Corrosion Science 87 (2014) 504–516
Contents lists available at ScienceDirect
Corrosion Science journal homepage: www.elsevier.com/locate/corsci
Corrosion mechanism of Al, Al–Zn and Al–Zn–Sn alloys in 3 wt.% NaCl solution S. Khireche, D. Boughrara ⇑, A. Kadri, L. Hamadou, N. Benbrahim Laboratoire de Physique et Chimie des Matériaux (LPCM), Université Mouloud Mammeri de Tizi-Ouzou, BP 17 RP, 15000 Tizi-Ouzou, Algeria
a r t i c l e
i n f o
Article history: Received 14 January 2014 Accepted 10 July 2014 Available online 18 July 2014 Keywords: A. Aluminum A. Alloy B. EIS B. Polarization B. SEM
a b s t r a c t The effect of zinc and tin addition to pure aluminum was investigated in 3 wt.% NaCl solution. The corrosion behavior of the elaborated samples (Al, Al–Zn and Al–Zn–Sn) was studied by open circuit potential, Tafel plot and electrochemical impedance spectroscopy. For the microstructure characterization, Scanning Electron Microscopy and Energy Dispersive X-ray Spectroscopy were used. The aluminum activation increases in the following order: Al < Al–5Zn < Al–5Zn–0.1Sn < Al–5Zn–0.2Sn < Al–5Zn–0.4Sn. The impedance measurements and the microscopic observations confirmed the great activity of Al–Zn and Al–Zn–Sn compared to pure Al. The segregation at the grain boundaries leads to intergranular corrosion. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The electrochemical behavior of aluminum (Al) and its alloys have been the subject of various studies, due to its wide applications (construction, metallurgy, mechanical engineering, transport vehicle industry from the car to the plane and in food production of the packaging material industry ...). These applications are appropriate through the stability of oxide film on aluminum spontaneously formed in air or in contact with aqueous solutions as well as by means of the anodizing process. Another attractive property of aluminum is its standard reversible potential (E0Al3þ =Al ¼ 1:66 V (vs. SHE)), used as anode material in a battery system, and sacrificial anode in cathodic protection of steel in seawater [1]. For seawater, zinc and aluminum are more reactive therefore preferred. However, the protective film, namely Al2O3 formed on the surface of pure aluminum, makes it unusable as a sacrificial anode in cathodic protection of steel in seawater [2]. Pure aluminum can be then used for corrosion resistance applications [3,4]. To make the aluminum more active, some small proportions of metallic elements must be added. The alloying elements commonly added to aluminum are Cu, Mn, Si, Mg, Hg, Zn, In, Ga or Sn. Indeed, when adding alloying elements, which are less active than aluminum, an increase in the corrosion potential (Ecorr) is observed. Silicon, copper and manganese are nobler than aluminum, therefore Ecorr increases. Zinc and magnesium are more active
⇑ Corresponding author. Tel.: +213 (0) 550537125. E-mail address:
[email protected] (D. Boughrara). http://dx.doi.org/10.1016/j.corsci.2014.07.018 0010-938X/Ó 2014 Elsevier Ltd. All rights reserved.
than aluminum thus, the corrosion potential of the aluminum matrix decreases. With the addition of suitable elements such as, Hg, In [5], Ti, Ga [6], Zn or Sn [7], the pitting potential of Al is shifted to more negative values which allows a use, of the alloy, as anode materials in batteries or cathodic protection systems for underwater installations. Sacrificial anodes based on aluminum are considered as the most used in cathodic protection technology as mentioned hereafter, Al–Zn–Mg [8–11] Al–Zn–In [12–19] Al–Zn–Sn [20–26], Al–Sn– In [1,27] and others [28–31]. Without activator, the corrosion potential would be 75 mV (vs. SCE) and therefore insufficient to provide cathodic protection. Thus, the role of these elements is to remove the passivation layer of alumina (make more electronegative potential of the anode improving their performance). In, Hg, Sn, Ga and Bi shift the corrosion potential to more negative values by more than 0.3 V, i.e., close to 1.1 V (vs. SCE) in sea water, while Zn, Cd, Mg and Ba lowered it by 0.1–0.3 V. Mercury is considered as the most reactive element that gives a potential of 1.5 V (vs. SCE), but it is the most harmful to health and to the environment as well. The addition of 5% of zinc activates Al by decreasing the potential of the anode without forming pits [32], but above this level, no advantage is obtained and below 0.9%, the zinc has little influence. The activation of aluminum can also be obtained by the addition of small amounts of metallic cations, such as In3+, Ga3+, Hg2+, Sn4+ or Sn2+ in the electrolyte [5,16,17,20,21,33–38]. The main purpose of this study is the activation of aluminum. For this fact, the effect of additions of zinc and tin on the electrochemical behavior of pure aluminum in NaCl solution (3 wt.%, pH = 6.5) was demonstrated using various electrochemical
505
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
techniques such as: open circuit potential (OCP), Tafel plots and electrochemical impedance spectroscopy (EIS). 2. Experimental procedure
with an exposed area of 0.5 cm2. These electrodes are then mechanically polished with emery paper with a decreasing grain size from 1000 to 2400, then passed on a felt soaked with a suspension of 0.3 lm alumina and finally degreased with acetone under ultra-sounds, rinsed with distilled water and dried in a warm air.
2.1. Preparation of the alloys 2.2. Electrochemical measurements Binary (Al–5 wt.% Zn) and ternary alloys (Al–5 wt.% Zn–x wt.% Sn), with x = 0.1, 0.2 and 0.4) were prepared from pure metals as received from ChemPur Heraeus Feinchemikalien und Forschungsbedarf Gmbh: aluminum (99.99 wt.%), zinc (99.99 wt.%) and tin (99.99 wt.%), in order to ovoid impurities. The samples were melted in an induction furnace having a maximum temperature of 1200 °C. Al was put in crucible (without cover) and gradually heated until 720 °C in order to facilitate the addition of Zn and Sn after Al melting (660 °C). The choice of this temperature may be explained by their low melting point corresponding to 419.53 °C and 232 °C for Zn and Sn respectively. The samples were then transferred into a graphite mold and left to cool down at room temperature. Optical examination of the as-cast-alloys (Fig. 1) showed structural imperfections, due to gas adsorption and oxidation of the Al matrix. In order to overcome this, the samples were subjected to a melt and deoxidation. This technique is called ‘‘bath treatment by a solid flow’’. A burner flame was used, supplied from a gas in order to be powerful and high heat. The sample was then melt in a crucible with addition of a small amount of a solid called ‘‘Couveral 55’’ which brings back many oxides being in the liquid metal to the surface forming a solid layer and insulating our casting from the free air. The solid oxide layer was then removed from the liquid surface before casting in sand mold and cooling to room temperature. The samples were then machined to form cylindrical rod of 8 mm diameter and 30 mm in length. These cylindrical rods were mounted in a mold with resin, configuring a working electrode
(a)
Potentiodynamic polarization measurements were performed using an Autolab PGSTAT-30 (Eco Chemie, the Netherlands), controlled by a PC and an electrochemical cell with a three-electrode arrangement. The potential measurements are recorded with respect to the saturated calomel electrode (SCE). A Pt sheet was used as a counter electrode. The measurements were performed in NaCl solution (Prolabo) under suitable conditions (3 wt.%, pH = 6.5), freely exposed to the atmosphere and ambient temperature). The OCP of the samples, in the test solution were continuously monitored for a period of 5 h. The anodic and cathodic polarization curves were registered by sweeping the potential at a scan rate of 0.5 mV s1. The EIS measurements were carried out over a frequency ranging from 10 kHz to 10 mHz with a 10 mV amplitude signal at OCP, using an Autolab PGSTAT-30 driven by FRA 4.9 Software. For EIS data modeling and curves fitting method, the fit program Equivcrt is employed. This program is based on non-linear least squares fitting which allows non-ideal electrochemical behavior to be modeled. 2.3. Metallography and morphology analyses The metallographic structure of the elaborated alloys was studied using optical microscope (Optika Microscopes), after the surfaces have been mechanically polished and etched by immersing in Keller’s reagent (1 mL HF, 1.5 mL HCl, 2.5 mL HNO3 and 95 mL H2O) for 10 s. The corrosion morphologies were examined with
(c)
50 µm
50 µm
(d)
(b)
50 µm
50 µm
Fig. 1. Optical micrographs of (a) as-cast Al–5Zn, (b) melt Al–5Zn, (c) as-cast Al–5Zn–0.4Sn, (d) melt Al–5Zn–0.4Sn.
506
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
camera infinity and precipitate distributions on the alloys were analyzed by scanning electron microscope/energy dispersive Xray detector (SEM/EDS, Philips ESEM XL 30 model). Microstructural characterization of the electrodes was also examined after measuring the OCP and after potentiodynamic polarization (PP).
3. Results and discussion 3.1. Microstructure The optical micrographs of Al–5Zn and Al–5Zn–0.2Sn as-cast alloys (image (a) and (c) respectively) were compared to that of the melt samples (image (b) and (d)), as shown in Fig. 1. Etching along the grain boundary for Al alloys can be clearly observed on the melt alloy. The as-cast alloys are full of defects. Grain size of Al–5Zn is refined (73 lm as cast and 47 lm after melting). The melt homogenizes the microstructure of the solid solution by decreasing the grain density, from 250 grain cm2 (as-cast) to 960 grain cm2 and decreases defaults by the deoxidation. Fig. 2 shows the SEM images and EDS analysis of the marked sites on the polished aluminum matrix (image (a)) and its alloys (images (b–e)) without etching. Image (a) shows circular pits formed at intermetallic phases of commercial Al sample. The adding elements may be in the solid solution or in isolated particles of a second phase, intermetallic compounds or inclusions. Zn and Sn are poorly soluble in Al. So, they precipitate in phases, during the elaboration of the alloys [39]. Al has the highest melting point, so it crystallizes the first. In other words, the elements having a lower melting point enrich the liquid at the solid–liquid interface during the process of solidification. Image (b) corresponding to Al–5Zn, shows the presence of a separated phase from the main structure, the Al matrix (the grey regions). The bright region shows a preponderance of Zn (Table 1). Indeed the addition of 5 wt.% Zn results in the formation of a phase (Al–Zn) alloy, which is followed by the formation of b phase (Al–Zn rich in Zn). Such intermetallic particles have been reported in the literature [40,41]. Images (c–e) of Al–Zn–Sn alloys, exhibit a structure where segregation enriched in Zn and Sn precipitates in a-Al phase (the grey regions) and the grain boundaries are clearly shown. The dark areas correspond to the pores (defects) formed during solidification [11,42,43]. Their appearance is unavoidable during the addition of a low proportion of the lower melting point element. The bright regions in Fig. 2c and d, correspond to a second phase which is called ‘phase Sn’, stretching along the grain boundaries. Also, the semi-quantitative EDS analysis revealed a grain boundary compounds composition enriched with Zn and Sn, as shown in Table 1. It may be noted, from images (c–e) that the grain size decreases by increasing the content of Sn, following the distribution of Sn in a-Al phase, which can lead to uniform corrosion. Fig. 3 shows the SEM images of Al–5Zn and Al–5Zn–0.2Sn alloys etched with Keller’s reagent with the marked sites of EDS analysis. Image (b) shows dissolution of Zn segregation along the grain boundaries. As shown in Table 1 of the EDS analysis, the grain boundaries are rich in elemental Zn before etch and it degreases after attack. Image (c) shows intergranular corrosion at grain boundary because Sn readily segregates to the surface when the alloy is above the melting point of Sn [44]. The average grain size for Al–Zn–Sn alloys (39–35 lm) was smaller than Al–Zn (49 lm) and the density of grains was greater. This result indicates that addition of Sn can decrease grain size and increase density of grains of Al–Zn–Sn alloys, as shown in Fig. 4. It is easy to consider the role of Sn inclusions, because of its standard potential (E0Sn2þ =Sn ¼ 0:136 V (vs. SHE)) that is more positive than that of a-Al matrix (E0Al3þ =Al ¼ 0:166 V (vs. SHE)). Therefore, it is clear that
Sn inclusions will act as cathodic sites in the galvanic corrosion cell. Thus the region around Sn-rich precipitates will dissolve preferentially by dealloying and give rise to localized corrosion phenomena [45]. The second phase precipitates along the grain boundary and homogeneously distributed within the grain (the matrix) suggest that surface should enhance uniform corrosion attack of the alloys. The high concentration of Cl ions detected by EDS analysis (Table 1) suggested the absorption of Cl which induce corrosion and the presence of a salt film associated with a significant dissolution rate. Fe and Si are present in the alloys as natural impurities (EDX analyses). They form cathodic intermetallic particles/second phase (noble potential against Al matrix) and can lead to the formation of microcells and localized attack of Al [46,47]. The Si-rich phase does not affect localized corrosion of aluminum alloys as it has been reported in the literature. Silicon potential (0.17 V/(vs. SCE)) is more noble compared to Al. Mizuno et al. [48] found that the activity of Si is null, and they proclaimed the formation of the oxide SiO on the Si-rich phase and the nature of this oxide can prevent the cathodic reaction. The Fe-rich phases (Al3Fe) more noble than the matrix (0.47 V/(vs. SCE)), play the role of cathodes. The potential difference induced galvanic currents, which results in a localized attack at the periphery of Fe-rich sites [49]. The presence of Si in the phase reduces the effect of Fe on both the anodic and cathodic reaction rates [50]. 3.2. Open circuit potential measurements (OCP) Fig. 5 shows the OCP for pure Al, Al–5Zn, Al–5Zn–0.1Sn, Al– 5Zn–0.2Sn and Al–5Zn–0.4Sn, after immersion in 3 wt.% NaCl solution. As seen in Fig. 5, in the early stage of immersion, the potential of pure Al shifts to anodic values until to reach steady state. This evolution of potential is synonym of oxide film formation. The latter is of Al2O3 and Al(OH)3, in neutral media, according to Pourbaix diagram [51]. On the other hand, potential fluctuations were observed on the plot and were due to transient process. The oxide layer film is generally porous and allows Cl ions penetration where local breakdown of the passive film gives initiation of pits. Some pits can be repassivate and others will give formation of AlCl-4 complexe [47]. Otherwise, the addition of 5 wt.% of Zn to pure Al and different Sn contents, affects the behavior of the passive aluminum electrode. Indeed, the alloys exhibit more electronegative potentials then that of aluminum electrode. The open-circuit potential value of 0.96 V (vs. SCE) was attained by Al–Zn while that of ternary alloys (Al–Zn–Sn) was between 1.10 and 1.13 V (vs. SCE). The active behavior may be attributed to the lowering of pH resulting from the formation of Zn(OH)2 as indicated in the Pourbaix diagram [51] of Zn in neutral water. In the presence of sodium chloride, the film becomes porous by chloride penetration and forms zinc hydroxide chloride (Zn5(OH)8Cl22H2O) [52]. 3.3. Surfaces morphology of the corroded materials During the OCP measurements, the surface appearances of the corrosion process on the alloys are evident with the hydrogen bubbles, followed by the formation of corrosion products. After one day of free immersion in chloride solution, the samples were rinsed with distilled water and then dried in a warm air. At the end, the samples lost their reflective surface and are damaged. The images taken with a camera (Fig. 6) and SEM micrographs of the corroded surfaces (Fig. 7) confirm the above visual results. Fig. 6a shows that localized pitting sites are distributed uniformly over the entire surface of the pure aluminum electrode. In contrast, there is a generalized attack of the alloy Al–5Zn
507
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
(a)
(b′)
(b)
α-Al matrix
Grain boundary
Defect
(c)
(c′)
Zn-Sn rich precipitates
(d′)
(d)
Second phase
Fig. 2. SEM image for (a) Al, (b, b0 ) Al–5Zn, (c, c0 ) Al–5Zn–0.1Sn, (d, d0 ) Al–5Zn–0.4Sn, without chemical etching.
(Fig. 6b) which becomes more severe with the addition of Sn (Fig. 6c). Indeed, a formation of a white gelatinous precipitate in the solution was observed. In addition, SEM observations show a surface of pure aluminum having a thin corrosion product layer
(Fig. 7a), being thicker with Zn and Sn additions (Fig. 7b–d). In the other hand, the bright spots were observed through the ternary alloys surfaces which indicate the presence of Zn and Sn at the grain boundaries which cause the intergranular corrosion. The
508
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
Table 1 The EDS analysis results of the samples. Samples
Position
Al polished Al–5Zn polished
– – – – – – – – – – – – – – –
Element (wt.%) Al
Al–5Zn etched Al–5Zn–0.1Sn polished Al–5Zn–0.1Sn polarized
Al–5Zn–0.4Sn polished Al–5Zn–0.4Sn etched Al–5Zn–0.4Sn polarized
Solid solution core Solid solution core Grain boundary compounds Grain boundary compounds Solid solution core Grain boundary compounds Grey area White compounds Black area Solid solution core Grain boundary compounds Grain boundary compounds Grey area White compounds Black area
87.90 77.37 79.63 45.08 82.11 70.34 87.66 50.70 57.48 81.32 49.14 21.53 34.86 22.97 58.48
(a)
Zn 13.40 13.67 7.76 10.93 15.92 3.98 4.9 12.23 12.21 10.61 9.73 6.36 21.24
(c)
Sn
O
Si
10.36
1.75 9.24
6.70 32.29 6.97 3.30 8.36 28.21 30.29 6.47 11.21 16.29 35.72 35.88 16.14
10.44 8.43
21.08 55.88 13.03 32.20 4.14
Fe
2.44
12.43
7.38
2.17 6.66
Cl
0.38
5.80 6.30
2.59
Grain boundary corrosion attack
Pitting attack Tin
Corroded area
(d)
(b)
Grain boundary corrosion attack
grain boundary etch
Fig. 3. SEM image for (a, a0 ) Al, (b, b0 ) Al–5Zn, (c, c0 ) Al–5Zn–0.1Sn, (d, d0 ) Al–5Zn–0.4Sn, alloy etched by Keller’s reagent.
deposition of a metal activator depends on its electromotive force in the corrosion cell. The standard potentials of tin and zinc are 1.36 V (vs. SHE) and 0.76 V (vs. SHE), respectively. So, the deposition of Zn is less favored. It should be noted that activation of Al is probably due to its oxide layer instability because of the ZnAl2O4 formation, according to Ma et al. [53] as well as the deposition of Sn. According to El Shayeb et al. [20,21], the main formed oxides are Al2O3, ZnO, ZnOX, SnO and SnO2. 3.4. Polarization measurements Potentiodynamic polarization tests were carried out by sweeping the potential at a scan rate of 30 mV min1. Fig. 8 shows the potentiodynamic polarization curves of Al–5Zn and Al–5Zn–xSn
with (x = 0.1, 0.2 and 0.4 wt.%) after 24 h of immersion in 3 wt.% NaCl solution. The Tafel slopes of the cathodic curves have a value close to 120 mV/dec, which may be attributed to the water reduction and/or to the oxygen reduction, in aerated NaCl solution. These reactions cause an enrichment in OH ions which causes a local increase in pH. Inside the pits the pH is acid, this which generates a secondary reduction reaction of hydrogen evolution. During this step, hydrogen bubbles evolution is observed on the surface of the electrode while inside pits, this phenomenon is accompanied by the dissolution of the aluminum. Uniform corrosion of Al generates the formation of Al2O3 passive film according to the following reactions:
2Al þ 3H2 O ! Al2 O3 þ 6Hþ þ 6e
E ¼ 1:55 Vðvs: SHEÞ
ð1Þ
509
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
(a) 50
Grain size (µm)
40
30
20
10
0 Al
Al5Zn
Al5Zn0.1Sn Al5Zn0.2Sn Al5Zn0.4Sn
Al alloys
(b) 12
d 10-4 (Grain/cm2)
10 8 6 4 2 0 Al
Al5Zn
Al5Zn0.1Sn Al5Zn0.2Sn Al5Zn0.4Sn
Al alloys Fig. 4. Grain size distribution of the alloys.
-0.8
E/V (vs. ECS)
Al Al5Zn Al5Zn0.2Sn
Al5Zn0.1Sn Al5Zn0.4Sn
-0.9
-1.0
-1.1
0
1
2
3
4
5
t/h Fig. 5. Measurements of the open circuit potential for Al, Al–5Zn, Al–5Zn–0.1Sn, Al– 5Zn–0.2Sn and Al–5Zn–0.4Sn, after one day of immersion in 3 wt.% NaCl solution.
The later is characterized by the lower passivation current density (ipass 0.8 lA cm2). It is interesting to note that the addition of Sn to Al–5Zn alloy shifts the polarization curves to more electronegative potential and therefore higher current densities, which suggests a more active behavior of Al–Zn–Sn alloys compared to pure Al. The kinetic parameters deduced from the polarization curves are summarized in Table 2. The anodic slope, ba, decreases as Sn and Zn are added to the aluminum matrix, which results accrue dissolution. Moreover, it is found that the addition of Zn and Sn results in the evolution of the corrosion potential to more electronegative values and higher corrosion current densities or a decrease in the polarization resistance, the highest value being obtained for the Al–5Zn–0.4Sn alloy. As Zn is more electronegative than Al and Sn elements, it may preferably dissolve at anodic potentials, at grain boundaries where Zn and Sn are detected by EDS analyzes (Table 1). Sn and Zn being poorly soluble in the Al matrix, so they precipitate as different phases during the elaboration of the alloys [35,39]. This leads to the formation of galvanic cells Sn(Zn)/Al, which in the presence of Cl ions enhances the dissolution process of Zn compared to Sn. Therefore, after immersion in the chloride solution, the alloy introduced Al3+ and Zn2+, according to the zinc dissolution. The anodic reaction is accompanied by hydrogen evolution reaction on the surface of Sn inclusions, thereby producing a local alkalization, which breaks up the oxide film Al2O3 passing into solution as AlO2 ions. Breakdown of the passive state of aluminum in a chloride medium occurs by interaction between Al(OH)3 film and Cl ions which can generate AlCl3 compound [54–56] after surface saturation with corrosion products in acidic and chloride confined environments and after sufficiently long time. Then AlCl3 dissolves as aluminum complex ion [AlCl4] above the pitting potential [57]. The complexe [AlCl4] is trapped in the pores of the deposits around the intermetallic [47]. Under these conditions the passivation of Zn and its alloys occurs by the formation of Zn(OH)2 and ZnO [58] passive film which breaks up by Cl ions giving a non protective zincate film ZnOZnCl2 [59]. For more long immersion time, the products formed on Zn include zinc hydroxide chloride Zn5(OH)8Cl2H2O (hydroxychloride) [52,54] and different corrosion product containing Cl, Al, Zn and O [60]. The anodic branches of the polarization curves of the ternary alloys Al–5Zn–xSn exhibit an abrupt increase in the current that can be attributed to the active dissolution of Sn to Sn(II) species: Sn(OH)2, SnO and SnO2 according to the reactions reported by Ahmido et al. [59]. SnO and/or SnO2 is a passive film formed on Sn that is gradually altered by the adsorption of Cl ions giving rise 2 to soluble complexes SnCl [61] or to SnCl2 com3 and/or SnCl6 pound which can react with dissolved oxygen to form the oxychloride Sn4[OH]6Cl2 [58]. Zn2+ ions may redeposit on the cathodic sites, therefore, the hydrogen evolution reaction decreases [27] that result in the following reaction:
2Al þ 3Zn2þ ! 2Al The Al2O3 film is not stable in water. It become Al(OH)3 by hydration. In this case, the accumulation of the oxide/hydroxide film on the electrode surface slows the Al dissolution, as well as H2 evolution by blocking active sites. Passive behavior is observed on the anodic polarization curves of pure Al and Al–Zn alloy. However, Al–Zn–Sn alloys exhibit no passive region as presented in Fig. 8. This passive state is followed by a sharp increase in the current at the pitting potential, Ep. Al–Zn alloy exhibits a passive region with a current density, ipass, higher than that of pure Al.
3þ
þ 3Zn
ð2Þ
The surface appearances of the alloys, after polarization measurements, show that thick gray films of corrosion products are developed. After etching of the surface, a rough morphology of the alloys has been observed. It can be noted that severe attacks are responsible for the deep pits formation which is controlled by adsorption and/or penetration of Cl ions on the defects in oxide layer [61]. The SEM observations (Fig. 9) with EDS analysis (Table 1) showed a surface covered with corrosion products containing Al, Zn, Sn and O. The spherical particles contain bright inclusions
510
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
(a)
(b)
Pitting attack
(c)
Fig. 6. Appearance of (a) Al, (b) Al–5Zn and (c) Al–5Zn–0.2Sn after one day of immersion in 3 wt.% NaCl solution.
of Zn or ZnSn with a large amount of Sn in the case of Al–5Zn–0.4Sn alloy. As shown in the magnified image, the surface is rough with Sn addition due to the physical lost of Sn inclusions during the intergranular corrosion. In other words, the ternary alloy is probably dissolved with release of the rich ZnSn zone indicating a ‘‘crater’’. This becomes more important for the alloy containing 0.4 wt.% Sn (image (b)) where a quasi-uniform dissolution of Al–5Zn–0.4Sn alloy can be observed. This is probably due to the distribution of Sn in the Al matrix. The high anodic current, combined with the existence of a thick film indicates that these films are rich in point defects [16,37]. The presence of chlorides, detected by EDS analyses (Table 1), suggests the presence of a salt film which is associated with the high anodic dissolution rate. This indicates a high adsorption of Cl ions in the presence of Sn, which causes the breakdown of the passive state at more electronegative potentials. 3.5. EIS measurements The electrochemical characterization of Al and its alloys is complemented by EIS study at corrosion potential. The advantage of this technique is the separation of kinetic regimes operating at the electrochemical interface. The Nyquist and Bode diagrams are recorded at the corrosion potential after one day of immersion in aerated 3 wt.% NaCl solutions. For pure Al the Nyquist plot (Fig. 10a) is characterized by two depressed semicircles. In Bode plot (Fig. 10b) one can discriminate between two frequency time constants. This enables the determi-
nation of two capacitive contributions involved in the measured impedance. Note also that, Bode diagrams reveal dispersion due to R-CPE combination (slope value of 0.83 was found in log|Z| = f(log(f) plot). The origin of the time constant at high frequencies has often been attributed to the interfacial reactions, in particular to the oxidation reaction of aluminum. Whereas, the low frequency time constant can be related to the properties of the oxide layer/solution interface. In the literature, the time constant at low frequencies was interpreted in terms of a destruction made in the passive film [27,63]. The impedance of CPE is given by the following expression [64]:
Z CPE ¼
1 n Q ðjxÞ
ð3Þ
where Q is a CPE parameter, connected to the surface properties and the electroactive species, jx is the complex variable of the sinusoidal perturbation with = 2pf; n is the CPE exponent (1 < n < 1). When n 1, 0.5, 0 and 1, the CPE is equivalent to a capacitor, the element Warburg diffusion, a resistance and an inductance, respectively. Appearance of the CPE is a consequence of the dispersion of time constants which can originate from different physical phenomena. The various origins of CPE behavior described in the literature can be classified into two- dimensional (2D) distributions of time constants, which originate from the roughness and inhomogeneity and/or three-dimensional (3D) distribution of time constants which may be due to the adsorption, the porosity of the
511
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
(a)
(c)
(b)
(d)
Fig. 7. SEM image for, (a) Al, (b) Al–5Zn and (c) Al–5Zn–0.2Sn and (d) Al–5Zn–0.4Sn, after 24 h of immersion in 3 wt.% NaCl solution.
i/µA cm-2
Al Al5Zn Al5Zn0.2Sn 10
5
10
4
10
3
10
2
10
1
10
0
10
-1
10
-2
-1.4
-1.3
-1.2
-1.1
-1.0
films and the variation in the composition of deposits on the electrode. Fitting of the Al impedance diagram was done with the equivalent circuit shown in the insert in Fig. 10a. Solid lines show the results from fitting the experimental data according to the presented model. In this model R1 is the electrolyte resistance, R2 characterizes the charge transfer process corresponding to the dissolution reaction. CPE1 is the pseudo double layer capacitance. R3CPE2 circuit refers to the second capacitor loop at lower frequencies and is related to the layer of adsorbed products on the surface of Al. Table 3 shows the fitting values according to the equivalent circuit. According to the fitting results, the electrolyte resistance, R1, is about 9.5 X cm2. n1 value for CPE1 of about 0.9 was found for the HF capacitive loop of the charge transfer process for pure Al. The calculated Cdl value was close to 10 lF cm2 and R2 was about 12 kX cm2. Those values are typical of a film covered surface. The double layer capacitance is deduced from the CPE parameters according to the Brug equation [65]:
Al5Zn0.1Sn Al5Zn0.4Sn
-0.9
-0.8
-0.7
E/V (vs. SCE) Fig. 8. Potentiodynamic polarization curves for Al, Al–5Zn, Al–5Zn–0.1Sn, Al–5Zn– 0.2Sn and Al–5Zn–0.4Sn, after one day of immersion in 3 wt.% NaCl solution.
1 C dl ¼ ½Q dl ðR1 s þ Rct Þ
n1 1=n
ð4Þ
The HF charge transfer resistance, R2, can be explained in terms of Al dissolution through localized regions where, the resistive film controls highly the dissolution. Consequently, while the charge
Table 2 Kinetic parameters deduced from Fig. 8. Samples
Ecorr (V (vs. SCE))
ba (mV dec1)
bc (mV dec1)
Ep (V (vs. SCE))
ipass (lA cm2)
icorr (lA cm2)
Rp (kX cm2)
Pure Al Al–5Zn Al–5Zn–0.1Sn Al–5Zn–0.2Sn Al–5Zn–0.4Sn
1.02 1.07 1.09 1.16 1.17
340 180 60 65 65
160 130 130 115 100
0.73 0.96 – – –
0.8 7.5 – – –
0.5 3 10 15 30
78 4.4 1.7 1.2 0.6
512
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
a
a'
b
b’
Fig. 9. SEM image of (a) Al–5Zn–0.1Sn, (b) Al–5Zn–0.4Sn after polarization in 3 wt.% NaCl solution.
(a)
(b) Magnitude
Phase
Fit
5
4 60 3 40 2
- phase/deg
log(|Z|/Ω cm2 )
80
20
1
0 -2
-1
0
1
2
3
4
log (f/Hz) Fig. 10. Impedance diagram of Al and simulated spectra, obtained at corrosion potential after one day of immersion in 3 wt.% NaCl solution. (a) Nyquist and (b) Bode representations.
Table 3 Electrochemical impedance parameters of pure Al. R2 (kX cm2) 12.32
CPE1 (X1 cm2 sn) 4
0.3578 10
n1
Cdl (lF cm2)
R3 (kX cm2)
CPE2 (X1 cm2 sn)
n2
0.86
9.8
23.45
0.406 103
0.72
transfer resistance is great, the corrosion rate tends to decrease. n2 value for CPE2 is close to 0.72. This value may be attributed to the inhomogeneity of the layer of the corrosion products. Concerning Al–5Zn alloy, the impedance diagram (Fig. 11) reveals a HF capacitive loop and two inductive loops at lower fre-
quencies. This behavior is typical for systems undergoing active pitting. The appearance of the inductive loops at low frequencies could be attributed to the formation of intermediates when the alloy undergoes active dissolution. The inductive loops observed in the low frequency region indicate that the reaction mechanism
513
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
of dissolution involves elementary steps [66], associated with the adsorption of reaction intermediates in the active dissolution [67,68]. The inductive loop in the intermediate frequency range can be attributed to the adsorbed species such as Zn(OH)2 or Zn(OH)+. To explain the origin of the inductive loops in the case of pitted aluminum, adsorption and penetration of chloride ions through the oxide film were reported [10,69]. The number of inductive loops was always referred to the number of adsorbed species [70]. In order to enable an accurate analysis of the impedance diagram, the equivalent circuit model of the corrosion system reported in the insert in Fig. 11a was used. In this circuit, R1 represents the electrolyte resistance. R2 and CPE1, are charge transfer resistance and pseudo double layer capacitance. L1, L2, R3 and R4, are the inductances and resistors of the adsorbed species. Table 4 shows the equivalent circuit elements accorded to the proposed model. One can notice that R2 value has been reduced compared to pure Al, to about 2 kX cm2 indicating increase dissolution of the sample. This result indicates also that the oxide film becomes more defective with the additions of Zn. The CPE exponent n1 is about 0.81, indicating a deviation from ideal capacitive behavior. This deviation is attributed to the heterogeneous nature of the surface due to the presence of precipitates in the alloy and to pitting and local defects. To further investigate the effect of Sn addition on the corrosion performance of Al–5Zn, impedance diagram of the ternary alloy Al–Zn–Sn, are presented in Fig. 12a and b. As can be seen, the shapes of the curves are identical and it consists of a large capacitive loop at higher frequencies, inductive loop at intermediate frequencies, followed by a second capacitive loop at lower frequencies. The absence of a secondary inductive loop in impedance diagram of Al–5Zn–Sn means that the rate of adsorption of species is only significant in the absence of Sn. It is also interesting to notice that the magnitude and diameter of the HF loop decreases with increasing amount of Sn, meaning that the corrosion of Al–5Zn–Sn accelerates. It is notable that, the impedance of a pure inductance must be in the positive imaginary part of the real and imaginary plane. This is not our case, probably is due to the influence of the double layer capacitance. The HF loop can be attributed to a charge transfer process and the LF capacitive loop, to the corrosion products precipitation. The inductive loop probably corresponds to the pitting corrosion
(a)
occurred on the Al–Zn–Sn alloys surfaces, as it has been reported by Sun et al. [11] and Liu et al. [29]. Based on the above analysis, the proposed equivalent circuit, shown in Fig. 12a, fits well the impedance diagrams. It was slightly modified from that proposed by Chen et al. [70,71] and it was simpler than that used by Liu et al. [29]. HF loop is represented by an RC circuit; an ionic charge transfer resistance, R2, in parallel with a double layer capacitance, C1. An electrochemical inductance, L1, is placed in parallel with a second charge transfer resistance R3, and then in series with a pseudo-capacitance CPE1 to represent the inductive loop at intermediate frequencies followed by a capacitive loop. The parameter values are listed in Table 5. According to the parameter values presented in Tables 3–5, the transition from the passive state to the active one is accompanied by a decrease of the total impedance while the capacity increases, which is explained by the increase of the active surface as well as the presence of a salt film [1,27]. As can be mentioned, the presence of Sn leads to higher dissolution of the alloy. The charge transfer resistance (R2) values for Al– Zn–Sn were lower compared for Al–5Zn and pure Al, indicating that the corrosion rate is increasing with addition of Sn. Pitting corrosion of Al and its alloys is due to the localized penetration of ions through the passive oxide film. Pitting corrosion on Al alloy starts with the adsorption of chloride ions on the oxide surface. The adsorption of chloride ions occurs preferentially at some specific sites, such as micro-structural electronic defects [72], inclusions [50–75], dislocations [76] and grain boundaries [13]. Muller et al. [77] have reported that alloying of aluminum with 5% zinc results in high aluminum activation without any substantial pit formation. Hence, the presence of Sn as an alloying element is responsible for pit formation and gives a significant increase in the rate of corrosion of the alloy. The presence of Sn leads on one part to the formation of localized defects. These defects may contribute to the surface activation and breakdown of the oxide film on the Al–5Zn alloy surface. Note also, as it was well reported before, that the solid solubility of Sn in Al matrix is very weak and Sn diffuse preferentially at the grain boundaries. Therefore, addition of Sn creates local cathodic action sites and enhances the adsorption process of chloride ions in these sites, making the alloy susceptible to pitting corrosion. One can also notice that as Sn content increases, R2 decreases. As shown in Table 5, a content of 0.4 wt.% Sn has the lower resistances. As Sn amount increases, more defects will be created and
(b) Magnitude
Phase
Fit
3.5 80 60
2.5
40
2.0
20
1.5
0
- phase/deg
log(|Z|/Ω cm2 )
3.0
-20
1.0
-40 0.5 -2
-1
0
1
2
3
4
log (f/Hz) Fig. 11. Impedance diagram of Al–5Zn alloy, and simulated spectra, obtained at corrosion potential after one day of immersion in 3 wt.% NaCl solution. (a) Nyquist and (b) Bode representations.
514
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
Table 4 Electrochemical impedance parameters of Al–5Zn alloy. R2 (kX cm2) 2.064
CPE1 (X1 cm2 sn) 4
0.7696 10
n1
Cdl (lF cm2)
R3 (kX cm2)
L1 (H cm2)
R4 (kX cm2)
L2 (H cm2)
0.812
14
0.745
137.6
0.068
104
(a)
(b) Al5Zn0.1Sn Al5Zn0.4Sn
3.0
Al5Zn0.2Sn Fit
70
50 2.0
40 30
1.5
20 1.0
-phase/deg
log(|Z|/Ω cm2 )
60 2.5
10 0 -2
-1
0
1
2
3
4
log (f/Hz) Fig. 12. Impedance diagrams of Al–5Zn–Sn alloys with different tin contents, and simulated spectra, obtained at corrosion potential after one day of immersion in 3 wt.% NaCl solution. (a) Nyquist and (b) Bode representations.
Table 5 Electrochemical impedance parameters of Al–5Zn–Sn alloys with different tin contents. Tin (wt.%)
R2 (X cm2)
C1 (lF cm2)
R3 (X cm2)
L1 (H cm2)
CPE1 (X1 cm2 sn)
n1
0.1 0.2 0.4
253 127 94
210 320 382
101 50 37
35.60 17.80 10.11
0.0113 0.0226 0.0276
0.28 0.28 0.28
more Sn will be distributed along the grain boundaries and within the grain body, thus creating and activating more local cathodic action sites at the alloy-solution interface. These sites are favorable for aggressive anion attack, resulting in a more corrosion. The Cdl values varied from 10 to 380 lF cm2. The higher double layer capacitance, C1, value for Al–Zn–Sn alloy is typical to high corroded surfaces. As Sn content increases, C1 increases, suggesting more corroding areas or more active sites involved in the electrochemical reaction at alloy-electrolyte interface, which in turn delivers less corrosion resistance to the alloy. Consequently Both Al–Zn and Al–Zn–Sn alloys show a high activity compared to pure Al. These results are consistent with the results mentioned above (microstructure observations, OCP evolution and potentiodynamic polarization). 4. Conclusions The following conclusions can be drawn: 1. The addition of Zn and Sn produces, in each case, a significant activation of aluminum. This activation is manifested by the shift of the corrosion and pitting potentials towards more electronegative values. The degree of activation depends on the element and its proportion in the alloy. It increases in the following order: Al < Al–5Zn < Al–5Zn–0.1Sn < Al–5Zn–0.2Sn < Al–5Zn–0.4Sn 2. EIS has been applied for the analysis of the Al activation. The results are interpreted using appropriate equivalent circuits. For pure Al, two time constants were obtained: the first is
related to the oxidation reaction of aluminum while the second frequency time constant can be interpreted in terms of a pitting made in the passive film. 3. Al–Zn and Al–Zn–Sn alloys reveal an inductive behavior of the intermediates in the total reaction and attributed to the adsorbed species. 4. Al–Zn and Al–Zn–Sn alloys compositions may be appropriate for active anodes. 5. A good correlation is obtained between the results found by the OCP, PP technique, EIS and SEM for the characterization of electrochemical behavior of Al and its alloys.
Acknowledgements The authors would like to thank Dr. S. Benterfaia (Centre de Recherche Nucléaire de Draria: CRND-Algiers) for the elaboration of the alloys and they also wish to acknowledge Mr. A. Saifi (University of Tizi-Ouzou) for the SEM analyses. References [1] S. Gudic´, I. Smoljko, M. Kliškic´, Electrochemical behavior of aluminium alloys containing indium and tin in NaCl solution, Mater. Chem. Phys. 121 (2010) 561. [2] B.M. Ponchel, R.L. Horst, Aluminium performance of Al–Zn–Sn, Mater. Protect. 7 (1968) 38. [3] G.D. Davis, W.C. Moshier, T.L. Fritz, G.O. Cote, Evolution of the chemistry of passive films of sputter-deposited, supersaturated Al alloys, J. Electrochem. Soc. 137 (1990) 422.
S. Khireche et al. / Corrosion Science 87 (2014) 504–516 [4] Z. Szklarska-Smialowska, Insight into the pitting corrosion behavior of aluminum alloys, Corros. Sci. 33 (1992) 1193. [5] S.B. Saidman, J.B. Bessone, Activation of aluminium by indium ions in chloride solutions, Electrochim. Acta 42 (1997) 413. [6] D.O. Flamini, S.B. Saidman, J.B. Bessone, Aluminium activation produced by gallium, Corros. Sci. 48 (2006) 1413. [7] J.B. Bessone, D.O. Flamini, S.B. Saidman, Comprehensive model for the activation mechanism of Al–Zn alloys produced by indium, Corros. Sci. 47 (2005) 95. [8] S. Pengfei, Y. Geping, X. Baojia, L.U. Guoqi, Studies on the anodic behavior of aluminum electrodes in alkaline solution, J. Power Sources 45 (1993) 105. [9] A. Barbucci, G. Cerisola, G. Bruzzone, A. Saccone, Activation of aluminium anodes by the presence of intermetallic compounds, Electrochim. Acta 42 (1997) 2369. [10] P.L. Cabot, J.A. Garrido, E. Perez, A.H. Moreira, P.T.A. Sumodjo, W. Proud, EIS study of heat-treated Al–Zn–Mg alloys in the passive and transpassive potential regions, Electrochim. Acta 40 (1995) 447. [11] H. Sun, L. Liu, Y. Li, L. Ma, Y. Yan, The performance of Al–Zn–In–Mg–Ti sacrificial anode in simulated deep water environment, Corros. Sci. 77 (2013) 77. [12] A.G. Muñoz, S.B. Saidman, J.B. Bessone, Corrosion of an Al–Zn–In alloy in chloride media, Corros. Sci. 44 (2002) 2171. [13] A.G. Muñoz, S.B. Saidman, J.B. Bessone, Corrosion mechanism of Al–Zn–In alloys in chloride solutions, Latin Am. Appl. Res. 33 (2003) 275. [14] A. Zazoua, N. Azzouz, An investigation on the use of indium to increase dissolution of AlZn anodes in sea water, Mater. Des. 29 (2008) 806. [15] M.R. Saeri , A. Keyvani, Optimization of manganese and magnesium contents in As-cast aluminium–zinc–indium alloy as sacrificial anode, J. Mater. Sci. Technol. 27 (2011) 785. [16] A. Venugopal, V.S. Raja, AC impedance study on the activation mechanism of aluminium by indium and zinc in 3.5% NaCl medium, Corros. Sci. 39 (1997) 2053. [17] S.B. Saidman, S.G. Garcia, J.B. Bessone, Electrochemical behaviour of Al–In alloys in chloride solutions, J. Appl. Electrochm. 25 (1995) 252. [18] H.A. El Shayaeb, F.M. Abd El Wahab, S. Zein El Abedin, Electrochemical behaviour of Al, Al–Sn, Al–Zn and Al–Zn–Sn alloys in chloride solutions containing indium ions, J. Appl. Electrochem. 29 (1999) 473. [19] H.A. El Shayaeb, F.M. Abd El Wahab, S. Zein El Abedin, Role of indium ions on the activation of aluminium, J. Appl. Electrochem. 29 (1999) 601. [20] H.A. El Shayeb, F.M. Abd El Wahab, S. Zein El Abidin, Effect of gallium ions on the electrochemical behavior of Al, Al–Sn, Al–Zn and Al–Zn–Sn alloys in chloride solutions, Corros. Sci. 43 (2001) 643. [21] H.A. El Shayeb, F.M. Abd El Wahab, S. Zein El Abidin, Electrochemical behavior of Al, Al–Sn, Al–Zn and Al–Zn–Sn alloys in chloride solutions containing stannous ions, Corros. Sci. 43 (2001) 655. [22] D.R. Salinas, J.B. Bessone, Electrochemical behavior of Al–5%Zn–0.1%Sn sacrificial anode in aggressive Media: influence of its alloying elements and the solidification structure, Corrosion 47 (1991) 665. [23] M. Kliskic´, J. Radosevic´, L.J. Aljinovic, Behaviour of Al–Sn alloy on the negative side of the open-circuit potential, J. Appl. Electrochem. 24 (1994) 814. [24] S. Gudic´, J. Radosevic´, M. Kliskic´, Impedance and transient study of aluminium barrier-type oxide films, J. Appl. Electrochem. 26 (1996) 1027. [25] S. Gudic´, J. Radosevic´, M. Kliskic´, Impedance transient study of barrier films on aluminium and Al–Sn alloys, in: Proceeding of Symposium on Passivity and its Breakdown, Electrochemical Society, INC, New Jersey, 1997, p. 689. [26] J.-G. He, J.-B. Wen, X.-D. Li, G.-W. Wang, C.-H. Xu, Influence of Ga and Bi on electrochemical performance of Al–Zn–Sn sacrificial anodes, Trans. Nonferr. Metal. Soc. 21 (2011) 1580. [27] S. Gudic´, I. Smoljko, M. Kliškic´, The effect of small addition of tin and indium on the corrosion behavior of aluminium in chloride solution, J. Alloys Compd. 505 (2010) 54. [28] D. O Flamini, S.B. Saidman, Electrochemical behavior of Al–Zn–Ga and Al–In– Ga in chloride media, Mater. Chem. Phys. 136 (2012) 103. [29] F. Liu, J. Zhang, C. Sun, Z. Yu, B. Hou, The corrosion of two aluminium sacrificial anode alloys in SRB-containing sea mud, Corros. Sci. 83 (2014) 375–381. [30] W. Xiong, G.T. Qi, X.P. Guo, Z.I. Lu, Corros. Sci. 53 (2011) 1298–1303. [31] J.-G. He, J.-B. Wen, X.-D. Li, Effects of precipitates on electrochemical performance of Al sacrificial anode, Corros. Sci. 53 (2011) 1948–1953. [32] J.C. Muller, J.R. Galvele, Pitting potential of high purity binary aluminium alloys-II. Al–Mg and Al–Zn alloys, Corros. Sci. 17 (1977) 995. [33] Q. Li, N.J. Bjerrum, Aluminum as anode for energy storage and conversion: a review, J. Power Sources 110 (2002) 1. [34] J.F. Equey, S. Muller, J. Desilvestro, O. Haas, Electrochemical properties of aluminum in weakly acidic sodium chloride solutions: II. Influence of the electrolyte additives Hg2+, In3+, Ga3+, and Sn2+, J. Electrochem. Soc. 139 (1992) 1499. [35] W.M. Carroll, C.B. Breslin, Activation of aluminium in halide solutions containing ‘activator ions’, Corros. Sci. 33 (1992) 1161. [36] C.B. Breslin, W.M. Carroll, The activation of aluminium by indium ions in chloride, bromide and iodide solutions, Corros. Sci. 34 (1993) 327. [37] A. Venugopal, V.S. Raja, The self regulating nature of In on the potential of Al in 3.5% NaCl solution, Corros. Sci. 39 (1997) 1285. [38] C.B. Breslin, A.L. Rudd, Activation of pure Al in an indium-containing electrolyte – an electrochemical noise and impedance study, Corros. Sci. 42 (2000) 1023. [39] M. Nestoridi, D. Pletcher, R.J.K. Wood, S. Wang, R.L. Jones, K.R. Stokes, I. Wilcock, The study of aluminium anodes for high power density Al/air batteries with brine electrolytes, J. Power Sources 178 (2008) 445.
515
[40] D.R. Salinas, S.G. Garcia, J.B. Bessone, Influence of alloying elements and microstructure on aluminium sacrificial anode performance: case of Al–Zn, J. Appl. Electrochem. 29 (1999) 1063. [41] S.-I. Pyun, W.-J. Lee, The effect of prior Cl ion incorporation into native oxide film on pure aluminium in neutral chloride solution on pit initiation, Corros. Sci. 43 (2001) 353. [42] G.T. Qi, Y.B. Qiu, Y.N. Zhao, Q.Z. Cai, The attack initiation of Al–Zn–In–Sn anode by the segregation concentrating Zn, Sn and In, Mater. Corros. 60 (2009) 206. [43] J. Ma, J. Wen, The effects of lanthanum on microstructure and electrochemical properties of Al–Zn–In based sacrificial anode alloys, Corros. Sci. 51 (2009) 2115. [44] J. Tan, K. Nisancioglu, Effect of small amounts of alloyed tin on the electrochemical behavior of aluminium in sodium chloride solution, Corros. Sci. 76 (2013) 219–230. [45] M. Mandal, A.P. Moon, G. Deo, C.L. Mendis, K. Mondal, Corrosion behavior of Mg–2.4Zn alloy micro-alloyed with Ag and Ca, Corros. Sci. 78 (2014) 172–182. [46] E.L. Din-Stirbu, Thesis of Doctor, Comportement à la corrosion des alliages d’aluminium utilisés dans l’industrie automobile pour la fabrication de radiateur de chauffage, Romania, Lyon-France 2005. [47] C. Vargel, Corrosion of Aluminium, Elsevier, 2004. [48] K. Mizuno, A. Nylund, I. Olefjord, Surface reactions during pickling of an aluminium–magnesium–silicon alloy in phosphoric acid, Corros. Sci. 43 (2001) 381–396. [49] K.A. Yasakau, M.L. Zheludkevich, S.V. Lamaka, M.G.S. Ferreira, Role of intermetallic phases in localized corrosion of AA5083, Electrochim. Acta 52 (2007) 7651–7659. [50] Z. Szklarska–Smialowska, Pitting corrosion of aluminum, Corros. Sci. 41 (1999) 1743–1767. [51] M. Pourbaix, Atlas d’équilibres électrochimiques, Cebelcor, Ed. GauthierVillars, 1963. [52] M. Mouanga, P. Berçot, J.Y. Rauch, Comparison of corrosion behaviour of zinc in NaCl and in NaOH solutions. Part I: corrosion layer characterization, Corros. Sci. 52 (2010) 3984–3992. [53] J. Ma, J. Wen, W. Zhai, Q. Li, In situ corrosion analysis of Al–Zn–In–Mg–Ti–Ce sacrificial anode alloy, Mater. Charact. 65 (2012) 86. [54] T. Kaewmaneekul, G. Lothongkum, Effect of aluminium on the passivation of zinc–aluminium alloys in artificial seawater at 80 C, Corros. Sci. 66 (2013) 67. 2 [55] S.-I. Pyun, S.-M. Moon, S.-H. Ahn, S.-S. Kim, Effects of Cl, NO 3 and SO4 ions on anodic dissolution of pure aluminium in alkaline solution, Corros. Sci. 41 (1999) 653–667. [56] S.S. Kale, V.S. Raja, A.K. Bakare, Environmentally assisted cracking behavior of Al–Zn–Mg–Cu–Cr alloy bonded with polymer matrix composite stiffener, Corros. Sci. 75 (2013) 318–325. [57] R.T. Foley, Localized corrosion of aluminum alloys – a review, Corrosion 42 (1986) 277–288. [58] U.S. Mohanty, K.-L. Lin, Electrochemical corrosion behaviour of lead-free Sn– 8.5 Zn–x Ag–0.1 Al–0.5 Ga solder in 3.5% NaCl solution, Mater. Sci. Eng. A 406 (2005) 34. [59] A. Ahmido, A. Sabbar, H. Zouihri, K. Dakhsi, F. Guedira, M. Serghini-Idrissi, S. El Hajjaji, Mater. Sci. Eng. B 176 (2011) 1032. [60] X. Zhang, C. Leygraf, I.O. Wallinder, Atmospheric corrosion of Galfan coating on steel in chloride-riche environments, Corros. Sci. 73 (2013) 62–71. [61] S.A.M. Refaey, G. Schwitzgebel, Electrochemical impedance spectroscopic investigation of dissolution, passivation and pitting corrosion of tin in Na2CO3 solution and the effect of Cl and I ions, Appl. Surf. Sci. 135 (1998) 243. [62] A.G. Muñoz, J.B. Bessone, Pitting of aluminium in non-aqueous chloride media, Corros. Sci. 41 (1999) 1447. [63] J.B. Bessone, D.R. Salinas, C.E. Mayer, M. Ebert, W.J. Lorenz, An EIS study of aluminium barrier-type oxide films formed in different media, Electrochim. Acta 37 (1992) 2283. [64] J.R. Macdonald, Impedance spectroscopy and its use in analyzing the steadystate AC response of solid and liquid electrolytes, J. Electroanal. Chem. 223 (1987) 25. [65] G. Brug, A.L.G. Van Den Eeden, M. Sluyters-Rehbach et J.H. Sluyters, The analysis of electrode impedances complicated by the presence of a constant phase element, J. Electroanal. Chem. 176 (1984) 275. [66] M. Itagaki, A. Taya, K. Watanabe, K. Noda, Deviations of capacitive and inductive loops in the electrochemical impedance of a dissolving iron electrode, Anal. Sci. 18 (2002) 641. [67] S.M.A. Shibli, S. George, Electrochemical impedance spectroscopie analysis of activation of Al–Zn alloy sacrificial anode by RuO2 catalytic coating, Appl. Surf. Sci. 253 (2007) 7510. [68] I. Epelboin, M. Keddam, J.C. Lestrade, Faraday impedances and intermediates in electrochemical reactions, Faraday Disc. Chem. Soc. 56 (1973) 264. [69] A.N. Ehteram, Evaluation of inhibitive action of some quaternary Nheterocyclic compounds on the corrosion of Al–Cu alloy in hydrochloric acid, Mater. Chem. Phys. 114 (2009) 533. [70] W.-C. Chen, T.-C. Wen, A. Copalan, Negative capacitance for polyaniline – an analysis via electrochemical impedance spectroscopy, Synth. Met. 128 (2002) 179. [71] W.-C. Chen, T.-C. Wen, C.-C. Hu, A. Copalan, Identification of inductive behavior for polyaniline via electrochemical impedance spectroscopy, Electrochim. Acta 47 (2002) 1305. [72] G.E. Thompson, G.C. Wood, in: J.C. Scully (Ed.), Treatise on Material Science and Technology, Academic Press, New York, vol. 23, 1983, p. 18.
516
S. Khireche et al. / Corrosion Science 87 (2014) 504–516
[73] G.O. Ilevbare, G.T. Burstein, The role of alloyed molybdenum in the inhibition of pitting corrosion in stainless steels, Corros. Sci. 43 (2001) 485. [74] T. Suter, R.C. Alkir, Microelectrochemical studies of pit initiation at single inclusions in Al 2024-T3, J. Electrochem. Soc. 148 (2001) B36–B42. [75] E. McCafferty, Introduction to Corrosion Science, New York-USA, 2010.
[76] D. Song, A.-B. Ma, J.-H. Jiang, Corrosion behavior of ultra-fine grained industrial pure Al fabricated by ECAP, Trans. Nonferrous Met. Soc. China 19 (2009) 1065– 1070. [77] I.L. Muller, J.R. Galvele, Pitting of high purity binary aluminium alloys – II. Al– Mg and Al–Zn alloys, Corros. Sci. 17 (1977) 995–1007.