Electrochemical corrosion characterization of nickel aluminides in acid rain

Nickel aluminides have been extensively studied in recent decades to replace superalloys, in some components of aircraft turbines, because they have excellent corrosion resistance. Many industrial cities have the problem of air pollution, which has forced the study of the degradation of these alloys in the presence of acid rain. The aim of this work is to study the electrochemical corrosion behavior of nickel aluminides in a medium of simulated acid rain. Potentiodynamic Polarization, Linear polarization resistance curves, Rp, Nyquist data and Bode curves obtained by electrochemical impedance spectroscopy were used to study the corrosion behavior of two intermetallic compounds. The polarization curves show that both intermetallic Ni3Al and NiAl have very similar corrosion potential, showing a slightly nobler behavior the NiAl intermetallic. The intermetallic Ni3Al has an active-passive behavior where the anode branch presents a general dissolution of the alloy, indicating that it is under cathodic control. NiAl intermetallic shows an active dissolution region, followed by a passive behavior. At longer immersion times, Ni3Al intermetallic has a higher polarization resistance, which means a lower corrosion rate.


INTRODUCTION
Nickel aluminides such as NiAl and Ni 3 Al have received considerable attention for high temperature structural and coating applications in aerospace field [1-3], and due to high temperature strength, high corrosion resistance, high thermal and electrical conductivity, high melting temperature (1638°C), high Young's module (240GPa) have been regarded as a promising candidate for the high temperature applications [4,5]. Ni-base superalloys, Co alloys and steels are widely used in applications such as aircraft turbine blades and power plants due to present good corrosion resistance at elevated temperatures, however, the lower density (5.86g/cm 3 ) than superalloys and excellent high temperature oxidation resistance, allowed nickel aluminides being extensively studied in the past decades to replace superalloys [6][7][8]. As the nickel aluminides also has good properties for low temperature applications, studies of corrosion resistance at low temperature in aqueous media is interesting, because these materials will not always be at the operating temperatures, and corrosion damage during manufacture or maintenance may result in catastrophic failure during service [9].
Nickel aluminides have excellent corrosion resistance in most environments, accordingly, are being used as biomedical metallic materials to replace alloys such as stainless steel, most of the used materials for human implants include. NiAl intermetallics compounds have been investigated by electrochemical experiment in Hank's solutions [10]. Trinstancho et al. investigated the possible mechanism controlling the rate of the corrosion processes in hot corrosion resistance of Inconel 718 from 588°C to 900°C in Na 2 SO 4 , 80V 2 O 5 -20Na 2 SO 4 , NaVO 3 and natural ash [11].
Acidic deposition in some areas of the world has received increasing interest in the raising emission of SO 2 and NO x . This can be attributed to the accelerated rate of economic development [12]. The acid rain is primarily caused by the precursors of strong acids such as H 2 SO 4 and HNO 3 , resulted from fossil fuel combustion [13]. pH values vary depending upon country as well as the area, for example, on the southwestern coast of Europe, the pH values of samples collected fluctuating from 4.48 to 6.88 [14]. In 2008, China reported high acidity in rain water, the pH value of rain sample was 3.7 [13]. Nickel aluminides are prone to be attacked in atmospheric corrosion in urban and coastal areas, industrial and marine environments corrosion are more aggressive due to the highly corrosive nature of air pollutants [15]. Electrochemical impedance spectroscopy (EIS) is a technique which has been widely used in the study of aqueous corrosion and has proved effective in determining and understanding reaction mechanism and kinetics of corrosion processes.
The purpose of the present research work is to evaluate the corrosion behaviors of Ni 3 Al and NiAl alloys in simulated acid rain by using the electrochemical techniques to establish the involved mechanism in the process.

Materials
Cast ingots of binary NiAl (Table 1) were produced as round 20 g buttons with 99.99% pure Al, 99.97% pure Ni, on a water-cooled copper crucible under an Ar atmosphere. The buttons were remelted at least 8 times to ensure sufficient homogeneity. Ingots were cooled inside the crucibles in the furnace at room temperature and in this condition were used. Chemical composition was determined by SEM-EDX (field emission, JSM-7401F). The phase composition of the alloys was determined by the XRD method with a PANalytical diffractometer using CuKα radiation. The data identification was carried out using the Match software.

Sample preparation
The test specimens were cut into small parallelepiped pieces using a diamond tipped blade. Subsequently, the samples were encapsulated in fast cure acrylic which, in turn, were performed by grinding the specimens from 180 to 4000 grit paper for mechanical properties measurement by nanoindentation technique. For electrical connection, specimens were spot-welded to a wire, then were encapsulated in fast cure acrylic. Sample surfaces were ground to 1000 grade grit paper and then rinsed with ethanol in an ultrasonic bath for 20 minutes. Specimens with this surface condition were employed as the working electrode (we) in the electrochemical tests.

Mechanical properties by Nanoindentation technique
Nanoindentation measurements were performed in the polished surface of the samples with a matrix array 4 by 4 to determine the mechanical properties such as hardness and Young's modulus. The nanoindentation tests were carried out by an Agilent Nano Indenter G200, with a Berkovich diamond indenter with radius of 20 nm and maximum load of 5 mN. The equipment was calibrated using a standard fused silica sample, following the method described elsewhere [16]. Tests parameters as the constants of area function were C0=24.08, C1= -179.336, C2= 6721.28, C3= -24409.2, and C5= 18691.8. The analysis of the load displacement curve was carried out by Oliver and Pharr method [17]. Residual indentation of samples was recorded by the AFM NanoVision system attached to the nanoindenter system.

Electrochemical techniques
In order to simulate the acid rain conditions, the chemical composition for the environment for in vitro corrosion studies at pH 2.3, is given in Table 2. It was prepared with distilled water and analytical grade reagents following previously-reported concentrations [18].
Electrochemical tests were performed using an ACM Instruments potentiostat controlled by a computer. All potentials were measured using a Saturated Calomel Electrode (SCE) as reference electrode and a graphite bar counter electrode. All tests were performed at 19°C. The volume of electrolytic solution into the cell was 60 ml. The electrochemical techniques included potentiodynamic polarization curves, linear polarization resistance (LPR) and electrochemical impedance spectroscopy (EIS) measurements. Corrosion rates were calculated in terms of the corrosion current, I corr , by using linear polarization resistance curves.
Potentiodynamic polarization curves were obtained by varying the applied potential with respect to the open circuit potential (OCP), from -1500 mV up to +1500 mV at a scan rate of 1 mV/s in according with ASTM G3-89. A delay time of 60 min was setup until a stable reading was achieved, after that, electrochemical tests started. EIS tests were carried out by using a signal with amplitude of 20 mV (RMS) vs. the OCP between a frequency interval of 0.05 Hz to 10 kHz obtaining a measurement at 30 min, 12 and 24 hours of immersion and the number of points per frequency decade was 9. Measurements of LPR were performed by polarizing the specimen from +15 to -15 mV vs. OCP at a scan rate of 1 mV/s each 15 minutes during 24 hours. Corrosion products analysis of as-cast corroded samples was performed by scanning electron microscope (SEM) and energy X-ray dispersive spectroscopy (EDS).

Microestructural characterization
The as-cast microstructures of the Ni 3 Al and NiAl alloys are shown in Fig. 1a and 1b, respectively. As show in Fig.1a, Ni 3 Al alloy has a dendritic morphology with a grain size bigger than 300 μm with the presence of pores; higher Ni contents and high cooling rates induce dendrite formation [1]. No precipitates were observed in this intermetallic alloy, confirmed by X-ray diffraction pattern, where crystal structure corresponding to Ni 3 Al alloy and no second phases were detected, see Fig. 2a. In Fig. 1b, NiAl alloy is composed of elongated and equiaxed grains with average size around 500-700 μm, with the presence of pores. X-ray diffraction pattern of NiAl intermetallic alloy confirmed the crystal structure corresponding to NiAl without the presence of second phases and precipitates, as can be seen in Fig. 2b.   Figure 3a shows the load-displacement curves for the samples Ni 3 Al and NiAl and both of them provides a good repeatability for the sixteen indentations applied. The values of elastic modulus and hardness obtained are present in Table 3, where this behavior agrees with that reported by DONGMIN SHI et al. [19]. In the notch tip Berkovich, upon unloading from a maximum load of 5 mN there are no irregularities in size, shape or cracking, as show in the AFM image of Fig. 3b.

Potentiodynamic Polarization Curves
Polarization curves for the system NiAl are shown in Figure 4. The results show that both intermetalics have very similar corrosion potentials, being observed a slightly nobler behavior for the NiAl intermetallic. The polarization curve for Ni 3 Al alloys present a dissolution continuously behavior where the anodic branch presents a general dissolution of the alloy, indicating that it is under cathodic control. NiAl alloy shows an active region of dissolution in a range of potential from -320 to -160 mV/SCE, followed by passive behavior in the range of -116 mV to -24 mV/SCE. During the increase of the overpotential around 32 mV/SCE, presents a semi-passive behavior and current limit with fluctuations of density over a range of 104 mV/SCE, approximately, where are presents a pitting potential similar to observed in a NiAl intermetallic [20]. The values of corrosion potential (E corr ) and corrosion current density (i corr ) obtained by the method of intercepts are showed in table 3. The Tafel slopes were determined by a linear array of the polarization data in the range of ±150mV of the E corr , can be observed in the intermetallic Ni 3 Al, in the beginning the reaction rate is relatively slow in the cathodic slope, where the reactants are abundant and the reaction products move easily, being controlled by an activation process. When a so great negative potential is applied in an acid medium, the activation energy is given by the reduction of cathodic hydrogen on the sample surface, this gaseous hydrogen formation causes the H + concentration decrease, increasing the local pH value.   Figure 5 shows the results of the variation of corrosion potential, E corr , with respect to time. Ni 3 Al intermetallic presents an increase of the value of E corr within the first 2 hours from -400mV to -338 mV/SCE, subsequently, stabilized in the latter value. Similarly, the NiAl intermetallic presents a gradual increase of E corr in a range from -355 mV to -229 mV/SCE during the first 2.4 hours approximately, showing a fluctuation between 4 and 8 hours. Subsequently, E corr values are stabilized for both intermetallic, which is related to a more homogeneous form of corrosion including the surface of the localized corrosion sites [9]. The interaction between the surface and Clleads to change of E corr to more negative values [21], as observed for the Ni 3 Al intermetallic.  Figure 6b shows the images of the corroded surfaces of NiAl intermetallic, where pitting corrosion sites are seen and caused by the presence of NaCl [2]. The presence of corrosion products onto the metallic surfaces was not observed, indicating that they were soluble and dissolved during the washing step of the working electrodes. Ni surface did not show a significant attack, since the presence of lines generated by the grinding process is still seen. In Figure 6a, it can be seen that the dissolution of the metal in the corrosion process of Ni 3 Al occurs preferentially in the interdendritic regions. From the results of linear polarization (LPR) resistance, the corrosion rate of kinetics was calculated (see Figure 7) according to the standard ASTM G102, where the corrosion rate is expressed by equation 1:

Long term tests
Where EW is the equivalent weight of the alloy, i corr is the corrosion current density, ρ = density of the alloy, and K = constant. The equivalent weight was calculated with equations 2 and 3.
Where f i = the fraction of mass of the element i-th in the alloy, W i = atomic weight of the i-th element in the alloy and n i = the valence of the i-th element of the alloy. The corrosion current density was calculated from measurements of the polarization resistance, as follows: Where B = constant Stern-Geary, which was obtained with the anodic ( a ) and cathodic ( c ) Tafel slopes, with the following equation: The kinetics of the corrosion rate of both alloys of the NiAl system is shown in Figure 7, obtained from LPR measurements. The results show that during the first 2 hours the values of the corrosion rate of the intermetallic Ni 3 Al decrease from 3.8 to 1.5 μm/year, remaining constant at this value. On the other hand, the NiAl intermetallic presents a decrease of 9.7 to 3 μm/year, with CR fluctuations in the first 12 hours, a slight decrease occurs for the remainder of the test with a tendency to estabilize from this time. The Ni 3 Al intermetallic has a lower corrosion rate. This behavior of the CR is an effect caused by the sulfuric acid in the corrosion rate of the base Ni intermetallic [2].  Figures 8 and 9 show the impedance behavior of the Ni 3 Al and NiAl intermetallics in the acid rain solution and both alloys exhibit capacitive behavior at middle frequencies, increasing in magnitude with the immersion time. In the first 30 minutes of immersion, both alloys have diminished semicircles close to high frequencies associated to the activation behavior followed by a resistive behavior in Ni 3 Al and inductive for NiAl (figures 8a and 9a) associated with localized corrosion [22]. EIS results show, at 12 and 24 hours of immersion, a modification in the final behavior for both alloys. In this sense, EIS results for Ni 3 Al at 12 hours of immersion showed a resistive behavior followed by a capacitive behavior at high and middle frequencies, respectively, the latter describing a loop at 7 Hz and 55° ( fig. 8b). Similarly, at 24 hours of immersion, a resistive and capacitive behavior was described at high and middle frequencies (with a loop at 1.3 Hz and 63°, figure 8b), respectively. At 12 and 24 hours of immersion, in both alloys a first semicircle (resistance and capacitance) is defined, indicating that it was controlled mainly by an activation process and the influence of a second semicircle attributed to the pitting and a diffusion effect on the surface. At 24 hours of immersion, the main control was due to activation at the intermediate frequencies, could suggest linear semi-infinite diffusion response or current limit at low frequencies. Similar to Ni 3 Al at 12 and 24 hours of immersion, NiAl showed a resistive behavior followed by one capacitive at high and middle frequencies, respectively; this latter describing a loop at frequencies of 1.8 Hz and 1.3 Hz to 12 and 24 hours of immersion, respectively, figure 9b. At low frequencies, at these immersion times, the behavior presented for NiAl intermetallic is similar to the intermetallic Ni 3 Al, describing an activation behavior. Results of NiAl alloy showed a change in the size and shape of the diagrams, the capacitive arcs increased due to an increase in polarization resistance [23], showing a better performance in this environment as observed by the CR, figure 7.

DISCUSSION
Ni 3 Al characteristic microstructure consists of two phases, Ni -' Ni3Al corresponding to matrix and NiAl -' Ni3Al which corresponds to the dendritic phase. According to the phase diagram for Ni-Al, a peritectic reac-of the electrochemical reactions occurring at the electrode surface is necessary, figure 10. For the Ni 3 Al alloy, R s is the solution resistance, CPE dl the double layer capacitance and R ct charge transfer resistance by roughness effect, CPE 1 and R 1 are the capacitance and resistance of the diffusive and localized corrosion processes occurring on the metal surface. While for NiAl alloy, the elements are Rs electrolyte resistance, CPE dl double layer capacitance and R ct charge transfer resistance. The values obtained by Zview2 software from the circuits are reported in Table 4.

CONCLUSIONS
From the current results obtained in this research work, it can be concluded the following:  The Ni 3 Al-based alloy showed only localized attack (dissolution) in dendritic phases, its type of preferential dissolution is observed due to the formation of micro galvanic cells.  NiAl have a higher corrosion current density in the presence of HNO 3 for equal concentrations of H 2 SO 4 and HNO 3 .  The variation of the E corr can be attributed to the low corrosion resistance of the Al-oxide layer in an acid medium and the difference of the Al composition between intermetallics.  The width of the phase angle spectrum is shifted to the high frequency region due to the formation of a porous oxide.

ACKNOWLEDGMENTS
Engelbert Huape thanks to the CONACYT of the grant to carry out his doctoral studies and the memory of Ph.D. Alberto Martinez Villafañe