Intergranular Corrosion of Cu-Al-Ni Alloy in 0.5 mol dm−3 H2SO4 Solution

The corrosion behaviour of Cu-Al-Ni alloy in 0.5 mol dm−3 H2SO4 solution was investigated by electrochemical methods including open circuit potential measurement, electrochemical impedance spectroscopy measurements, linear and potentiodynamic polarization. Measurements were performed in 0.5 mol dm−3 H2SO4 at temperatures 20 and 40 °C. After polarization testing, corroded electrode surfaces were ultrasonically cleaned in deionized water and examined by light and scanning electron microscopy, while the elemental composition at individual points of the alloy surface was determined by EDS analysis. The results of the investigations revealed the occurrence of intergranular corrosion on the Cu-Al-Ni surface, which became more pronounced with increasing temperature.


Introduction
Cu-Al-Ni alloys belong to the group of shape memory alloys (SMAs), a relatively new class of functional materials that can change from one crystallographic structure to another, in response to temperature changes or to other applied stresses. [1][2][3][4] If such alloys are plastically deformed at low temperature, they will completely recover their original shape when raised to a higher temperature. In the process of recovering their shape, these alloys can produce a displacement or a force as a function of temperature. This unique effect of returning to an original geometry after a large inelastic deformation is known as the shape memory effect (SME). The shape memory phenomenon results from crystalline phase changes known as "thermoelastic martensitic transformation". At temperatures below the transformation temperature, shape memory alloys are martensitic. In this condition, their microstructure is characterized by "self-accommodating twins". The martensite is soft and can easily be deformed by de-twinning. Heating above the transformation temperature converts the material to its high strength, austenitic phase. [5][6][7][8] Ni-Ti alloys are one of the most commonly used shape memory alloys in practice due to their outstanding properties, such as excellent shape memory effect, unique superelasticity, low elastic modulus, high corrosion resistance and biocompatibility. [9][10][11] Their disadvantages lie in their high production costs and low transformation temperature range (−100 to 100 °C), which is why they are now being replaced by cheaper alternatives like Cu-SMA alloys, for less demanding applications. The main advantages of Cu-Al-Ni alloys are their low price, relatively simple fabrication procedure, better machinability, better work/cost ratio, and high electrical and thermal conductivity compared to other shape memory alloys. 4,5,[12][13][14][15] Also, the stability of the two-way shape memory effect, which is very important in its application as actuators, is superior to that of Ni-Ti alloy. 16 Actuators based on Cu-Al-Ni alloy have been tested for usage as prosthetic actuators, and also actuators made from Cu-Al-Ni single crystals (produced by Nimesis Special Industries) have been reported by NASA to be very good candidates for space applications that need trigger temperatures between 100 and 200 °C. [17][18] Besides usage in the production of actuators, Cu-Al-Ni shape memory alloys have found their applications in various engineering devices, sensors, automotive and aerospace industries, robotics, etc. 3,12,13,19 Most studies on Cu-Al-Ni shape memory alloys have focused on their microstructure, mechanical and shape memory properties, and possible fields of application, but have neglected their corrosion behaviour, which is very vital for their practical use. The corrosion resistance of Cu-Al alloys has been attributed to the formation of a protective layer of alumina along with copper chlorides and oxides on the alloy surface. 20,21 Aluminium has a greater affinity towards oxygen than copper, and higher stability of the Al 2 O 3 than Cu 2 O. Some researchers have attributed the enhanced corrosion resistance to the formation of duplex-layered surface oxides composed of Cu 2 O × Al 2 O 3 × xH 2 O. 22 The presence of nickel is also important in the passivation of CuNi alloys because of its incorporation in the Cu(I) oxide, which reduces the number of cation vacancies that normally exist in Cu(I) oxides. 20,21,23 Again, a large number of the publications related to the corrosion properties of Cu-Al-Ni alloys have focused on the influence of chloride ions on the corrosion behaviour of Cu-Al-Ni alloys. 4,6,7,[23][24][25] The present investigation aims to determine the corrosion behaviour of Cu-Al-Ni alloy in 0.5 mol dm −3 H 2 SO 4 solution.

Materials and test solution
The investigated Cu-Al-Ni alloy, composed of 83.1 % Cu, 12.8 % Al, and 4.1 % Ni in wt%, was produced by melting pure components (Cu, Al, and Ni) in a vacuum induction furnace in a protective argon atmosphere. The melting furnace was connected to a device for continuous vertical casting, so the melting and casting of the alloy was carried out in an argon protective atmosphere. The resulting 8 mm diameter bars were cut with a precision Isomet cutter made by Buehler USA to small rollers, 1 cm in height. The preparation of the electrodes for electrochemical measurements consisted of soldering Cu-Al-Ni alloy rollers to insulated copper wires, followed by their insulation with Polyrepar S acrylate, to leave only one non-insulated roller base of 0.502 cm 2 bare, to be used as working electrode surface in contact with the electrolyte.
Mechanical preparation of samples before each experiment involved wet grinding of electrode surfaces with successive grades of emery papers (down to 2000 grit), using a Metkon Forcipol 1 V grinding/polishing machine. The surfaces were subsequently polished with Al 2 O 3 suspension (particle size of 0.3 µm), and then ultrasonically washed in ethanol solution and with deionized water.
The 0.5 mol dm −3 H 2 SO 4 was prepared from the 96 % H 2 SO 4 obtained by Kemika.

Electrochemical measurements
The electrochemical experiments were carried out in a three-electrode glass cell in which Cu-Al-Ni electrode served as a working electrode, saturated calomel electrode as reference electrode and platinum sheet electrode as an auxiliary electrode. The electrochemical cell was a jacketed model available for temperature control by connecting to a Huber Kiss circulator. A PAR 273A potentiostat/ galvanostat connected to a PAR M5210 lock-in amplifier was used for all electrochemical measurements. The open circuit potential measurements (E OC ) were recorded in 60 min time period immediately after electrode immersion in the electrolyte. Electrochemical impedance spectroscopy measurements were performed at E OC in a frequency interval from 0.01 Hz to 50 kHz using the signal amplitude of 10 mV. Linear polarization measurements were performed with the scanning rate of 0.2 mV s −1 in the potential region of ±20 mV around the corrosion potential, while potentiodynamic polarization experiment was performed in a wide potential range (−0.250 V (vs. E OC ) to 1.250 V), at a scan rate of 0.5 mV s −1 .

Surface study
Corrosion surface morphology, after potentiodynamic polarization measurements, was investigated using an optical microscope MXFMS-BD (Ningbo Sunny Instruments Co.) and with the scanning electron microscope TESCAN VEGA TS 5136 MM equipped with an EDS system for elemental analysis.

Results and discussion
The open circuit potential of Cu-Al-Ni alloy in 0.5 mol dm −3 H 2 SO 4 solution was measured immediately after immersion of the electrode in the solution at temperatures of 20 °C and 40 °C. The obtained E OC vs. time plots are shown in Fig. 1. In the Bode format (logarithm of impedance, Z, and phase angle respectively vs. logarithm of frequency, f) at high frequencies (f > 1 kHz), the impedance response is dominated by the electrolyte resistance. In the medium frequency region, the linear log |Z| vs. log f relationship with a slope close to −1 and a phase angle of ≈ −70°, reflect the capacitive behaviour of the system. At low frequency, the slope of ≈ −0.5 and the phase angle of ≈ −30° point towards the presence of a slow diffusion process. The overall impedance of the system and hence corrosion resistance of the alloy, decreases with increasing temperature, indicating that the electrode surface gets less protection.
The impedance response also shows more than one time constant, which reflects the diversity of the interfacial phenomena in the system under investigation. Fig. 3 shows the equivalent circuit proposed to fit the experimental data, which consists of an electrolyte resistance R el (≈ 7 Ω cm 2 ) connected with two time constants. The first time constant observed in the high frequency region results from the fast charge transfer process in the metal dissolution reaction. In this case, R 1 represents the charge transfer resistance, and Q 1 represents the constant phase element and replaces the capacitance of the electrical double layer. To account for the surface layer of corrosion products and diffusion process in the low frequency region, additional equivalent circuit parameters were introduced, such as R 2 for the surface layer resistance, Q 2 for constant phase element of the surface layer (Q 2 replaces the capacitance of surface layer), and a Warburg impedance W for the diffusion process.
where the constant Q accounts for a combination of properties related to both the surface and the electroactive species; jω is the complex variable for sinusoidal perturbation with ω = 2πf , and n is the exponent of CPE.
CPEs are used in the analysis of impedance data to account deviations that occurred by surface roughness on a microscopic level. For a smooth electrode, n approaches the unity and CPE is equivalent to a capacitor. Porous surfaces show lower n values.
The parameters of the equivalent circuit were evaluated using a simple least square fit procedure, and are presented in Table 1. The results obtained indicate that an increase in electrolyte temperature leads to a corresponding decrease in charge transfer resistance (R 1 ) and surface layer resistance (R 2 ), while the capacitance of the double layer (Q 1 ), the capacitance of the surface layer (Q 2 ) and the diffusion element (W) increase. This direction of change is attributed to the negative influence of temperature increase on the protective properties of the electrode/solution interphase.
Polarization measurements were conducted in a narrow range of potentials, around E OC (±20 mV), to determine the values of polarization resistance that corresponds to the corrosion resistance of the alloy. The values of the polarization resistance were determined from the slopes of the linear part of the curves shown in Table 2. Thereafter, potentiodynamic polarization measurements were performed over a wide range of potentials (from −250 mV to 1.250 V) to determine the corrosion parameters. The results are shown in Fig. 4.  Fig. 4 reveals that the anodic parts of the curve have a socalled "S-shaped" profile 28 , and can be divided into three separate areas; the Tafel region, the pseudo-passive area where the maximum and minimum anode current density occur (i.e., the decrease in the value of the anode current density due to the formation of a surface layer of corrosion products that slow down the alloy dissolution process), and a third region where the anode current density increases again. 26 Similar polarization behaviour was observed for Cu-Al-Ni alloy in NaCl solutions. 4,6,24 The slope of the anode Tafel direction of about 60 mV dec −1 suggests that the anodic process is determined by the diffusion of soluble compounds from the electrode surface into the bulk of the solution. 20 Awad and associates 29 have proposed the following copper dissolution mechanism: The adsorption of sulphate on the copper surface over a wide range of anode potentials has been experimentally demonstrated. 30 According to this mechanism, which has been accepted by some other authors, the disproportionation of adsorbed Cu (I) compounds leads to the formation of soluble Cu (II) compounds.
Corrosion parameters were determined from the polarization curves shown in Table 2. Corrosion rate v corr was calculated according to ASTM G 59-97 standard procedure, 31 wherein the corrosion rate (mm per year) can be determined from Eq. (5), in which E.W. is the equivalent weight of the corroding species in grams, and ρ is the density of the corroding material in g cm −3 .
From Table 2, it can be seen that increasing the electrolyte temperature leads to an increase in corrosion current density and a decrease in polarization resistance, corresponding to more severe corrosion of the Cu-Al-Ni alloy.
After potentiodynamic polarization measurements, the corroded Cu-Al-Ni electrodes were cleaned thoroughly in deionized water in an ultrasonic bath, dried, and examined using a light microscope in the dark field mode (Fig. 5). The images show intense corrosion at the grain boundaries (intergranular corrosion), which in some places goes into localized corrosion. Shallow round damage due to localized corrosion was also observed on the surface away from the grain boundaries.
After polarization measurements, a more detailed insight into the state of the electrode surfaces was obtained by SEM/EDS analysis (Figs. 6-8). Fig. 6a) and Fig. 7 show the state of the Cu-Al-Ni alloy surface after potentiodynamic polarization in 0.5 mol dm −3 H 2 SO 4 at a temperature of 20 °C, while Fig. 6b) and Fig. 8 show the electrode surface after polarization at 40 °C.
The surfaces are clearly without any corrosion products, and the images show occurrence of intense intergranular corrosion at the grain boundaries, which in some places is combined with severe localized corrosion, characterized by deep pits.
EDS analysis in marked points has shown that all alloying elements were present on the surface, with the percentage of aluminium and nickel being slightly lower than their percentage in the original alloy before corrosion dissolution.
In the investigation of the corrosion behaviour of Cu-Al and Cu-Al-Be alloys in H 2 SO 4 media, Kuo et al. had found the formation of the intergranular corrosion resulting from the preferential etching of grain boundaries when samples were polarised at 0.6 V. 28 Therefore, the reduction of Al content on the grain boundaries as well as on the spot nearby the grain boundary suggests preferential dissolution  The surface image of the Cu-Al-Ni alloy after polarization measurements by light and SEM microscopy indicated the appearance of intergranular corrosion along with pitting corrosion. Increasing the electrolyte temperature led to greater damages to the surface.
EDS elemental surface analysis at a certain point revealed the presence of all alloying elements on the surface. At higher temperatures, the copper content was slightly higher and the aluminium content lower, which could indicate its higher dissolution from the surface.