Elsevier

Corrosion Science

Volume 52, Issue 1, January 2010, Pages 90-103
Corrosion Science

Stable pit formation on AA2024-T3 in a NaCl environment

https://doi.org/10.1016/j.corsci.2009.08.043Get rights and content

Abstract

Scanning vibrating electrode (SVE), particle induced X-ray emission spectroscopy (PIXE) and standard electrochemical measurements were used to study the establishment of stable pits on AA2024-T3 in neutral sodium chloride solution (0.1 M NaCl). Pits were allowed to develop until hydrogen evolution was observed. Typical current at the mouth of the pits were in the vicinity of 1 mA. PIXE maps revealed the intermetallic (IM) particle distributions in the surface as well as significant chloride buildup around the pits. A significant fraction of the small selection of pits examined here appeared to have an S-phase particle (or remnant) within 20 μm of a AlCuFeMn type IM particle suggesting a coupling between the two. The electrochemistry of the coupling between different IM particle types was further investigated using potentiodynamic scans in 0.1 M aqueous NaCl solution of macroscopic electrodes made according to the IM particle compositions. Current densities at the open circuit potential of AA2024-T3 were largest, typically (0.2 up to 1 mA/cm2) for phases that were anodic with respect to AA2024-T3. Coupling of the IM particles was examined by comparing the degree of clustering around chloride attack sites to the average particle density for each map. There were significantly higher number of IM particles surrounding pit sites than the average IM particle densities indicating that local clustering played an important role in pit initiation.

Introduction

Pitting caused by chloride attack is one of the most serious forms of corrosion for aluminium alloys and can lead to other types of corrosion as well as structural degradation. Pitting in chloride electrolytes has been the subject of intense study with the aim of understanding the mechanism of initiation and propagation of pits on aluminium and its alloys [1], [2], [3], [4]. The mechanism of initiation depends on the purity of the underlying alloy. For very high purity aluminium, breakdown of the surface oxide may occur via one (or a combination) of the following mechanisms (i) penetration of the oxide by chloride ions, (ii) thinning of the oxide by chloride ions or (iii) through flaws in the oxide at the nanoscopic level [2], [3], [4], [5], [6], [7]. Aluminium alloys, however, have a range of intermetallic (IM) particles with compositions that depend on the alloy additions and processing conditions [8], [9]. These particles are generally the initiation sites for corrosion in commercial alloys since flaws in the oxide surrounding these IM particles and the galvanic coupling between the IM particles and the surrounding matrix promote localized corrosion [8], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31].

In recent years there has been a trend towards a statistical description of pit initiation using electrochemical and microstructural approaches. The electrochemical approach describes pitting in terms of nucleation events, metastable pitting and stable pitting, both in ferrous-based metals [32], [33] and aluminium and its alloys [34], [35], [36], [37]. The focus of this approach is the “identification of electrochemical factors that promote the transition from metastable to stable pit growth” [34]. The use of this terminology comes from potentiostatic measurement of currents made at either open circuit potential (OCP) or potentials between the OCP and the breakdown potential [34], [35], [36], [37]. For aluminium and ferrous metals, characteristics of these measurements include current transients and increases in total current which have been described as nucleation, metastable and stable pitting events.

For ferrous metals these events have been described as follows: Nucleation events are spontaneous corrosion and passivation events which typically have a lifetime of a few seconds [32]. Fig. 1 displays a schematic for the metastable and stable pitting events based on the authors’ experience with AA2o24-T3 and reported elsewhere [35]. The growth in current (Δtg) for a metastable pitting event is steep but peaks and relaxes (Δtr) more slowly than during growth maintaining a corrosion current over tens of seconds, possibly through a ruptured oxide coating [32]. Repassivation occurs when this oxide is fully ruptured and the diffusion path is not long enough to maintain critical pitting conditions at the head of the pit. For these events, it is thought that corrosion occurs beneath the surface oxide to the point at which there is hydrogen production which bursts the oxide and dissipates the anolyte solution into the general solution causing repassivation.

Stable pitting again begins in a similar fashion to metastable pitting events, but has a higher initial current density. It is suggested that when the oxide cap over the pit fully ruptures then the pit is deep enough to maintain a concentration gradient and an acidic anolyte solution is established at the pit head. Current transients of this type have been observed for both pure Al [34], [35], [36], AA2024-T3 [37], [38] and AA7010-T76 [39]. The characteristics of these events may involve a step increase in the baseline current and also current transients on top of the step in baseline current. The initiation of localized corrosion under the surface oxide will precede the detection of the stable pitting event as all the corrosion is occurring beneath the surface oxide at a rate (current) that increases from an initially undetectable value. The length of the time delay from corrosion initiation to detection of a metastable pitting event is unknown, but clearly, if an oxide ruptures, then that can only occur as a result of prior corrosion. The subject of this paper is the identification of the initiation sites for corrosion.

In this electrochemical approach, characteristics of the current transients such as peak current and size are used to determine the pit stability product. As early as 1976, Gavele [40] proposed the quantity i · rpit (pit stability product) where i is the pit current density and r is the pit depth, in modeling the chemistry within pits and the establishment of chemical conditions for the propagation of a stable pit. He estimated that i · rpit should be ∼10−2 A/cm for the establishment of a stable pit. For pure aluminium, Pride et al. [34] proposed that the boundary between metastable and stable pitting occurred when the rise in the pit current was steep enough to maintain a i · rpit > 10−2 A/cm.

The first limitation of the purely electrochemical approach for an alloy like AA2024-T3 is that it provides no microstructural information on the nature of the corrosion sites that lead to a metastable current response. While there have been numerous studies of localized corrosion in the form of trenching and S-phase dissolution on commercial alloys such as AA2024-T3 [14], [16], [18], [23], [29] the recent work of Ilevbare et al. [20] attempted to make a connection between the microstructure and the types of corrosion events behind the current transients. They concluded that metastable pitting occurred at, or adjacent to IM particles in the surface. The metastable pitting resulted in local dissolution at both cathodic and anodic IM particles which eventually stopped. Even in the case of S-phase particles, corrosion only resulted in etchout of the particle since the i · rpit was generally below that required to etchout the adjacent matrix. So the connection between the electrochemical definition of stable pitting and microstructural observations have never really been established for aluminium alloys.

The second drawback of the purely electrochemical approach also revolves around the relationship between the microstructure and the observation of a stable pitting current. There is an assumption that because the metastable and stable pitting events “look” the same electrochemically, i.e., a sharp rise in current characterizes both events, the population of stable pitting events is simply on the extreme end of the population statistics of metastable pitting events. This view appears to be confirmed for steels since the metastable and stable pit nucleation frequencies appear to be have a similar behaviour [41]. This interpretation has been applied to pure aluminium by Pride et al. [34], as well as by the authors for AA2024-T3 for modeling for structural health management in airframes [37], [38], [42], [43]. However, Sasaki et al. [44] have found, that while the current transients observed during the early stages of pitting on pure aluminium might be due to metastable pitting, at longer times they are due to activation and passivation events in pits and even between pits.

Thus there is a growing interest in characterizing and understanding the types of sites that lead to stable pitting. In this context clustering of IM particles has emerged as an area of importance in understanding pit stabilization however, only a small number of studies allude to clustering as part of the stable pitting process [12], [13], [18], [20], [22], [24], [45]. Clustering may occur at several different length scales and perhaps even times scales and in a separate report the authors also explore the influence that much larger clusters (a few hundred particles) have on stable pit initiation through the establishment of corrosion features such as rings of corrosion product a few hundred microns in diameter [46]. The current work on clustering, however, focuses on the influence that small clusters (perhaps 10–20 IM particles) have on corrosion initiation. The central premise of previous reports on the role of clustering in stable pit initiation is that clusters drive the electrochemical dissolution of the surrounding matrix with the interaction between particles having no formal role. In this context Wei and co-workers [18], [22], [45] have reported that clustering on AA2024-T3 and AA7075-T6 may be responsible for stable pitting as have Juffs and co-workers [47], [48]. Ilevbare et al. [20] also suggested that large clusters of particles were necessary to establish conditions where S-phase etchout could continue on to stable pitting. These studies [18], [20], [22] conclude that clustering influences stable pit initiation through excessive trenching which leads to particle fallout and other attack. In the work quoted above the clustered IM particles attack the matrix, however, as suggested by Chen et al. [18], coupling between IM particles might also lead to stable pitting. In the most comprehensive work to date in this area, Leblanc and Frankel [49] have examined the influence of AlCuFeMn IM particles on the matrix and S-phase particles using a masking technique where they opened ‘windows” over IM particles using an AFM. They found that while S-phase particles may leave a Cu-rich layer after dealloying, not all this Cu promotes further corrosion. They also noted that AlCuFeMn particles drove both dissolution of the matrix and S-phase particles. The area of matrix that is attacked increases faster than the area percent of the AlCuFeMn particles [49]. They also noted that no attack occurred in areas that only contained AlCuFeMn IM particles (for times typically between 30 and 120 min that they examined). In other localized studies using microelectrodes, Suter and Alkire [50] demonstrated that the local OCP of the surface was dominated by the weakest point, thus an area with more S-phase particles will have a lower OCP and be more susceptible to attack.

The aim of this paper is to examine the influence that coupling of anodic to cathodic IM particle types has on corrosion in AA2024-T3. To achieve this objective both the stable pitting in commercial AA2024-T3 and the electrochemical characteristics of intermetallic compounds (IMC) were examined. For the former, the degree of clustering of IM particles in AA2024-T3 and the characterization of pits developed on polished AA2024-T3 exposed to NaCl solution were examined using the nuclear microprobe technique of particle induced X-ray emission (PIXE) spectroscopy to identify IM particles in the surface and then this information was used to determine particle densities around the attack sites which were compared to average particle densities to identify the degree of clustering. Individual pits, developed after immersion of samples in 0.1 M NaCl at ambient temperature, were examined using the scanning vibrating electrode (SVE) technique [51]. SVE was used both to locate the individual corrosion pits and for direct quantitative measurement of the current associated with the individual pit. For the IM particle studies a range of IMCs were made according to particle compositions identified by Buchheit et al. [52] for AA2024-T3. These include Al3Fe, Al2CuMg, Al6(CuFeMn), Al2Cu and Al7Cu2Fe. A preliminary examination for a more detailed study of the composition of IM particles in AA2024-T3 has been published in this journal [53].

Section snippets

Sample preparation

For the SVE and PIXE experiments, samples were prepared by pressing discs out of AA2024-T3 sheet with a thickness of 1.6 mm. They were ground on silicon carbide papers to P1200 and then diamond polished to 0.25 μm. They were finally wiped with methyl ethyl ketone prior to installation in the cell of the SVE equipment. Corrosion of the sample in the cell was initiated by exposing it to aerated 0.1 M NaCl solution. In a second series of experiments, the polished discs of AA2024-T3 were immersed in

General observations

The polished AA2024-T3 samples were immersed in 0.1 M NaCl solution at ambient temperature (21 °C). Immersion experiments were performed in the SVE cell, where the samples were immersed in 25 ml of 0.1 M NaCl solution. The sample surfaces were 10 mm below the surface of the solution. After an induction period of a few minutes, corrosion in the form of darkening around IM particles was observed on the surface. This was followed by hydrogen evolution from a number of sites. White gelatinous corrosion

Discussion

In the first sections of this paper SVE and PIXE were used to characterize pits evolving H2. It is assumed that the pits were stable since it took over 90 min for the current to stabilize and the pits were evolving H2 which is only likely from an acidic environment such as that established in a stable pitting event, i.e., where a long diffusion path maintains a low pH. The PIXE results showed that there was a Cu-rich particle which was either an S-phase particle or it’s remnant adjacent to a

Conclusions

Stable pit initiation was studied on AA2024-T3. SVE measurements were made to determine pit currents and current densities for pits where H2 evolution was observed. PIXE was used to determine the elemental distribution around stable pits. These studies showed that the stable pits sites often had IM particles with opposite electrochemical activity (S-phase and AlCuFeMn IM particles). The electrochemical characteristics of IMC compositions similar to those in AA2024-T3 were measured. The anodic

Acknowledgements

Part of the work was completed under the Corrosion Prediction Modeling program between BAE SYSTEMS, CSIRO and DSTO (1999 – 2003). The authors would like to thank Mr. A. Stonham of BAE Systems, Salisbury, Australia and Drs. S. Harris, D. Dixon, P. Morgan, M. Hebbron and S. Church for useful discussions. BAE Systems are acknowledged for providing financial and technical support for part of this work through the Corrosion Prediction Modeling project. The remainder of this work was funded

References (75)

  • Z. Szklarska-Smialowska

    Corros. Sci.

    (1999)
  • E. McCafferty

    Corros. Sci.

    (2003)
  • K. Shimizu et al.

    Corros. Sci.

    (1997)
  • N. Sato

    Electrochim. Acta

    (1971)
  • N. Birbilis et al.

    Corros. Sci.

    (2006)
  • A. Conde et al.

    Corros. Sci.

    (1997)
  • P. Campestrini et al.

    Corros. Sci.

    (2000)
  • G.T. Burstein et al.

    Corros. Sci.

    (1993)
  • P.C. Pistorius et al.

    Corros. Sci.

    (1994)
  • M.A. Amin

    Electrochim. Acta

    (2009)
  • A.R. Trueman

    Corros. Sci.

    (2005)
  • D.E. Williams et al.

    Corros. Sci.

    (1994)
  • Y.L. Cheng et al.

    Corros. Sci.

    (2004)
  • D.N. Jamieson et al.

    Nucl. Instrum. Methods B

    (2002)
  • C.G. Ryan et al.

    Nucl. Instrum. Methods

    (1990)
  • C.G. Ryan et al.

    Nucl. Instrum. Methods

    (1993)
  • B. Rout et al.

    Nucl. Instrum. Methods B

    (2003)
  • Weilong Zhang et al.

    Electrochim. Acta

    (2003)
  • J.M.C. Mol et al.

    Corros. Sci.

    (2004)
  • F.H. Scholes et al.

    Corros. Sci.

    (2006)
  • S.A. Furman et al.

    Corros. Sci.

    (2006)
  • M. Fishkis et al.

    Wear

    (1997)
  • X. Zhou et al.

    Corros. Sci.

    (2003)
  • M. Janik-Czachor et al.

    Br. Corros. J.

    (1980)
  • G.S. Frankel

    J. Electrochem. Soc.

    (1998)
  • R.T. Foley

    Corros. Sci.

    (1995)
  • T.H. Muster et al.

    Copper distributions in Al-alloys

  • J.E. Hatch

    Aluminum: Properties and Physical Metallurgy

    (1984)
  • R.M. Ryders et al.

    J. Electrochem. Soc.

    (1994)
  • K. Nisancioglu

    J. Electrochem. Soc.

    (1990)
  • J.O. Park et al.

    J. Electrochem. Soc.

    (1999)
  • J.O. Park et al.
  • R.G. Buchheit et al.

    Corros. Sci.

    (1990)
  • A. Kolics et al.

    J. Electrochem. Soc.

    (2001)
  • V. Guillaumin et al.

    Corros. Sci.

    (1999)
  • N. Dimitrov et al.

    J. Electrochem. Soc.

    (1999)
  • G.S. Chen et al.

    Corrosion

    (1996)
  • Cited by (196)

    View all citing articles on Scopus
    1

    Tel.: +1 940 369 8127 (O), +1 940 765 9150 (Cell); fax: +1 940 565 2515.

    View full text