Continuous and discontinuous localized corrosion of a 2xxx aluminium–copper–lithium alloy in sodium chloride solution

Continuous and discontinuous localized corrosion of a 2xxx aluminium–copper–lithium alloy in sodium chloride solution

Accepted Manuscript Continuous and Discontinuous Localized Corrosion of a 2xxx Alumnium-copperlithium Alloy in Sodium Chloride Solution Chen Luo, Serg...

2MB Sizes 1 Downloads 47 Views

Accepted Manuscript Continuous and Discontinuous Localized Corrosion of a 2xxx Alumnium-copperlithium Alloy in Sodium Chloride Solution Chen Luo, Sergiu P. Albu, Xiaorong Zhou, Zhihua Sun, Xiaoyun Zhang, Zhihui Tang, George E. Thompson PII:

S0925-8388(15)31441-9

DOI:

10.1016/j.jallcom.2015.10.185

Reference:

JALCOM 35740

To appear in:

Journal of Alloys and Compounds

Received Date: 23 April 2015 Revised Date:

18 October 2015

Accepted Date: 20 October 2015

Please cite this article as: C. Luo, S.P. Albu, X. Zhou, Z. Sun, X. Zhang, Z. Tang, G.E. Thompson, Continuous and Discontinuous Localized Corrosion of a 2xxx Alumnium-copper-lithium Alloy in Sodium Chloride Solution, Journal of Alloys and Compounds (2015), doi: 10.1016/j.jallcom.2015.10.185. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Continuous and Discontinuous Localized Corrosion of a 2xxx Alumnium-copper-lithium Alloy in Sodium Chloride Solution

Chen Luoa,*, Sergiu P. Albub, Xiaorong Zhoub, Zhihua Suna, Xiaoyun Zhanga, Zhihui

a

RI PT

Tanga, George E. Thompsonb

Aviation Key Laboratory of Science and Technology on Advanced Corrosion and

Protection for Aviation Materials, AVIC Beijing Institute of Aeronautical Materials,

b

SC

Beijing 100095 (P.R. China)

School of Materials, The University of Manchester, Manchester M13 9PL (U.K.)

Abstract A

heterogeneous

M AN U

*Corresponding author. E-mail address: [email protected]

microstructure

aluminium-copper-lithium

alloy

is

intentionally

during

solidification

developed and

in

2A97-T3

thermomechanical

TE D

processes to obtain good mechanical properties. As a consequence, the alloy is susceptible to localized corrosion. Electron microscopy was employed to observe intermetallic particles and their periphery and the initiation and development of particle

induced

localized

corrosion

in

2A97-T3

EP

intermetallic

aluminium-copper-lithium alloy. In-situ optical microscopy monitoring, energy dispersive X-ray spectroscopy and electron backscatter diffraction were also used to

AC C

provide supportive evidence. Compared with the small number of continuous localized corrosion events, discontinuous localized corrosion event is relatively common. They are associated with corroded Al2Cu IM particles and Al-Cu-Fe-Mn-(Si) IM particles, as well as corrosion pits that are formed by particle fall-out due to dissolution of surrounding aluminium matrix. Discontinuous localized corrosion is confined within the shallow near-surface region of aluminium matrix. Triggered immediately after immersion, hydrogen gas evolution developed in form of bubbling at a continuous localized corrosion site which is associated with severe surface etching and sub-surface attack. Intergranular corrosion initiated from the corrosion pit 1

ACCEPTED MANUSCRIPT bottom, connects to the corrosion pit via small openings, and developed into the large network buried underneath the alloy surface. T1 phase precipitate remnant and corroded IM particles at grain boundary induced dissolution in the periphery of the particle, drive intergranular corrosion to propagate. Copper is much less oxidized than

RI PT

aluminium and lithium during grain boundary attack, and therefore accumulated at the corrosion product - aluminium matrix interface in the intergranular corrosion filament. Then, the copper enrichment band acts as further local cathode to support reduction

SC

reactions.

Keywords: localized corrosion; aluminium-copper-lithium alloy; intermetallic

M AN U

particle; intergranular corrosion 1 Introduction

2xxx aluminium alloy has been extensively used as structural material due to its high strength and damage-tolerance. However, addition of alloying elements such as

TE D

copper, magnesium, lithium, and silicon results in poor corrosion resistance of the alloy [1]. Continuous and discontinuous localized corrosion are classified in terms of time duration. Although less common, continuous localized corrosion has great impact on the application of engineering components by introducing significant

EP

reduction in their cross-sections. Microstructure that leads to such event is not clear. While with discontinuous localized corrosion we are looking for typical

AC C

microstructures, with continuous localized corrosion we need to look at exceptional microstructures.

Prior to the discussion of continuous localized corrosion, an important issue needs to be addressed is that corrosion is unstable during the initiation and early growth stage. Szklarska-smialowska [2, 3] reported that fresh corrosion pit repassivates because the pitting current (which represents the alloy dissolution rate) is too low to sustain the aggressive solution (and low pH) within the pit. In fact, the product of current density and pit diameter (ir) should be higher than 10-2 A/cm to maintain a stably growing 2

ACCEPTED MANUSCRIPT pitting event. Also, local high concentration of chloride and hydrogen ions may be swept away by stray convection currents in the solution since protection does not exist in a shallow pit cavity. Thus, new corrosion pits have been observed to become

RI PT

inactive after short time of growth [4].

Ilevbare et al. studied intermetallic (IM) particle induced pitting and trenching adjacent to the particles in AA2024 aluminium-copper-magnesium alloy [5]. Corrosion within most S-phase particles was localized to the particle itself.

SC

Calculation of the product of dissolution rate and pit dimension (ir) showed that pit growth in the Al-Cu-Mg particle occurs at values of much lower that required for the

M AN U

aluminium matrix. Even after extensive growth under open-circuit conditions, the corrosion did not spread into the matrix. According to Boag et al. [6-12], high pitting current density can be obtained at a corrosion site where coupling of IM particles is present. Their in-situ observation of pitting events has shown that Al-Cu-Fe-Mn particles in conjunction with S-phase particles have a high correlation with onset of

TE D

pitting corrosion. EELS elemental mapping further showed that near the Cu-rich regions are where there are high levels of oxide. The location of these two areas suggests that Cu is a local cathode in the corrosion pit and oxide is where there is

EP

anodic dissolution of aluminium. The key role of IM particle cluster in triggering the formation of deep corrosion pits is also confirmed by Wei et al. [13-18]. A series of research on the microstructure of AA2024 and AA7074 aluminium alloys and

AC C

relevant corrosion phenomena revealed two types of corrosion events. General pitting corrosion occurs at isolated surface particles and severe pitting corrosion develops at subsurface particle clusters. Particles in these clusters can gain access to the electrolyte through aluminium matrix dissolution around the surface particles in the same cluster. Then corrosion in aluminium matrix propagates from the bottom of an IM particle cluster to another cluster buried in a deeper position.

In this research, electron microscopy was employed to observe IM particles and their periphery and monitor the initiation and development of IM particle induced localized 3

ACCEPTED MANUSCRIPT corrosion in recently developed 2A97-T3 aluminium-copper-lithium alloy. In-situ optical microscopy, energy dispersive X-ray spectroscopy and electron backscatter diffraction were also used to provide supportive evidence.

RI PT

2 Materials and methods 1.2 mm thick, cold-rolled sheet 2A97-T3 aluminium alloy (Li 0.8-2.3 wt.%, Fe 0.15 wt.%, Si 0.15 wt.%, Cu 2.0-3.2 wt.%, Mn 0.20-0.6 wt.%, Be 0.001-0.10 wt.%, Zn 0.17-1.0 wt.%, Mg 0.25-0.50 wt.%, Ti 0.001-0.10 wt.%, Zr 0.08-0.20 wt.%, other

SC

elements total ≤ 0.15 wt.%, Al rem.) was employed. Specimens, of dimensions of 20 × 20 × 1.2 mm, were mechanically ground with 800, 1200 and 4000 grit silicon

M AN U

carbide paper and polished sequentially using 6, 3 and 1 µm diamond paste with water-free polishing liquid as lubricant, and then cleaned ultrasonically in an acetone bath and dried in a cool air stream. Electron probe microanalysis (EPMA) was performed on JEOL JXA-8100 at an accelerating voltage of 20 kV, with a 2×10-8 A beam current. Grain boundary precipitates was characterized using a Tecnai G2

TE D

F30S-TWIN transmission electron microscopy (TEM) operating at 300 kV. Electron transparent specimen was generated by twin-jet electropolishing in a solution of 80

EP

vol.% methanol + 20 vol.% nitric acid.

Potentiodynamic polarization experiments were performed on 2A97-T3, -T6, 2060-T8, AA2099-T83 and AA2024-T3 specimens in naturally aerated 3.5 wt.% sodium

AC C

chloride solution (analytical grade) to test the susceptibilities of the 2xxx aluminium alloys. Immersion testing of the 2A97 alloy was carried out in the same solution. Specimens before and after immersion testing were examined using a FEI QUANTA 600 scanning electron microscope (SEM) equipped with energy dispersive X-ray (EDX) facilities. For secondary electron (SE) and backscattered electron (BSE) imaging the incident electron beam was kept to 15 kV. Philips XL30 FEG-SEM was employed for electron backscatter diffraction (EBSD) mapping. EBSD data was recorded using a CHANNEL EBSD system produced by HKL Technology and was processed using in-house software VMap to determine the orientation of grains. 4

ACCEPTED MANUSCRIPT

To generate ultramicrotomed specimens, the alloy sheet was cut to dimensions of 20 mm × 7 mm × 1.2 mm and subsequently trimmed to a sharp tip with a glass knife exposing a cutting area of about 200 × 50 µm2. The block face remaining after

RI PT

sectioning was examined by SEM. Also, after a series of cutting at a step size of 15 nm, thin sections of the alloy were spread in water, collected onto 400 mesh nickel grids, and then examined using TEM.

SC

3 Results and discussion 3.1 Microstructural characterization

M AN U

3.1.1 Intermetallic particle distribution

Figure 1 a shows the scanning electron micrograph of the alloy sheet surface, revealing a uniform distribution of coarse IM particles with size ranging from 0.23 µm to 19.8 µm. Chemical compositions of nearly 120 IM particles from Figure 1 a were determined using EPMA (Table 1). Two IM phases were identified. They are

TE D

θ-phase that is rich in aluminium and copper only, and α-phase that is rich in aluminium, copper, iron, and manganese with or without small amount of silicon. Typical EDX spectra of θ-phase and α-phase are displayed in Figure 1 g and h. 8.4%

EP

of the IM particles is θ-phase. The other 91.6% of the IM particles is generally considered as α-phase (Al-Cu-Mn-Fe-(Si)), though variation in the amount of manganese and iron elements makes it difficult to identify their specific phase

AC C

composition. A cluster of Al-Cu-Fe-Mn-(Si) particles is shown in the scanning electron micrograph of Figure 1 b, with its largest dimension of 17.3 µm from bottom to top part of the cluster. Increased copper, iron and manganese yield were verified by EDX mapping of the IM particles (Figure 1 c - f). Smaller dispersoid particles, ranging from 100 nm to 500 nm, are also evident in Figure 1 b. These submicron particles formed during homogenization of aluminium alloy ingots by solid state precipitation. The intermetallic compounds contain elements of modest solubility and slow diffusion rate in solid aluminium. Once formed, these particles resist dissolution and coarsening. They serve to retard recrystallization and grain growth during 5

ACCEPTED MANUSCRIPT processing and heat treatment of the alloys. Normally, these dispersoids exhibit a relatively inert nature during corrosion testing in sodium chloride solution [31].

3.1.2 Grain boundary precipitates

RI PT

Figure 2 a shows scanning electron micrograph of triple points of grain boundaries. Needle-like particles of around 50 nm in length, 10 nm in width are well separated from each other at the grain boundaries. Interestingly, no particle is found at grain boundary AB, indicating that the distribution of such needle-like grain boundary

SC

precipitates is not uniform. Bright field TEM image of a precipitate is displayed in Figure 2 b at increased magnification. The particle is 175 nm in length with a sharp

M AN U

tip on two ends. EDX analysis was carried out. The precipitate is rich in copper and depleted in Al (Figure 2 c). According to many studies on the effect of heat treatment on the microstructure of aluminium-copper-lithium alloy [19, 20], and considering the incapability of EDX in lithium detection, it is highly likely that a T1 phase precipitate

TE D

is detected.

3.2 Electrochemical polarization

The polarization curve for 2A97-T3 alloy in 3.5 wt.% sodium chloride electrolyte is

EP

presented in Figure 3, along with that for 2A97-T6, 2060-T8, AA2099-T83 and AA2024-T3 alloys in the same electrolyte overlaid on the same plot for comparison. Similar electrochemical response is seen for these alloys but over a markedly wide

AC C

range of potentials.

All the polarisation curves exhibit a cathodic current plateau corresponding to the oxygen reduction reaction (ORR). Results show that the ORR plateau is higher for the 2A97-T6 and AA2099-T83 alloys than for 2A97-T3, 2060-T8 and AA2024-T3 alloys which could be related to enhanced kinetics of the ORR on the first two in comparison to the kinetics on the other three. Corrosion mechanisms was proposed on aluminium alloy [28,29,30]: the ORR takes place preferentially on the copper-rich IM particles such as S-phase (Al2CuMg), θ-phase (Al2Cu), α-phase (Al-Cu-Fe-Mn-(Si)), 6

ACCEPTED MANUSCRIPT T1 phase (Al2CuLi) particles. Comparison of the cathodic current density for 2A97 alloy in T3 temper to that for T6 temper confirms this since the higher population of copper containing precipitates in the alloy the higher the cathodic current density. Concerning the anodic domain, no passive region plateau is present for any of the

10-3 A/cm2 above the open circuit potential (OCP).

RI PT

alloys. The current density value for the alloys increased rapidly from 10-6 A/cm2 to

The OCPs of the aluminium-copper-lithium alloys displayed marked difference in the

SC

diagram, with values around -0.62 V/SCE for the 2A97-T3 alloy and ranging from -0.64 to -0.68 V/SCE for the 2A97-T6, 2060-T8 and AA2099-T83 alloys. Further, the

M AN U

OCP of the 2A97-T3 alloy was more negative than that for AA202-T3 alloy. Thus, the newly discovered alloy is good in corrosion resistance comparing to other 2000 serials aluminium-copper-lithium alloys, but more susceptible to corrosion than AA2024-T3 aluminium alloy.

TE D

3.3 Immersion testing in sodium chloride solution 3.3.1 General view of localized corrosion sites

During corrosion testing, faster generation of Al3+ (1) at local anodes resulted in a

EP

lower pH in corrosion front environment (2) - (4), promoting hydrogen gas evolution (5) - (6). Thus, hydrogen gas evolution indicated local sites where electrochemical reactions were taking place and can be used to position continuous localized

AC C

corrosion.

Al = Al3+ + 3e-

(1)

Al3+ + H2O = Al(OH)2+ + H+

(2)

Al3+ + 2H2O = Al(OH)2+ + 2H+

(3)

Al3+ + 3H2O = Al(OH)3 + 3H+

(4)

6H+ + 2Al = 2Al3+ + 3H2↑

(5)

2H+ + 2e- = H2↑(at corrosion front)

(6)

7

ACCEPTED MANUSCRIPT Figure 4 a - d show the optical micrographs of a 12 × 10 mm2 alloy surface in 3.5% sodium chloride solution. Triggered immediately after immersion, hydrogen gas evolution developed in form of continuous bubbling at a number of local sites. Gas bubbles grew to as big as 300 µm in diameter before they detached from the alloy

RI PT

surface. Then, a new bubble formed at the same position, grew large and detached the alloy surface again.

Localized corrosion was confirmed by ex-situ SEM examination. The corrosion sites

SC

where gas evolution persisted during the course of testing represent continuous localized corrosion. The corrosion sites where gas evolution was not evident are

M AN U

considered as discontinuous localized corrosion. Scanning electron micrographs of 2A97-T3 alloy after immersion for 45 minutes and 2 hours are displayed in Figures 5 a, 6 a and 7 a, respectively. It is clearly revealed that, compared with the large number of discontinuous localized corrosion events, continuous localized corrosion event is

TE D

relatively rare, with only one or two present in the examined regions.

3.3.2 Discontinuous localized corrosion sites An area including continuous and discontinuous localized corrosion sites is illustrated

EP

in Figure 5 a. The central and surrounding area of the continuous localized corrosion site is covered by corrosion product. Discontinuous localized corrosion sites are associated with corroded IM particles and corrosion pits. Interestingly, no corrosion

AC C

product is found at the sites. The matrix at the vicinity of IM particles was also attacked during testing, forming trenches.

Cross-sectional examination was carried out at a corrosion pit (line A), as displayed in Figure 5 b. It is located very close to the continuous localized corrosion site but exhibited no evident gas bubbling during immersion. A cavity in hemispherical shape exhibits wide opening to the outside environment. IM particle is absent in the corrosion pit, indicating the pit is formed by particle fall-out due to dissolution of surrounding aluminium matrix. Unlike the continuous localized corrosion site in 8

ACCEPTED MANUSCRIPT Figure 5 c, no penetration into the alloy regions deep beneath the surface occurred.

A galvanic coupling is established between α-phase IM particles and the aluminium matrix when the alloy is exposed to sodium chloride solution. Since copper, iron and

RI PT

maganese have a more positive electrode potential than aluminium, Al-Cu-Fe-Mn-(Si) particle is expected to serve as local cathode in a micro electrochemical cell [32-35]. The aluminium matrix in the surrounding region is anodic in electrochemical nature due to the lack of noble elements. As a consequence, due to the limited area of local

SC

cathode to support oxygen / H+ reduction, anodic dissolution of the surrounding

M AN U

material proceeds slowly.

2H+ + 2e- = H2↑(at cathodic IM particles)

(6)

O2 + 2H2O + 4e- = 4OH-

(7)

The reduction reaction (6) and / or (7) caused local alkalinisation around cathodic

TE D

particles. Aluminium oxide is not stable in alkaline environment (pH > 9) so that corrosion product could not form on the Al-Cu-Fe-Mn-(Si) particles [21]. Not covered by corrosion product, these particles drive aluminium matrix in the surrounding

EP

region further anodically dissolved in the solution. Aluminium element in the IM particles also dissolved selectively. Thereby the particle surface is enriched with Cu, Fe and Mn, leading to increased cathodic activity [22]. Etching of the aluminium

AC C

matrix around the particles detached the particles from the alloy surface (Figure 5 b), which results in reduction in the driving force for the acidic reaction.

Oxidation of aluminium in the periphery of IM particles produces Al3+ (1). Rapid hydrolysis of Al3+ produces H+ ions (2) - (4), resulting in acidification of the corrosion front environment in the trenches. Chloride ions migrate into the corrosion front region to balance the positive charge produced by reaction (1) - (4). The corrosive condition (H+ and Cl-) must be maintained to avoid repassivation of aluminium at the discontinuous localized corrosion site. 9

ACCEPTED MANUSCRIPT

Then, the phenomenon is similar to the concept of stable and metastable pitting in pure aluminium [23]. The localized corrosion in pure aluminium proceeds to a stage of metastable growth after initiation [2]. With a small number of metastable corrosion

RI PT

sites continuing to the next phase of stable growth, most of them repassivate at this stage. According to Szklarska-Smialowska [2, 3], several factors cause dying out of metastable corrosion events, e.g. when a pit is small and shallow, corrosive anolyte at corrosion front is easily swept out by the convection of bulk solution. Then, further

SC

dissolving of aluminium ceases and the corrosion pits repassivate.

M AN U

Thus, as shown in Figures 5 b, localized attack is restricted in the shallow near-surface region of aluminium matrix at Al-Cu-Fe-Mn-(Si) particles during immersion of the alloy in sodium chloride solution, with the maximum depth of attack around 2.8 µm. Al-Cu-Fe-Mn-(Si) particles could not solely promote continuous

TE D

localized corrosion.

3.3.3 Continuous localized corrosion sites Figure 6 a and b show the scanning electron micrographs of 2A97-T3 alloy after

EP

immersion in 3.5% sodium chloride solution for 2 hours, revealing another continuous localized corrosion site. Cross section of the continuous localized corrosion site was examined in order to get insight into the propagation mechanism, as displayed in

AC C

Figure 6 c. The corrosion site is associated with severe surface etching and sub-surface attack. An irregular-shape pit is evident with depth of 1.8 µm and width of 10.4 µm. Remnant of IM particle is also present in the surrounding region of the corrosion pit, appearing as a number of smaller particles.

It is also evident that intergranular corrosion connects to the corrosion pit via small openings, suggesting that intergranular corrosion initiated from the corrosion pit bottom, and developed into the large network buried underneath the alloy surface. SEM images of further cross-section (line C) taken from the same site are shown in 10

ACCEPTED MANUSCRIPT Figures 6 d. It is also clearly revealed that the intergranular corrosion in Figure 6 d is not directly connected to the outside environment. However, the intergranular corrosion that intersects the alloy surface and the intergranular corrosion that is buried beneath the alloy with no direct connection to the outside environment belong to the

RI PT

same intergranular corrosion network. A part of the network, which served as the path of corrosion propagation, is missed from the current cross-sectional examination. These intergranular corrosion filaments serve as connection between testing solution

SC

and the beneath-surface corrosion front.

It is clear that the geometry of continuous localized corrosion site is different from

M AN U

that at a discontinuous localized corrosion site, as shown in Figures 5 - 6. The latter has a relatively shallow reaction volume that is openly connected to the testing solution, resulting in ready access of bulk solution to the local region and pH increase. The former has a relatively large reaction volume beneath the alloy surface with a small area of connection to the testing solution. Such structure restricts diffusion

TE D

between the local region and the bulk solution, therefore, maintaining the necessary acidity of the solution within the corrosion front region for continuous corrosion propagation. When the active corrosion event develops larger than certain depth,

EP

corrosive environment at corrosion front is not interrupted by solution convection. In

AC C

such cases, the corrosion event survives and propagates as a stable one.

3.3.4 Intergranular corrosion Plan-view and cross-sectional SEM examination (line D) of a further continuous localized corrosion site is displayed in Figure 7 a and b. Similar to the one shown in Figure 6 a, a corrosion pit was found in the centre of the corrosion site (Figure 7 a). Attacked grain boundaries are evident from both cross-section and alloy surface. Beneath the alloy surface, the matrix at the vicinity of IM particles was also attacked. These particles are detached from aluminium matrix, with the space filled with corrosion product. EDX analysis was carried out on the remnant of IM particle, as shown in Figure 7 d. The spectrum exhibits increased yield copper, indicating the 11

ACCEPTED MANUSCRIPT presence of θ-phase. Attacked grain boundary established the connection between the separated corroded IM particles.

Normally, grain boundary precipitates are the preferential driving force for

RI PT

intergranular corrosion to propagate [24-27]. Consistent with this, it is revealed in Section 3.1.2 that T1 phase precipitate was present at the grain boundary. Actually, Li et al. [36, 37] evaluated the electrochemical behaviour of T1, T2 and θ′ phases with respect to the matrix and their studies showed that T1 precipitates have a more

SC

negative corrosion potential than the matrix in 4% sodium chloride solution. Therefore, these precipitates were associated with dissolution phenomena at grain

M AN U

boundaries. Although corrosion mechanism of T1 phase precipitate is still not clear, magnesium dealloying and copper liberation in similar S-phase have been reported in the previous studies. The work done by Buchheit et al. [38, 39] show that S-phase particles under 50 A˚ show dissolution behavior that is essentially identical to that of particles orders of magnitude larger in size. After potentiostatic exposure, one particle

TE D

is dealloyed leaving a porous remnant that is essentially pure copper at the original location of the particle. Supposedly, T1 phase actively undergo severe lithium dealloying and then switched from an anode to a cathode after the particle is enriched

EP

with copper. The precipitate remnant and corroded IM particles induced dissolution in the periphery of the particle. However, after certain time exposure, the remnant particle became electrically isolated from the matrix via the development of a

AC C

continuous oxide over the whole particle - aluminium matrix interface. The isolated remnant particle is susceptible to dissolution and broke up into smaller pieces.

Moreover, Figure 8 a shows the SE image of further intergranular corrosion site. Corrosion product was removed from the alloy surface by gentle mechanical polishing and subsequently argon plasma sputtering in order to reveal the surface morphology underneath corrosion product. Corrosion pits and attacked grain boundaries are found, as indicated by the arrows. The framed area is examined at increased magnification, as shown in Figure 8 b, indicating grain / subgrain boundary 12

ACCEPTED MANUSCRIPT attack as well as crystallographic pitting (grain body dissolution). Grain and subgrain orientation maps of the attacked region are displayed in random colour in Figures 8 c and 8 d, respectively. Grain / subgrain boundaries are presented by black lines in the map. Interestingly, a few attacked subgrain boundaries are not displayed in the

RI PT

subgrain orientation map, indicating that T1 phase grain boundary precipitates did not solely drive intergranular corrosion. Thus, further microstructural aspect that leads to the establishment of intergranular corrosion other than phase composition should be

SC

discussed.

Bright field TEM images of Figures 9 a and b reveal the inner part of an intergranular

M AN U

corrosion filament. The interface between intact aluminium matrix and corrosion product is decorated with a dark band, indicating re-distribution of alloying element(s). With limited amount of precipitates found in the alloy matrix, grain body is supposed to be rich in lithium and copper. Lithium, together with aluminium, was dissolved from grain matrix during grain boundary attack. As the other major alloying element,

TE D

copper is of relatively high Gibb’s free energy with respect to aluminium and lithium, and therefore much less oxidized than aluminium and lithium, and accumulated at the corrosion product - aluminium matrix interface. Interestingly, the arrangement of

EP

copper band is not smooth, indicating an uneven dissolution rate of aluminium matrix in different positions at the corrosion front when grain boundary attack propagated. The copper enrichment band acts as local cathode to support oxygen / H+ reductions

AC C

and balance anodic dissolution of aluminium.

4 Conclusions

Compared with the limited number of continuous localized corrosion events, discontinuous localized corrosion event is relatively common. They are associated with corroded Al2Cu particles and Al-Cu-Fe-Mn-(Si) particles, as well as corrosion pits that are formed by particle fall-out due to dissolution of surrounding aluminium matrix. Discontinuous localized corrosion is confined within the shallow near-surface region of aluminium matrix. 13

ACCEPTED MANUSCRIPT

Triggered immediately after immersion, hydrogen gas evolution developed in form of bubbling at a continuous localized corrosion site which is associated with severe surface etching and sub-surface attack. Intergranular corrosion initiated from the

RI PT

corrosion pit bottom, connects to the corrosion pit via small openings, and developed into the large network buried underneath the alloy surface. T1 phase precipitate remnant and corroded IM particles at grain boundary induced dissolution in the periphery of the particle, drive intergranular corrosion to propagate. Copper is much

SC

less oxidized than aluminium and lithium during grain boundary attack, and therefore accumulated at the corrosion product - aluminium matrix interface in the intergranular

support reduction reactions.

Acknowledgement

M AN U

corrosion filament. Then, the copper enrichment band acts as further local cathode to

The authors wish to thank the National Natural Science Foundation of China Program

TE D

Grant (No. 51201157) and National Defense Technology Foundation Project (H052013A003) for provision of financial support for the work.

EP

References

[1] A. E. Hughes, B. Hinton, S. A. Furman, I. S. Cole, D. Paterson, A. Stonham, G. McAdam, D. Dixon, S. J. Harris, A. Trueman, M. Hebbron, C. Bowden, P.

AC C

Morgan and M. Ranson, Corrosion Reviews 25 (2007) 275.

[2] M. G. Fontana and N. D. Greene, Corrosion Engineering, third ed., McGraw-Hill, London, 1967.

[3] Z. Szklarska-Smialowska, Corrosion Science 41 (1999) 1743. [4] Z. Szklarska-Smialowska, Corrosion Science 33 (1992) 1193. [5] G. O. Ilevbare, O. Schneider, R. G. Kelly and J. R. Scully, Journal of The Electrochemical Society 151 (2004) B453. [6] A. Boag, A. E. Hughes, A. M. Glenn, T. H. Muster and D. McCulloch, Corrosion Science 53 (2011) 17. 14

ACCEPTED MANUSCRIPT [7] A. Boag, R. J. Taylor, T. H. Muster, N. Goodman, D. McCulloch, C. Ryan, B. Rout, D. Jamieson and A. E. Hughes, Corrosion Science 52 (2010) 90. [8] A. M. Glenn, T. H. Muster, C. Luo, X. Zhou, G. E. Thompson, A. Boag and A. E. Hughes, Corrosion Science 53 (2011) 40.

RI PT

[9] A. Hughes, T. H. Muster, A. Boag, A. M. Glenn, C. Luo, X. Zhou, G. E. Thompson and D. McCulloch, Corrosion Science 52 (2010) 665.

[10] A. E. Hughes, A. Boag, A. M. Glenn, D. McCulloch, T. H. Muster, C. Ryan, C. Luo, X. Zhou and G. E. Thompson, Corrosion Science 53 (2011) 27.

SC

[11] A. E. Hughes, J. M. C. Mol, B. R. W. Hinton and S. van der Zwaag, Corrosion Science 47 (2005) 107.

Science 46 (2004) 1201.

M AN U

[12] J. M. C. Mol, A. E. Hughes, B. R. W. Hinton and S. van der Zwaag, Corrosion

[13] G. S. Chen, M. Gao and R. P. Wei, Corrosion (Houston) 52 (1996) 8. [14] M. Gao, C. R. Feng and R. P. Wei, Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science 29 (1998) 1145.

TE D

[15] C. M. Liao, J. M. Olive, M. Gao and R. P. Wei, Corrosion 54 (1998) 451. [16] C. M. Liao and R. P. Wei, Electrochimica Acta 45 (1999) 881. [17] R. P. Wei, Scripta Materialia 44 (2001) 2647.

EP

[18] R. P. Wei, C. M. Liao and M. Gao, Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science 29 (1998) 1153. [19] Z. Yuan, Z. Lu, Y. Xie, S. Dai, C. Liu, Chinese Journal of Nonferrous Metals 16

AC C

(2006) 2027.

[20] Y. Lin, Z. Zheng, Y. Han, H. Zhang, Chinese Journal of Nonferrous Metals 22 (2012) 2181.

[21] D. D. MacDonald and P. Butler, Corrosion Science 13 (1973) 259. [22] G. Svenningsen, Corrosion of Aluminium Alloys, SINTEF, Trondheim, 1993. [23] B. W. Davis, P. J. Moran and P. M. Natishan, Corrosion Science 42 (2000) 2187. [24] W. L. Zhang, G. S. Frankel, Electrochimica Acta 48 (2003) 1193. [25] V. Guillaumin, G. Mankowski, Corrosion Science 41 (1999) 421. [26] J. R. Galvele, S. M. de De Micheli, Corrosion Science 10 (1970) 795. 15

ACCEPTED MANUSCRIPT [27] N. Birbilis, M. K. Cavanaugh, L. Kovarik, R. G. Buchheit, Electrochemistry Communications 10 (2008) 32. [28] R. G. Buchheit, R. P. Grant, P. F. Hlava, B. McKenzie, G. L. Zender, Journal of The Electrochemical Society 144 (1997) 2621-2628.

The Electrochemical Society 146 (1999) 4424-4428.

RI PT

[29] R. G. Buchheit, L. P. Montes, M. A. Martinez, J. Michael, P. F. Hlava, Journal of

[30] L. Lacroix, L. Ressier, C. Blanc, G. Mankowski, Journal of The Electrochemical Society 155 (2008) C131-C137.

SC

[31] V. Guillaumin and G. Mankowski, Corrosion Science 41 (1998) 421.

[32] P. Schmutz and G. S. Frankel, Journal of The Electrochemical Society 145 (1998)

M AN U

2285.

[33] P. Leblanc and G. S. Frankel, Journal of The Electrochemical Society 149 (2002) B239.

[34] N. Birbilis and R. G. Buchheit, Journal of The Electrochemical Society 155 (2008) C117.

TE D

[35] O. Schneider, G. O. Ilevbare, J. R. Scully and R. G. Kelly, Journal of The Electrochemical Society 151 (2004) B465. [36] J. F. Li, C. X. Li, Z. W. Peng, W. J. Chen, Z. Q. Zheng, Journal of Alloys and

EP

Compound 460 (2008) 688.

[37] J. F. Li, Z. Q. Zheng, N. Jiang, S.C. Li, Materials and Corrosion 56 (2005) 192. [38] N. Birbilis, M. K. Cavanaugh, L. Kovarik, R.G. Buchheit, Electrochemistry

AC C

Communications 10 (2008) 32-37.

[39] R. G. Buchheit, M. A. Martinez and L. P. Montes, Journal of The Electrochemical Society 147 (2000) 119-124.

16

ACCEPTED MANUSCRIPT

Table 1 EPMA results of IM particles on the 2A97-T3 aluminium alloy surface Average chemical composition / wt.% Phase

Chemical

Particle

formula

percentage

Atomic ratio Al

Cu

Fe

Mn

Si

Total

43.782

49.628

(0.033)

(0.083)

(0.078)

93.604

θ-phase: Al/Cu = 2.1:1.0

RI PT

Al-Cu

8.4%

Al2Cu

Al/Cu/Fe/Mn =

α-phase:

13.6:2.8:1.7:1.0

Al6(Fe,Mn),

Al/Cu/Fe/Mn/Si

Al3(Mn,Fe),

Al-Cu50.444

24.227

12.911

7.528

(0.036)

95.146

Al-CuFe-Mn

54.235

24.311

11.577

5.795

=

Al7Cu2Fe,

20.1:3.8:2.1:1.1:

Al12Si(Mn,

98.712

M AN U

-Si

2.794

SC

Fe-Mn

1.0

Table 2 EDX analysis of a grain boundary precipitate wt.%

at.%

Al

77.17

88.84

Cu

22.83

11.16

AC C

EP

TE D

Element

17

89.9%

1.7% Fe)

ACCEPTED MANUSCRIPT List of Figures

Figure 1 (a) Scanning electron micrograph of mechanically polished 2A97-T3 aluminium alloy surface, (b) Al-Cu-Mn-Fe-(Si) particles, (c) - (f) EDX elemental

RI PT

maps, and (g) - (h) EDX spectra of typical IM phases. Figure 2 (a) SEM image of grain boundaries, (b) transmission electron micrograph of grain boundary precipitate, and (c) EDX spectra of the precipitate.

Figure 3 Voltage – current density curves of potentiodynamic polarization of 2A97-T3,

SC

-T6, AA2024-T3, 2060-T8 and AA2099-T83 aluminium alloys in 3.5 wt.% sodium chloride solution.

M AN U

Figure 4 Video frames of 2A97-T3 aluminium alloy immersed in 3.5% sodium chloride solution after (a) 5 minutes, (b) 10 minutes, (c) 20 minutes and (d) 30 minutes.

Figure 5 (a) Scanning electron micrograph of 2A97-T3 aluminium alloy after immersion in sodium chloride solution for 45 minutes, showing sites of continuous

magnification.

TE D

and discontinuous localized corrosion, (b) cross-section, (c) plan-view at increased

Figure 6 SEM images of a continuous localized corrosion site after 2 hours immersion

EP

in sodium chloride solution, (a) plan-view, SE, (b) plan-view, BSE, (c) and (d) cross sections (the sectioning direction and position are indicated by line B and C). Figure 7 Scanning electron micrograph of a continuous localized corrosion site after 2

AC C

hours immersion, (a) plan-view, (b) cross section, (c) framed area at increased magnification, and (d) EDX spectrum of grain boundary IM particles. Figure 8 SEM images of a continuous localized corrosion site, (a) SE image; (b) framed area at increased magnification, and EBSD maps, (c) grain orientation map in Euler’s colour, and (d) sbugrain orientation map in random colour. Figure 9 Transmission electron micrographs of an intergranular corrosion site in 2A97-T3 aluminium alloy after immersion in 3.5% sodium chloride solution, and (b) framed area at increased magnification, indicating re-distribution of copper.

18

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 1 (a) Scanning electron micrograph of mechanically polished 2A97-T3 aluminium alloy surface, (b) Al-Cu-Mn-Fe-(Si) particles, (c) - (f) EDX elemental maps, and (g) - (h) EDX spectra of typical IM phases. 1

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 2 (a) SEM image of grain boundaries, (b) transmission electron micrograph of grain boundary precipitate, and (c) EDX spectra of the precipitate.

2

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 3 Voltage – current density curves of potentiodynamic polarization of 2A97-T3, -T6, AA2024-T3, 2060-T8 and AA2099-T83 alloys in 3.5 wt.% sodium chloride

AC C

EP

TE D

solution

3

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 4 Video frames of 2A97-T3 aluminium alloy immersed in 3.5% sodium

AC C

chloride solution after (a) 5 minutes, (b) 10 minutes, (c) 20 minutes and (d) 30 minutes.

4

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 5 (a) Scanning electron micrograph of 2A97-T3 aluminium alloy after immersion in sodium chloride solution for 45 minutes, showing sites of continuous and discontinuous localized corrosion, (b) cross-section, (c) plan-view at increased magnification. 5

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

6

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 6 SEM images of a continuous localized corrosion site after 2 hours immersion in sodium chloride solution, (a) plan-view, SE, (b) plan-view, BSE, (c) and (d) cross sections (the sectioning direction and position are indicated by line B and C).

7

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

8

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 7 Scanning electron micrograph of a continuous localized corrosion site after 2 hours immersion, (a) plan-view, (b) cross section, (c) framed area at increased magnification, and (d) EDX spectrum of grain boundary IM particles.

9

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

10

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 8 SEM images of a continuous localized corrosion site, (a) SE image; (b)

framed area at increased magnification, and EBSD maps, (c) grain orientation map in Euler’s colour, and (d) sbugrain orientation map in random colour.

11

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 9 Transmission electron micrographs of an intergranular corrosion site in 2A97-T3 aluminium alloy after immersion in 3.5% sodium chloride solution, and (b) framed area at increased magnification, indicating re-distribution of copper.

12

ACCEPTED MANUSCRIPT Highlights

Observation of continuous and discontinuous localized corrosion in an Al-Cu-Li alloy

RI PT

OM monitoring followed by in-situ cross-sectional SEM examination of corrosion sites

SC

Discontinuous corrosion is confined within shallow near-surface region of the alloy

M AN U

Continuous corrosion is associated with trenching adjacent to IM particles and IGC

AC C

EP

TE D

Grain boundary IM phases and copper enrichment band act as cathode to support IGC