The Estimation of Corrosion Behavior of NiTi and NiTiNb Alloys Using Dynamic Electrochemical Impedance Spectroscopy

1 Faculty of Chemical Engineering and Environmental Protection, “Gheorghe Asachi” Technical University of Iasi, 73 Professor Dr. Doc. D. Mangeron Boulevard, 700050 Iasi, Romania 2 Faculty of Mechanical Engineering, “Gheorghe Asachi” Technical University of Iasi, 61-63 Professor Dr. Doc. D. Mangeron Boulevard, 700050 Iasi, Romania 3 Faculty of Materials Science and Engineering, “Gheorghe Asachi” Technical University of Iasi, 61A Professor Dr. Doc. D. Mangeron Boulevard, 700050 Iasi, Romania


Introduction
Ti and its alloys have become one of the most attractive biomaterials due to their better corrosion resistance, biocompatibility, good fracture toughness, and relatively low modulus [1][2][3][4].
Nevertheless, a major concern on the dissolution properties of NiTi alloy still remains owing to the high nickel content; nickel science has been shown to determine several adverse biological effects [13,14].
�sually, NiTi alloys present a super�cial thin layer with TiO 2 as the main component [15] protecting them from dissolution, as for commercial pure titanium [16].e potential of this passive �lm breakdown is, sometimes very low for NiTi alloys, leading to active dissolution processes.ird alloying element was added to NiTi alloy to enhance its mechanical or corrosion properties.NiTiNb alloys have been developed recently as potential orthodontic arch wire materials; unfortunately little published works can be found so far as concerning the corrosion resistance of these alloys in physiological solution [17].
e electrochemical impedance spectroscopy (EIS) is an important technique for the examinations of passive layer.In the classical EIS technique the system can be considered as a stationary one, in which case the investigated process should remain constant during the time of full impedance experiment.Dynamic electrochemical impedance spectroscopy (DEIS) is a method of impedance measurements and creates a new possibility of corrosion investigation.In this method, impedance spectra are determined under potentiodynamic conditions (nonstationary condition).
In the present work, the dynamic electrochemical impedance spectroscopy (DEIS) measurements were carried out to investigate the passive �lm behavior under linear potential change for NiTi and NiTiNb alloys.

Materials
A NiTiNb alloy produced by National Institute of Research and Development for Non-ferrous and Rare Metals, Bucharest, Romania, was used for the study; it had the following chemical composition (% by weight): Ti = 37.8, Nb = 14.5, and Ni = balance.An equiatomic NiTi alloy produced by the same supplier was used in some tests as a reference material.Both alloys were synthesized by electron beam melting technique.
e samples were cut into 0.21 cm 2 sizes, and brass nut was attached to the sample using conductive paint to ensure electrical conductivity.e assembly was then embedded into an epoxy resin disk [18].
en the samples were grounded with SiC abrasive paper up to 2000 grit; �nal polishing was done with 1 m alumina suspension.e samples were degreased with ethyl alcohol followed by ultrasonic cleaning with deionised water and dried under an air stream.
Electrochemical corrosion tests were performed for both samples in aerated physiological 0.9 wt% NaCl solution (B.Braun Melsungen AG, Germany).e pH of this solution was 6.9.

Electrochemical Measurements.
Electrochemical measurements were carried out in 0.9 wt% NaCl solution at 37 ∘ C using a Princeton Applied Research potentiostat (Model 263 A) connected with a Princeton Applied Research 5210 lock-in ampli�er controlled by a personal computer and a speci�c soware package called Electrochemistry Power Suite (Princeton Applied Research).A glass corrosion cell kit with a platinum counter-electrode and a saturated calomel reference electrode (SCE) was used to perform the electrochemical measurements.All potentials referred to in this paper are with respect to SCE.
Aer the open circuit potential (OCP) was held for 1 hour, the linear anodic polarization test was performed in 0.9 wt% NaCl solution at a sweep rate of 1 mV/s, from OCP to 500 mV.
Dynamic electrochemical impedance spectra (DEIS) were acquired from the open circuit potential to the dissolution region with a step potential of 20 mV in the frequency range of 50 kHz-0.01Hz with a 10 mV amplitude sine wave.
In order to supply quantitative support for discussioning the experimental impedance spectra results, an appropriate model (ZSimpWin-PAR, USA) for equivalent circuit (EC) quanti�cation has also been used.e usual guidelines for the selection of the best-�t EC were followed: a minimum number of circuit elements are employed, the  2 error was suitably low ( 2 < 10 −4 ), and the error associated with each element was up to 5%.Instead of pure capacitors, constant phase elements (CPEs) were introduced in the �tting procedure to obtain good agreement between the simulated and experimental data.

Scanning Electron Microscopy of Corroded Surfaces.
To observe the surface morphology of the alloys aer linear anodic potentiodynamic polarization test, the specimens were ultrasonically cleaned in deionized water, dried under an air stream, and then examined by a scanning electron microscopy (SEM).To perform this Vega Tescan, scanning electron microscope (model VEGA II LMH) was used.

Results and Discussion
Plots in semilogarithmic scale of current densities corresponding to NiTi samples in aerated physiological 0.9 wt% NaCl solution at 37 ∘ C traced between open circuit potential to 500 mV with 1 mV/s potential sweep rate are displayed in Figure 1.Prior to the beginning of the linear anodic polarization procedures, the samples were kept in the solution for 1 hour.
e nature of the linear anodic potentiodynamic polarization curves indicated that both samples have been passivated immediately aer the immersion in physiological 0.9 wt% NaCl solution.ese samples translated directly from the "Tafel region" into a stable passive state, without exhibiting an active-passive transition.e addition of Nb on NiTi alloy shied to the positive (noble) direction the anodic polarization curve.Figure 1 clearly proves that NiTiNb alloy had a better corrosion resistance than that of NiTi alloy, and this results in smaller values for the passive current densities ( pass ).Passive current density was determined from the linear anodic potentiodynamic polarization curve and is obtained around the middle of the passive range.Table 1 lists the mean values of corrosion parameters measured for the NiTi and NiTiNb alloys in physiological 0.9 wt% NaCl solution.
In the case of NiTiNb alloy from open circuit potential up to −50 mV, anodic curve indicates typical behaviour of activation polarization.From −50 mV up to 500 mV the curve shows a passive behaviour.In contrast with the NiTiNb sample, the NiTi alloy exhibited a breakdown in 0.9 wt% NaCl solution.On anodic polarization from open circuit potential, an initial passive region was observed prior to current increase, which is associated with breakdown potential ( bd ) for localized corrosion.Breakdown is evident from the sharp increase in anodic current around 170 mV.
Within the human body, commercial pure titanium (Cp-Ti) may be subjected to potentials of up to 500 mV [13].
Unfortunately, the same information has not been found for NiTi and NiTiNb alloys.By comparing the  bd value of NiTi alloy recorded in this work with the potential for Cp-Ti (500 mV), we can conclude that appreciable attention should be given to the passivation of NiTi alloy.
e corroded surfaces of both alloys were observed by means of scanning electron microscopy (SEM).SEM images, Figures 2(a role that Nb plays as an alloying element is by increasing the resistance of NiTi alloy to pitting corrosion. e impedance spectroscopic analyses were carried out on both samples for every 20 mV increase in the potential, beginning from OCP to the dissolution potential for NiTi alloy or to the 500 mV for NiTiNb alloy.Figures 3(a) and 3(b) depict a three-dimensional representation of the resulting impedance spectra of NiTi and NiTiNb alloys.From Figure 3, a successive change in the shape of each impedance spectrum with increase in potential was observed for both alloys.
In order to have a better understanding of the corrosion process, electrochemical impedance spectra at a potential in each region, namely, OCP, passive, and dissolution were selected from the potentiodynamic impedance spectra.
From the potentiodynamic impedance spectra three potentials were selected for NiTi alloy: OCP, −30 mV (passive region), and 200 mV (dissolution region), and the corresponding Bode plots are presented in Figures 4(a)-4(c).e advantage of the Bode plot is that the data for all measured frequencies are shown and that a wide range of impedance values can be displayed.
e Bode-phase plots for NiTi alloy at OCP show two relaxation constants, that is, two peaks are observed in the Bode-phase plots.A proposed equivalent circuit (EC) for the electrochemical interface may be seen in Figure 5(a).e EC is characterized by two parallel combination terms (RQ) in series with the resistance of the solution ( sol ).e two RQ elements can be attributed to charge transfer and separation processes at the alloy/oxide layer interface ( 1  1 ) and to the oxide layer formed on the surface ( 2  2 ).Large values of  2 , at OCP (order 10 5 Ω cm 2 ), were obtained con�rming the formation of a passive oxide layer with good corrosion protection ability.As the potential changes from OCP to −30 mV (passive region), the Bode-phase plots are in agreement with an EC with one time constant (Figure 5(b)).e EC consists of the parallel combination terms ( 2  2 ) in series with the resistance of the solution ( sol ) occurring between the sample and the reference electrode.e parameters  2 and  2 describe the properties of the passive oxide layer formed on NiTi alloy.As the potential changed from OCP to −30 mV,  2 increased (from 0.2 MΩ cm 2 to 1.1 MΩ cm 2 ) and  2 decreased.ese results seem to correspond to a slight thickening of the oxide layer.For the NiTi alloy polarized in physiological 0.9 wt% NaCl solution at 200 mV (dissolution region), within the low frequency range, an inductive behavior is observed.In this case the impedance data were �tted with the EC presented in Figure 5(c).For the NiTi alloys polarized in physiological 0.9 wt% NaCl solution at 200 mV, the value of  1 , representing the charge transfer resistance, is quite low, about 100 Ω cm 2 .e  3  element is usually attributed to the relaxation of the corrosion products on the electrode surface [19].In this case, the protectiveness of oxide layer is no longer present.In Figures 4(a All the resulting EIS parameters for the NiTi alloy in physiological 0.9 wt% NaCl solution are given in Table 2. Figures 6(a)-6(b) show the Bode plot of NiTiNb alloys at OCP and the passive region (300 mV).
e impedance spectra �tting for NiTiNb alloy at OCP in physiological 0.9 wt% NaCl solution was carried out using the EC presented in Figure 5(a).Table 2 shows the results of the �tting.e EIS spectra for NiTiNb alloy polarized in physiological 0.9 wt% NaCl solution at 300 mV (passive region) could be analyzed in terms of the EC for a passive layer that is depicted in Figure 5(b).e higher magnitude of impedance (around 10 6 Ω cm 2 ) was observed at the passive range of potential.Hence, the resistance of the formed layer in the passive region is higher than the resistance of the NiTiNb alloy at OCP, accounting for the higher corrosion resistance of the NiTiNb alloy.
T 2: Impedance parameters of NiTi and NiTiNb alloys in physiological 0.9 wt% NaCl solution, at 37 ∘ C.

Sample
Imposed potential  1 (kΩ cm 2 ) 10  e role of Nb plays as an alloying element is by increasing the resistance of NiTi alloy to pitting corrosion.NiTiNb alloy exhibited high corrosion resistance within the potential range used in the present study.

Conclusions
e addition of Nb to NiTi alloy facilitates passivation, produces the more stable passive layer by reducing anodic current density, and �nally promotes the corrosion resistance.
e dynamic electrochemical impedance spectroscopic (DEIS) technique has been used to evaluate the corrosion resistance of the NiTi and NiTiNb alloys in physiological 0.9 wt% NaCl solution.
e Bode plot for various regions, namely, OCP, passive region, and dissolution region was evaluated under the potentiodynamic condition, and each of these regions was discussed.For NiTi alloy a low and dangerous breakdown potential value (around 170 mV) may be recorded.e SEM images of corroded surfaces on NiTi alloy aer anodic potentiodynamic polarization test con�rm the pitting dissolution.

F 1 :
Linear anodic potentiodynamic polarization curves measured for NiTi and NiTiNb alloys aer immersion in physiological 0.9 wt% NaCl solution.

F 2 :F 3 :
SEM images of corroded surfaces on (a) NiTi and (b) NiTiNb alloys aer anodic potentiodynamic polarization tests.e impedance versus potential diagram of (a) NiTi and (b) NiTiNb alloys.T 1: e mean values of corrosion parameters measured for both samples in physiological 0.9 wt% NaCl solution.Samples pass (A/cm 2 )  bd

F 4 :
) and 2(b), were obtained aer anodic potentiodynamic polarization tests of both samples.e SEM images analysis indicated a different behavior of NiTINb alloy in comparison with the NiTi alloy characterized by pitting corrosion.When the NiTiNb alloy was polarized up to 500 mV, the dendritic morphology can be observed.e Bode plots for EIS data of NiTi alloy at (a) open circuit potential, (b) passive region, and (c) dissolution region.

F 5 :F 6 :
Equivalent circuits used for �tting the measured impedance spectra.Bode plots for EIS data of NiTiNb alloy at (a) open circuit potential and (b) passive region.
)-4(c), the experimental data are shown as individual points, while the theoretical spectra resulting from the �ts with ECs models are shown as lines.
) and 6(b), the experimental data are shown as individual points, while the theoretical spectra resulting from the �ts with ECs models are shown as lines.