The Electrochemical Behavior of the NiTi Alloy in Different Simulated Body Fluids

The NiTi alloy is extensively used in the biomedical field due to its notable mechanical properties of superelasticity and shape memory1. Comparative studies of biocompatibility between alloys2 show that the solubilized nickel may present several adverse reactions. Despite the reports of some authors3,4 of the lack of cytotoxicity of alloys, a study conducted by Fili5 found that Ni-rich precipitates in NiTi osseous implants chemically passivized with HF, HNO3 and H2O 300 days after implantation. A number of methodologies were proposed with the objective of creating a barrier that would block the release of nickel, resulting in an increased biocompatibility of the alloy and corrosion resistance: anodization6,7, heat treatments for superficial oxidation8,9, sputtering10,11, plasma electrolytic oxidation12, plasma immersion ion implantation13, hydroxiapatite/zirconia composite coating14, TiN coating15, sol-gel technology16 metal vapor vaccum arc plasma source17. The correlation of the results from these studies in regards to the corrosion resistance is complex due to the different simulated body fluids that were used as electrolytes. The American Society for Testing and Materials (ASTM) standard to determine the corrosion susceptibility of small implant devices does not specify the electrolyte formulation – only an ion concentration range is mentioned18. However, the different electrolyte compositions may result in distinct aggressiveness and corrosion mechanisms. Some of the most used solutions to simulate body fluids include: Ringer’s solution19, Hanks solution20,21, Hanks balanced salt solution (HBSS)22, saline body fluid solution (SBF)23. A study of Qiu et al.14 used mechanically polished and hydroxiapatite/zirconia (HAP/ZrO2) electrodeposited NiTi. The electrochemical corrosion test was performed in an SBF solution. These authors observed a corrosion current density (icorr) of 3.98 × 10 –7 A/cm2 for the NiTi uncoated sample, and of 7.00 × 10–9 A/cm2 for the HAP/ZrO2 coated sample, which is almost 60 times lower14. They also found that the breakdown potential (Ebr) of the bare NiTi sample appeared at nearly 600 mV, while coated samples did not break at studied potentials14. Zhao et al.17, on the other hand, used a commercial NiTi polished alloy samples, as well as samples polished by metal vapor vaccum arc plasma (MEVVA 100) with different parameters. The electrochemical tests were conducted in Hank’s solution. The Electrochemical Behavior of the NiTi Alloy in Different Simulated Body Fluids


Introduction
The NiTi alloy is extensively used in the biomedical field due to its notable mechanical properties of superelasticity and shape memory 1 .Comparative studies of biocompatibility between alloys 2 show that the solubilized nickel may present several adverse reactions.Despite the reports of some authors 3,4 of the lack of cytotoxicity of alloys, a study conducted by Fili 5 found that Ni-rich precipitates in NiTi osseous implants chemically passivized with HF, HNO 3 and H 2 O 300 days after implantation.A number of methodologies were proposed with the objective of creating a barrier that would block the release of nickel, resulting in an increased biocompatibility of the alloy and corrosion resistance: anodization 6,7 , heat treatments for superficial oxidation 8,9 , sputtering 10,11 , plasma electrolytic oxidation 12 , plasma immersion ion implantation 13 , hydroxiapatite/zirconia composite coating 14 , TiN coating 15 , sol-gel technology 16 metal vapor vaccum arc plasma source 17 .The correlation of the results from these studies in regards to the corrosion resistance is complex due to the different simulated body fluids that were used as electrolytes.The American Society for Testing and Materials (ASTM) standard to determine the corrosion susceptibility of small implant devices does not specify the electrolyte formulation -only an ion concentration range is mentioned 18 .However, the different electrolyte compositions may result in distinct aggressiveness and corrosion mechanisms.Some of the most used solutions to simulate body fluids include: Ringer's solution 19 , Hanks solution 20,21 , Hanks balanced salt solution (HBSS) 22 , saline body fluid solution (SBF) 23 .
A study of Qiu et al. 14 used mechanically polished and hydroxiapatite/zirconia (HAP/ZrO 2 ) electrodeposited NiTi.The electrochemical corrosion test was performed in an SBF solution.These authors observed a corrosion current density (i corr ) of 3.98 × 10 -7 A/cm 2 for the NiTi uncoated sample, and of 7.00 × 10 -9 A/cm 2 for the HAP/ZrO 2 coated sample, which is almost 60 times lower 14 .They also found that the breakdown potential (E br ) of the bare NiTi sample appeared at nearly 600 mV, while coated samples did not break at studied potentials 14 .Zhao et al. 17 , on the other hand, used a commercial NiTi polished alloy samples, as well as samples polished by metal vapor vaccum arc plasma (MEVVA 100) with different parameters.The electrochemical tests were conducted in Hank's solution.
Nb-NiTi samples exhibit much higher E corr (-396 mV) and E br (1094 mV) than the untreated NiTi samples (E corr = -478 mV, E br = 420 mV).The authors state that E br is an indication to evaluate the susceptibility to pitting corrosion, thus, higher E br values means higher resistance to pitting corrosion 17 .They also found that the i corr values of the Nb-NiTi samples (5.89 × 10 -10 A/cm 2 ) were much lower than that of the untreated NiTi (1.07 × 10 -8 A/cm 2 ), suggesting a much slower eroding rate.In both studies, the developed surface treatment improved the corrosion resistance of the NiTi.Even though the Nb alloy treatments developed lower current densities, it is important to notice that the corrosion densities found by Zhao et al. 17 were also lower than the values obtained by Qiu et al. 14 for the NiTi without surface treatment.Therefore, when the same substrate (NiTi) is analyzed in different electrolytes that simulate body fluids, there may be problems in interpreting the results obtained by different authors.
Shahrabi et al. 11 studied bulk NiTi, polished to 1500 grit, in both Ringer's and Hank's solutions.The authors observed a higher E br in Ringer's solution (747 mV).Furthermore, the passive region of the bulk in Ringer's solution has shifted to the right (higher current density) in comparison with Hank's solution 11 .Another work, by Liang and Mou 24 , studied the effects of different electrolytes (Ringer's, PBS and Hank's solution) on anticorrosion biometallic materials (SUS316L stainless steel, Co-Cr alloy and Ti-6Al-4V).The results indicate that the corrosion caused by the Ringer's solution is the strongest, followed by the PBS and Hank's solutions.In addition, the decrease of the pH of the solution significantly increased the corrosion rate and the susceptibility to localizated corrosion of SUS316L and Co-Cr.On the other hand, Ti-6Al-4V alloys exhibited stability and only a slight increase of corrosion rate with decreasing pH.
I n g e n e r a l , a s m e n t i o n e d b e f o r e , s t u d i e s envolving biomedical alloys already showed different electrochemical results when distincts electrolytes were used to simulated the body fluids, which indicates the important influence of the electrolyte composition.Besides, different alloys can present distint behavior in the same electrolyte.In this context, the aim of this work is to evaluate the influence of the different electrolytes used to simulate the body fluids on assessing the NiTi corrosion resistance.For this purpose, the Ringer, Hanks, HBSS and SBF solutions were chosen.Additionally a 0.9% NaCl solution was used following some authors suggestions that its corrosion potential is similar to the body fluid due to its high chloride concentration 1,25,26 .

Sample preparation
The near-equiatomic superelastic NiTi (Ni 55.8 wt%) alloy was used for this study.NiTi samples were cut using the Electric Discharge Machining (EDM).The test specimen dimensions were 15 mm x 15 mm x 1 mm.All samples were polished with silicon carbide sandpaper up to the 600 grade and subsequently isolated with Araldite  , defining an area of 0.91 cm² for electrochemical tests.

Preparation of electrolytes
The following electrolytes were prepared according to published articles: Ringer 27 , Hanks 28 , HBSS 22 and SBF 29 .A 9 g.L -1 (153.mmol/L) NaCl (0.9% NaCl) distilled water solution was prepared.The ion concentration of each prepared solution is described on Table 1.For comparison purposes, the human blood plasma ion concentrations is also demonstrated on Table 1 28 .

Electrochemical characterization
Open circuit potential (OCP) measurements and potentiodynamic polarization curves were performed threefold, under an inert atmosphere, in an Autolab PGSTAT302N potenciostat/galvanostat, according to the ASTM F2129-08 standard 18 .The test temperature was of 25°C (0.9% NaCl), with a maximum variation of 1°C, and of 37 °C -for the other solutions, with the same variation.The 0.9% NaCl solution was prepared at room temperature to reproduce the methodology described by several authors 1,25,26 .The OCP and potentiodynamic polarization were performed in a conventional three-electrode electrochemical cell.The working electrode was the NiTi samples, the counter electrode was a platinum wire, and the reference electrode was a satured calomel electrode (SCE).The OCP was monitored for 1 hour, the period that was necessary to stabilize the electrochemical system.The cyclic voltammetry was performed from the OCP potentials until 2 V in reference to the SCE with a scan rate of 1 mV.s -1 .

Superficial characterization
The superficial characterization was performed in polished NiTi after potenciodynamic polarization in 0.9% NaCl, Hanks, HBSS, Ringer and SBF solution using an scanning electron microscopy (SEM), JEOL 6060.Characterization was also made in polished NiTi without the electrochemical test for comparison purposes.

Electrochemical characterization
The OCP values that were obtained are shown in Figure 1.The lowest OCP value, measured at 3600 seconds (-291 mV), which is the most active potential, was developed in HBSS solution.Considering the electrolytes studied at 37 °C, HBSS, SBF and Ringer demonstrated the highest corrosion potential, which may be associated to a higher Cl -concentration in these electrolytes (Table 1).NiTi alloy is susceptible to pitting corrosion in chloride containing solutions 30 .Attack by Cl -in NiTi result in Ni being released into the solution and decrease in the local Ni concentration at the pitting sites 31 .The remaining Ti reacts with dissolved oxygen from the solution to form titanium oxides in the corroded area 31 .The corrosion product layer expands over the entire surface and is composed of TiO 2 , Ti 2 O 3 , and TiO with depleted Ni 31 .The change in the Ni release rate is related to the corrosion defects such as pitting pores on the NiTi specimens 31 .In the early immersion period, nickel ions area released gradually into the surrounding solution 31 .As the immersion continues, corrosion results in the formation of pitting pores that promote nickel release, leading to a high Ni release rate 31 .However, if the immersion time is long enough, the corrosion-induced titanium oxide layer seals the pitting pores on the NiTi surface, thereby reducing the Ni release rate 31 .Early pitting corrosion takes place on many sites on the surface.Early corrosion process take place on surface defects initially and involves the breakdown of original thin oxide layer and Ni-Ti bond after attack by Cl -, and this process releases Ni ions 31 .As time elapses, the pitting sites propagate and noticeable pores form on the surface 31 .Ni released from the bulk materials into the solution reduces the local concentration of Ni in the materials 31 .Meanwhile, the remaining Ti reacts with dissolved oxygen in the solution to form titanium oxides around the pitting sites and pores 31 .As the corrosion process proceeds, the titanium oxide layer grows and propagates over the entire surface of the NiTi specimen 31 .The pores formed during corrosion are possibly blocked by the titanium oxides formed inside the pores.As a result, a uniform and dense oxide layer is formed 31 .After the titanium oxide layer has formed on the entire surface of the NiTi specimen, it will serve as a passivation film to retard the corrosion process 31 .This susceptibility increases when the chloride concentration is higher 30 .In the Hanks solution, the NiTi alloy developed a less active potential (-214 mV). Figure 2 presents the voltammograms of NiTi in different studied solutions.A slight increase in the current density occurred in the anodic potential scan, indicating a superficial oxide disruption allowing the current flow.This phenomenon of anodic dissolution in the surface indicates the beginning of the oxide layer breakdown.When the applied potential reaches 2 V, it is reversed, and for all systems the current density of the inverse scan is higher (Figure 2).This behavior is associated to the pitting corrosion of the NiTi alloy in the studied solutions 27 .
In the 0.9% NaCl solution, the NiTi samples developed current densities that were one order of magnitude lower than the others studied solutions (Figure 2) indicating that a lower temperature (25 °C) and the absence of other ions, such as HCO 3 -1 , can make the medium less aggressive to this alloy.However the difference in temperature, 25 °C for the test in 0.9% NaCl solution compared to 37 ºC for the other electrolytes can contribute for the differences observed between NaCl solution and the other solutions.Few studies have investigated the effects of temperature on the corrosion resistance of nitinol, but a study in Hank's physiological solution at temperatures ranging from 10 to 80 °C show that the ability of NiTi to repassivate is significantly reduced by an increase in temperature 32 .Studies using other alloys proved that a bicarbonate ion can induce pitting corrosion 33 .Furthermore, in a 0.9% NaCl solution it is possible to observe the disruption of the passive layer at 757 mV (Figure 2).Considering the solutions that were analyzed at 37 °C, the Hanks solution demonstrated the lowest current densities in the passive zone of the polarization curve (Figure 2).This is probably associated to HCO 3 -ion concentration in the Hanks solution (Table 1).The oxide layer was firstly ruptured, in Hanks solution, when the applied potential reached 99 mV (current density increased from 1.55 × 10 -7 A/cm 2 to 8.72 × 10 -7 A/cm²) while this value was -78 mV (current density increased  from 3.20 × 10 -6 A/cm² to 1.17 × 10 -5 A/cm 2 ) for Ringer, -119 mV (current density increased from 2.80 × 10 -6 A/cm² to 1.17 × 10 -5 A/cm 2 ) for HBSS and -148 mV (current density increased from 2.80 × 10 -6 A/cm 2 to 1.39 × 10 -5 A/cm²) for SBF.In the Hanks solution the NiTi shows a lower tendency to corrosion due to a minor passive current and a higher pitting potential, when compared to other solutions studied at 37°C.This results demonstrated a lower corrosive tendency of NiTi in Hanks solution in comparison with other studied solutions.The variation in the passive zone of the polarization curve of different alloys due to the concentration of bicarbonate in the electrolyte was described by other authors 34 .Ions bicarbonate accelerated the cathodic reactions, promoting an oxide growth on the surface, and resulting in lower anodic current densities.Despite the curves in the other solutions having similar behavior (Figure 2), the potentials in which sudden increases of current density occurred varied.This fact suggests that it is not possible to compare the results of studies performed in different electrolytes, even if they have all been performed in simulated body fluids.A standardization of all of the parameters for corrosion tests is important to compare the different studies about the protection of NiTi surfaces.It should be mentioned that, according to some authors, it is difficult to choose an ideal solution to simulate the body fluids due to the complexity of the human physiological enviroment 24 .Other authors suggest that organic acids found in the blood should also be considered in the preparation of simulated body fluids 27 .
In the present paper, the Hanks and the HBSS solutions demonstrated a good reproducibility of the results.
Considering that all samples were made of the same alloy and were submitted to the same superficial treatment, and the electrochemical tests were performed under the same parameters (temperature, sealing and scan rate), the corrosion potential variance is probably due to the different aggressiveness levels of the studied electrolytes, despite the slight difference in the pH among the solutions, which can influence the cathodic curves, thus affecting the corrosion potential.Nevertheless, this influence is probably not relevant, since all of the solutions are in neutral pH range.A study developed in Hank's physiological solution with pH varying from 1 to 9 showed that corrosion resistance of NiTi is not significantly affected by pH values 32 .The Hanks solution showed the lowest aggressiveness in comparison to the other simulated body fluid solutions, which is evidenced by the less active corrosion potential and by the evolution of minor current densities.Other authors also demonstrated that the Ringer solution is stronger than the Hanks.This variation was associated to the presence of Na 2 PO 4 and KH 2 PO 4 , which benefits the biomaterials against corrosion, in Hanks solution 11,24 .It is important to notice that the Cl -ions concentration is lower in the Hanks solution, in comparison to the others studied solutions.Studies indicate that the NiTi alloy is suscetible to pitting corrosion in chloride solutions, and in increasing chloride concentration, the pitting potential decreases 30 .In the HBSS solution the NiTi alloy developed a higher current density and more active potentials (Figure 2), which was expected due to the results obtained by the OCP (Figure 1).

Superficial characterization
After the electrochemical tests, the NiTi samples were analyzed by scanning electron microscopy in order to determine the relation to the corrosion mechanism found in the electrochemical tests (Figures 3-8).Although the electrochemical results demonstrated that the NiTi alloy shows the same corrosion mechanism (pitting) in all studied simulated body fluids, only with the HBSS solution it was possible to notice the presence of oxidation products in the surface (Figure 6), resulting from more active potentials (Figure 2) and a higher current density; such products have been described by other authors 35,36 as titanium hydroxides, which are more soluble, less protective than the oxides, and are known by an increase in the current between ~1,0 and      1,5 V. Other studies 37 have showing pitting corrosion on NiTi, in physiological solution, at potentials between +190 and +280 mV versus SCE.Oshida et al. 38 , evidenced NiTi low pitting potential values, and presence of numerous oscillations of anodic passive current for potentials lower than 500 mV.

Conclusions
Evaluating coatings through an appropriated method is as important as developing surface treatments for NiTi alloys.Considering that the corrosion resistance is a major issue in these alloys, it is relevant to standardize the electrochemical procedure parameters, as the electrolytes, allowing a comparison between the various coatings developed by several researchers.This study compared the most commonly used electrolytes to simulated body fluids: Hanks, HBSS, SBF, Ringer and 0.9% NaCl.Electrochemical and morphological results, show that the NiTi presents the same corrosion mechanism (pitting induced by Cl -and HCO 3 -ions) in all studied simulated body fluids.However, the Hanks solution demonstrated less aggressiveness in comparison to other simulated body fluids, which is evidenced by its less active corrosion potential and by the development of lowest current densitie values, probably due to a lower chloride concentration.In the HBSS, SBF and Ringer solutions, the NiTi showed higher current densities and a more active potential.The Hanks and HBSS solutions presented the best reproducibility of results.Considering that the HBSS represents an extreme environment, this solution seems to be the most indicated to study the corrosion behavior of NiTi treated surfaces.

Figure 1 .
Figure 1.Open circuit potentials in different simulated body fluids at 37 °C and in 0.9% NaCl at 25 °C.

Figure 2 .
Figure 2. Potentiodynamic polarization in different simulated body fluids at 37 °C and in 0.9% NaCl at 25 °C.Voltammogram detail showing the passive region of NiTi in (a) HBSS solution, (b) SBF solution and (c) Ringer solution.

Figure 3 .
Figure 3. Mechanically polished NiTi images obtained by scanning electron microscope.

Figure 4 .
Figure 4. NiTi samples after electrochemical tests in Hanks solution, by scanning electron microscope.

Figure 5 .
Figure 5. NiTi samples after electrochemical tests in SBF solution, by scanning electron microscope.

Figure 6 .
Figure 6.NiTi samples after electrochemical tests in HBSS solution, by scanning electron microscop.

Figure 7 .
Figure 7. NiTi samples after electrochemical tests in Ringer solution, by scanning electron microscope.

Table 1 .
Simulated body fluids and human blood plasma ion concentrations.