cm 2 culture flasks with a growth medium (DMEM/F12 supplemented with 10% FBS and 1% ABAM, all from Gibco, Thermo Fisher, MA, USA) and incubated at 37 • C in a humidified 5% CO 2 atmosphere. The cells were passaged regularly upon reaching 80% confluence. The culture medium was changed every 2-3 days. After the third passage, HGCs were used in the study.
For the assessment of mitochondrial activity after direct exposure to the tested materials the medium was discarded and a medium containing 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyltetrazolium bromide (MTT, 0.5 mg/mL) (Sigma-Aldrich, St. Louis, MO, USA) was added to each well and incubated. After 4 h the supernatant was discarded and dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO, USA) was added to each well. The plate was placed on a shaker for 20 min at 250 rpm, in the dark, at 37 • C. The extracted colored solutions from 12-well plates were transferred into a new 96-well plate. The optical density was measured at 550 nm using a microplate reader, RT-2100c (Rayto, Shenzhen, China). As a control, cells were seeded onto sterile glass discs, identical in size and shape to the experimental discs. The percentage of mitochondrial activity was calculated as the difference to the control group.
Preparation of the samples for SEM observations: After 7 days, the cells were fixed in 2% glutaraldehyde for 2 h at 40 • C, dehydrated with increasing concentrations of ethanol (30%, 50%, 70%, 90%, 100%) with 10 min for each concentration, transferred to a critical point dryer for 30 min, and gold-coated before a scanning electron microscopic evaluation in a JEOL JSM-6610LV machine (Dearborn, MI, USA).

Statistical Analysis
The data were analyzed using IBM SPSS Statistics v22 software (SPSS Inc., Chicago, IL, USA). An independent samples t-test was used for evaluating the differences in the hardness between the NiTi-1 and NiTi-2 samples.
The results of the formatted oxide layers' thickness were shown as quantitative data and as mean ± standard deviation (SD). A quantile-quantile (Q-Q) graphical technique was used to determine if the data were normally distributed.
For the purpose of comparing the results of cell viability, an independent samples t-test was used for evaluating the differences between the NiTi-1 and NiTi-2 samples. A one-way analysis of variance (ANOVA) test was used for evaluating the differences between both the NiTi samples and the control. A p-value less than 0.05 was considered to be statistically significant.

Chemical Composition (SEM/EDX and SEM/XRF Analyses)
The results of the EDX and XRF analyses revealed that the presence of nickel was higher in the NiTi-2 samples. Contrary to that, both analyses demonstrated that the presence of titanium was higher in the NiTi-1 sample. The analyses detected iron only in the NiTi-2 sample (Table 1).

Phase Identification (EBSD Analysis)
The phase distribution in the microstructure of the classical cast and continuous cast samples was analyzed using EBSD as shown in Figures 1 and 2. The results demonstrated the presence of NiTi-cubic and Ni 3 Ti-hexagonal phases in the NiTi-1 sample. On the other hand, an additional NiTi 2 cubic phase was detected in the NiTi-2 sample.
to be statistically significant.

3.1.Chemical Composition (SEM/EDX and SEM/XRF Analyses)
The results of the EDX and XRF analyses revealed that the presence of nic higher in the NiTi-2 samples. Contrary to that, both analyses demonstrated that t ence of titanium was higher in the NiTi-1 sample. The analyses detected iron on NiTi-2 sample (Table 1). The phase distribution in the microstructure of the classical cast and continu samples was analyzed using EBSD as shown in Figures 1 and 2. The results demo the presence of NiTi-cubic and Ni3Ti-hexagonal phases in the NiTi-1 sample. On t hand, an additional NiTi2 cubic phase was detected in the NiTi-2 sample.   between both the NiTi samples and the control. A p-value less than 0.05 was considered to be statistically significant.

3.1.Chemical Composition (SEM/EDX and SEM/XRF Analyses)
The results of the EDX and XRF analyses revealed that the presence of nickel was higher in the NiTi-2 samples. Contrary to that, both analyses demonstrated that the presence of titanium was higher in the NiTi-1 sample. The analyses detected iron only in the NiTi-2 sample (Table 1). The phase distribution in the microstructure of the classical cast and continuous cast samples was analyzed using EBSD as shown in Figures 1 and 2. The results demonstrated the presence of NiTi-cubic and Ni3Ti-hexagonal phases in the NiTi-1 sample. On the other hand, an additional NiTi2 cubic phase was detected in the NiTi-2 sample.

Grain Size Measurement
The microstructure was examined with optical microscopy (Figure 3) on the 500 um scale with a magnification of 50, and the grain size was determined microscopically according to the ASTM Standard. The microstructural ASTM analysis showed that, for the NiTi-1 alloy, the grain number (G) was 5 and there were 256 grains per 1 mm 2 , which indicates a typical deformed state, while in NiTi-2 the grain number was 7 and there were 1.024 grains per 1 mm 2 . The ASTM grain number increased with the decreasing grain size ( Table 2).

Grain Size Measurement
The microstructure was examined with optical microscopy (Figure 3) on the 500 scale with a magnification of 50, and the grain size was determined microscopically cording to the ASTM Standard. The microstructural ASTM analysis showed that, fo NiTi-1 alloy, the grain number (G) was 5 and there were 256 grains per 1 mm 2 , w indicates a typical deformed state, while in NiTi-2 the grain number was 7 and there w 1.024 grains per 1 mm 2 . The ASTM grain number increased with the decreasing grain ( Table 2).

Sample Hardness
The descriptive statistics for the hardness measurements are given in Table 3. results obtained on eight measurements showed that the mean hardness value for NiTi-2 sample was two times higher (624 HV) than for NiTi-1 (317 HV). The results o hardness measurements obtained by the independent samples t-test by group (Ni NiTi-2) showed that there was a statistically significant difference in the hardness betw NiTi-1 and NiTi-2 (p ≤ 0.001).  Figure 4 presents the optical microstructure of indentation where the length o diagonal of the impression was created as a negative (copy) of the indenter. As it ca seen, the impression of NiTi-2 is smaller than in the NiTi-1sample.

Sample Hardness
The descriptive statistics for the hardness measurements are given in Table 3. The results obtained on eight measurements showed that the mean hardness value for the NiTi-2 sample was two times higher (624 HV) than for NiTi-1 (317 HV). The results of the hardness measurements obtained by the independent samples t-test by group (NiTi-1, NiTi-2) showed that there was a statistically significant difference in the hardness between NiTi-1 and NiTi-2 (p ≤ 0.001).  Figure 4 presents the optical microstructure of indentation where the length of the diagonal of the impression was created as a negative (copy) of the indenter. As it can be seen, the impression of NiTi-2 is smaller than in the NiTi-1sample.   Table 4 shows the presence of released ions, measured by the ICP-MS analysis, for both NiTi-1 and NiTi-2 samples in the solutions of artificial saliva and lactic acid. Immersion in a solution of artificial saliva with a pH 6.5 after 7 days did not induce significant ion release in neither of samples (NiTi-1 and NiTi-2). Al and Cu were not detected. Moreover, it should also be noted that Fe, although present in NiTi-2, was not detected in the solutions, i.e., there was no release from the sample surface. The concentration of Ni ions was slightly higher in the solution with the NiTi-1 samples (0.05 μg/cm 2 ) in comparison to the solution with the NiTi-2 samples (0.04 μg/cm 2 ). The concentration of Ti ions was below the detection limit in both groups. In the acidic environment, the suppression of Al ions was lower in the NiTi-1 alloy (0.01 μg/cm 2 ) than in the NiTi-2 alloy (0.06 μg/cm 2 ). Considering the migration of the Cu ions from the samples, the NiTi-1 samples released lower values (0.37 μg/cm 2 ) as opposed to NiTi-2 (0.045 μg/cm 2 ). The concentration of Ni ions was two times higher in the solution with the classical cast alloy NiTi-1 (2.33μg/cm 2 ) than the corresponding values in the continuous cast alloy NiTi-2 (1.2 μg/cm 2 ), despite the fact that the content of nickel in the NiTi-2 alloy was higher. As far as Ti ions are concerned, the suppression was highly reduced in the NiTi-2 samples (0.63 μg/cm 2 ) in comparison to NiTi-1 (2.17 μg/cm 2 ).

FIB Cross-Section Analysis
The descriptive statistics of the thickness of the oxide layer that formed on the samples' surfaces after the immersion testing are presented in Table 5. The results obtained  Table 4 shows the presence of released ions, measured by the ICP-MS analysis, for both NiTi-1 and NiTi-2 samples in the solutions of artificial saliva and lactic acid. Immersion in a solution of artificial saliva with a pH 6.5 after 7 days did not induce significant ion release in neither of samples (NiTi-1 and NiTi-2). Al and Cu were not detected. Moreover, it should also be noted that Fe, although present in NiTi-2, was not detected in the solutions, i.e., there was no release from the sample surface. The concentration of Ni ions was slightly higher in the solution with the NiTi-1 samples (0.05 µg/cm 2 ) in comparison to the solution with the NiTi-2 samples (0.04 µg/cm 2 ). The concentration of Ti ions was below the detection limit in both groups. In the acidic environment, the suppression of Al ions was lower in the NiTi-1 alloy (0.01 µg/cm 2 ) than in the NiTi-2 alloy (0.06 µg/cm 2 ). Considering the migration of the Cu ions from the samples, the NiTi-1 samples released lower values (0.37 µg/cm 2 ) as opposed to NiTi-2 (0.045 µg/cm 2 ). The concentration of Ni ions was two times higher in the solution with the classical cast alloy NiTi-1 (2.33 µg/cm 2 ) than the corresponding values in the continuous cast alloy NiTi-2 (1.2 µg/cm 2 ), despite the fact that the content of nickel in the NiTi-2 alloy was higher. As far as Ti ions are concerned, the suppression was highly reduced in the NiTi-2 samples (0.63 µg/cm 2 ) in comparison to NiTi-1 (2.17 µg/cm 2 ).

FIB Cross-Section Analysis
The descriptive statistics of the thickness of the oxide layer that formed on the samples' surfaces after the immersion testing are presented in Table 5. The results obtained on twenty measurements showed almost the same mean values (34 nm for NiTi-1 and 35 nm for NiTi-2), while the oxide layer thickness ranged between 27-55 nm for NiTi-1 and between 25-50 nm for NiTi-2. The NiTi-2 samples showed more homogeneous data than NiTi-1 regarding the depth of the oxide layer ( Figure 5).   20 20 Note: N = number of measurements.
The NiTi-2 samples showed more homogeneous data than NiTi-1 regarding th depth of the oxide layer ( Figure 5).

Biocompatibility Results
After 24h of direct exposure of human gingival cells (HGCs) to the tested material increased mitochondrial activity was observed in both tested groups (NiTi-1 and NiTiin comparison to the untreated cells (the control group) (Figure 6). Higher activity wa observed in cells seeded directly on the NiTi-1 samples but without statistically significan differences between NiTi-1 and NiTi-2 (p = 0.069). Comparing the cells' mitochondrial a tivity between groups (the control group, the cells seeded on NiTi-1 and the cells seede on NiTi-2 samples), the ANOVA test showed that there was no statistically significan increase of mitochondrial activity (p = 0.64).
After seven days of direct exposure, mitochondrial activity was still increased in th NiTi-1 and NiTi-2 samples in comparison to the untreated cells ( Figure 6), with high proliferation on NiTi-2, but, again, without a statistically significant difference betwee the groups NiTi-1 and NiTi-2 (p = 0.168). The ANOVA results and post hoc compariso tests showed that there was a statistically significant increase in mitochondrial activi

Biocompatibility Results
After 24 h of direct exposure of human gingival cells (HGCs) to the tested materials, increased mitochondrial activity was observed in both tested groups (NiTi-1 and NiTi-2) in comparison to the untreated cells (the control group) (Figure 6). Higher activity was observed in cells seeded directly on the NiTi-1 samples but without statistically significant differences between NiTi-1 and NiTi-2 (p = 0.069). Comparing the cells' mitochondrial activity between groups (the control group, the cells seeded on NiTi-1 and the cells seeded on NiTi-2 samples), the ANOVA test showed that there was no statistically significant increase of mitochondrial activity (p = 0.64).

Discussion
In the present study we compared the biofunctional properties of nickel-titanium alloys obtained by two different production processes. Contrary to conventional casting, continuous casting was previously reported to be a production process that was able to produce stands of relatively small diameter with an acceptable surface quality [17,25,26]. Even though the correlation between production processes and material properties has After seven days of direct exposure, mitochondrial activity was still increased in the NiTi-1 and NiTi-2 samples in comparison to the untreated cells ( Figure 6), with higher proliferation on NiTi-2, but, again, without a statistically significant difference between the groups NiTi-1 and NiTi-2 (p = 0.168). The ANOVA results and post hoc comparison tests showed that there was a statistically significant increase in mitochondrial activity between the NiTi-1 samples (mean = 114.44) and the control (mean = 100) (p < 0.01) and between the NiTi-2 samples (mean = 114.42) and the control (mean = 100) (p < 0.01).
SEM images of fibroblasts grown on the NiTi-1, NiTi-2, and control samples are given in Figure 7. Fibroblasts were created a thick cellular layer over the whole surface of the alloys. SEM images of fibroblasts grown on the NiTi-1, NiTi-2, and control samples are given in Figure 7. Fibroblasts were created a thick cellular layer over the whole surface of the alloys.

Discussion
In the present study we compared the biofunctional properties of nickel-titanium alloys obtained by two different production processes. Contrary to conventional casting, continuous casting was previously reported to be a production process that was able to produce stands of relatively small diameter with an acceptable surface quality [17,25,26]. Even though the correlation between production processes and material properties has

Discussion
In the present study we compared the biofunctional properties of nickel-titanium alloys obtained by two different production processes. Contrary to conventional casting, continuous casting was previously reported to be a production process that was able to produce stands of relatively small diameter with an acceptable surface quality [17,25,26]. Even though the correlation between production processes and material properties has been proven in numerous studies [27,28], there is still no clearly defined consensus on the accepted model for the production of these alloys for application in dentistry.
From the perspective of the specimens' chemical composition, it is evident that there is a difference between NiTi-1 and NiTi-2 in nickel and titanium content, which is the consequence of different production processes. Classical casting into ingots requires remelting in a vacuum furnace, due to Ti ions' tendency to oxidize. Additionally, in this procedure, it is necessary to ensure appropriate mixing conditions for both components. High equilibrium conditions are usually achieved during remelting, with the formation of a practically homogeneous alloy, according to the NiTi phase diagram. In order to achieve the required dimensions for the usage in dentistry, nitinol products obtained by classical casting into ingots, have to undergo secondary fabrication methods (thermo-mechanical treatments). During rolling or drawing of the material, large deformations occur, as confirmed by microstructural analysis. On the other hand, a significantly lower mass (20 kg) of the nitinol alloy is used within the single procedure of continuous casting. The higher nickel and lower titanium content in NiTi-2 alloys is the consequence of mixing conditions of the melt. Namely, vacuum induction melting in the crucible is performed at a medium frequency (f = 8000 Hz) and additional mixing of Ni and Ti components would probably be necessary with use of lower frequencies (f = 2000 Hz). Even though continuous casting technology results in the different chemical contents of the Ni and Ti components, this fact does not negatively affect the biological characteristics of the obtained alloy. Additionally, since there is no need for secondary production processes, the microstructure of the continuous cast specimens is appropriate and better in comparison to the classical cast alloy.
Regarding the mechanical properties of nickel-titanium alloy, the hardness should be adapted to their clinical purpose. Given that nitinol is used predominantly in orthodontics for archwires, high hardness values are desirable. The problem with components of orthodontic arches with low hardness values is that they might compromise the transfer of torque force from an activated archwire to the bracket, which can cause possible plastic deformation of the wings [29]. Our study demonstrated a higher hardness of NiTi-2 in comparison to the NiTi-1 alloy. This can be explained by a higher nickel content, and/or iron content, as shown by EDX/XRF analysis, which is in agreement with previous reports [25,26]. An additional EBSD analysis revealed that there were two phases in the microstructure of NiTi-1 samples (NiTi-cubic and Ni 3 Ti hexagonal), while an additional NiTi 2 -cubic phase was detected in the NiTi-2 microstructure. It can be concluded that this phase was able to increase the hardness. It could also be concluded that the finer the microstructure, the higher the hardness.
Immersion tests performed for seven days estimated the quantity of Ni and Ti ions released in pH neutral and acidic environments. Different concentrations of released ions were only detected in the acidic environment. NiTi-1 and NiTi-2 alloys in an artificial saliva solution, which has a neutral pH, showed no significant difference in the concentration of released ions, nor were these amounts of clinical significance. These results were expected, considering that the good corrosion resistance in pH neutral solutions of nickel-titanium alloys is well documented [9,30,31]. Maximum allowable doses of nickel ions for humans are 0.5 µg/cm 2 /week [32]. This is 10 times higher than the results that we obtained in the artificial saliva. Our results indicate that the nickel ion release in artificial saliva was 0.04 µg/cm 2 /week from NiTi-2 samples and 0.05 µg/cm 2 /week from NiTi-1.
However, various studies have demonstrated lower corrosion resistance of these alloys in acidic solutions and chloride-containing environments [31,33,34]. In general, it is well documented that the corrosion potential increases as the pH value decreases [33][34][35].
The acidic environment had a significant effect on increasing the release of toxic ions from both groups of samples. However, the Ni ion release was two times lower, and the Ti ion release was three times lower from the NiTi-2 than from the NiTi-1 samples. Even though the oral cavity pH is predominantly neutral, changes towards the acidic environment may happen, but are short in duration, depending mainly on the consumption of specific foods. However, the duration of the immersion tests both in neutral and acidic environments should be extended in future experiments.
After the immersion test was performed in lactic acid, a surface oxide layer was formed in both groups of samples, NiTi-1 and NiTi-2. The results obtained with the FIB cross-section showed similar thicknesses of the oxide layer formed on samples' surfaces, even though ICP analysis showed different concentrations of metal ions released from the samples. The lower values of the standard deviations for the depth of the oxide layer in the NiTi-2 samples also indicated a greater data consistency in continuous casting, even though the thickness of the corrosive layer was almost the same in both samples.
In terms of Ni and Ti ion release, the results from both groups of samples showed that the release of nickel ions was higher than of titanium ions, and this supports the previous studies [36,37]; it can be explained by the fact that, after the particular dissolution of the surface TiO 2 layer, the inner layer, which is an Ni-rich layer, is exposed [37,38]. According to the literature data, the surface characteristics of materials act like ion diffusion barriers [27,28,38]. It is hard to obtain such a surface integrity without secondary surface modification processes and, to our knowledge, this is one of the first studies that compares classical and continuous casting as the primary manufacturing process. The present research established that the NiTi-2 alloy, which has a more stable microstructure with smaller grains and a higher grain number, is more biologically resistant. It also showed that Ni ion release was not proportional to the content of nickel in the alloy samples, a finding that is in accordance with previous investigations [34,39].
The present study clearly demonstrated the biocompatibility of nickel-titanium alloy manufactured by continuous casting. To be able to perform biocompatibility testing on appropriate cells, it is necessary to know the different cellular behaviors for the different material surface properties. Fibroblasts are known to be present in almost all types of tissues and to play an important role in tissue inflammatory reactions. Moreover, these cells prefer smoother surfaces, showing better adhesion to them [27]. Accordingly, we prepared samples by the machine polishing of their surfaces. Generally, the mitochondrial activity of cells grown directly on the tested samples was greater at both points in time compared to the control cells with a proliferation tendency during the observation time. However, further investigations are needed for the evaluation of the potential genotoxic effects of these alloys.

Conclusions
On the basis of the present study, it can be concluded that there is a significant difference in the biofunctional properties between nickel-titanium alloys obtained by classical and by continuous casting. Nickel-titanium alloys obtained by continuous casting have a more stable microstructure, higher hardness, and better resistance. They also have an additional NiTi 2 -cubic phase in their structures that contributes to greater stability and biocompatibility.

Institutional Review Board Statement:
The study, conducted in accordance with the Declaration of Helsinki, was reviewed and approved by the institutional Ethical Committee (approval number 36/7). Informed Consent Statement: All patients signed a written informed consent.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available as they are part of a yet undefended PhD thesis.

Acknowledgments:
The authors gratefully acknowledge Rebeka Rudolf, University of Maribor, and Aleš Stambolić, ITM Ljubljana, for the EBSD analysis.