Corrosion Science 102 (2016) 413–424

Contents lists available at ScienceDirect

Corrosion Science
journal homepage: www.elsevier.com/locate/corsci

The effect of artificial ageing heat treatments on the

corrosion-induced hydrogen embrittlement of 2024 (Al–Cu)
aluminium alloy
Nikolaos D. Alexopoulos a,∗ , Zaneta Velonaki a , Constantinos I. Stergiou b ,
Stavros K. Kourkoulis c
a
University of the Aegean, Department of Financial Engineering, 821 00 Chios, Greece
b
Technological Educational Institute of Piraeus, Department of Mechanical Engineering, 250 Thivon Ave, 12244 Athens, Greece
c
National Technical University of Athens, Department of Mechanics, Laboratory of Testing and Materials, 9 Heroes Polytechniou Str., 15780 Athens, Greece

a r t i c l e i n f o a b s t r a c t

Article history: This work is focused on the role of microstructure (S-type particles) that are precipitated in the aluminium
Received 19 August 2015 matrix due to artificial ageing on the tensile ductility decrease of 2024 aluminium alloy when exposed
Received in revised form 21 October 2015 to exfoliation corrosion solution. A common, extremely short corrosion exposure time was selected to
Accepted 23 October 2015
corrode the already heat-treated specimens in different conditions, including under-, peak- and over-
Available online 28 October 2015
ageing so as to avoid pitting formation and to concentrate only to hydrogen embrittlement degradation.
Corrosion-induced yield stress decrease was less sensitive to the change of microstructure. Corrosion-
Keywords:
induced tensile ductility decrease was correlated with the precipitation of S-phase particles, showing
A. Aluminium
A. Alloys
that the specimens at the peak-ageing (S and S precipitates) and over-ageing (S precipitates) exhibited
B. SEM the lowest decrease (∼11% and ∼22%, respectively) due to corrosion exposure. The highest ductility
C. Exfoliation corrosion decrease (∼26%) was noticed for the T3 condition (as received) that limited GPB zones were formed on
C. Hardening the grain boundaries. Tensile ductility decrease was also ageing temperature sensitive, as the lowest
C. Pitting corrosion temperature gave the least ductility decrease, thus giving proof of the role of the precipitates on the
C. Hydrogen embrittlement corrosion mechanism. The results were explained based on the hydrogen trapping at the interphases of
the S-type precipitates.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction
Many decades now, the aluminium alloy 2024 is widely used
in the aeronautical industry due to its high specific mechani-
cal properties and damage tolerance capabilities. New versions
and derivatives of AA2024 with lower compositions of Fe and Si
have been investigated during the last decade to improve fatigue,
fracture toughness and resistance to corrosion. AA2024-T3 is a
high strength alloy with a complex microstructure, comprising the
aluminium matrix as well as a number of different intermetal-
lic particles. Several articles tried to characterise the resulting
microstructure of the alloy and the respective composition of
the intermetallic particles [1,2]. Almost 61.3% of the intermetal-
lic particles are S-phase (Al2 CuMg), 12.3% are Al6 (Cu,Fe,Mn), 5.2%
are Al7 Cu2 Fe and 4.3% are (Al,Cu)6 Mn particles in AA2024 [2].
Starke and Staley [3] suggested that, in addition to S-phase and -
phase (Al2 Cu) particles, the intermetallic particles included AlCuFe
and AlFeMnSi. In a recent work, Boag et al. [1] showed that the
microstructure of AA2024-T3 is very complicated, exhibiting mul-
tiphase particles, periphery phases around composite particles and
clustering. A shape-based classification of the particles was used
to distinguish between Al20 (Cu,Mn,Fe)5 Si type particles and the
rest intermetallic particles, however, this could not confidently be
used to distinguish between cathodic and anodic particles. Hence,
the authors suggested that a compositional classification would
be appropriate, since it better reflects the likely electrochemical
properties.
It is well known that AA2024 is strengthened by microstructure
evolution (precipitates) during ageing. In the 1950s, Bagaryatsky
[4] proposed the following precipitation sequence:

SSS → GPBzone → S → S  → S,

∗ Corresponding author. where SSS stands for supersaturated solid solution and GPB stands
E-mail address: nalexop@aegean.gr (N.D. Alexopoulos). for Guinier–Preston–Bagaryatsky; GPB is considered to be a short

http://dx.doi.org/10.1016/j.corsci.2015.10.034
0010-938X/© 2015 Elsevier Ltd. All rights reserved.
414 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424
range ordering of Cu and Mg solute atoms while S are very small
precipitates fully coherent with the Al matrix. The S phase is an
equilibrium phase and is incoherent with the Al matrix. These
S-phase particles are round particles and 60% of the particles
are larger than 0.5 ␮m [2]. In contrast to the S-phase particles,
the Al(Cu,Mn,Fe,Si) intermetallics present irregular shapes and
size typically larger than 5 ␮m. The S phase has generally been
considered as semi-coherent with the matrix, having the same
structure as the S-phase but with slightly different lattice param-
eters [4,5]. Ringer et al. [6,7] confirmed that GPB zones and other
precipitate-structures prior to the S-phase formation are the dom-
inant precipitates at the strengthening regime of the ageing curve,
while the S-phase appears in the softening regime (over-ageing
condition). In a most recent work, Wang and Starink [8] showed
that actually two types of the S precipitate are formed, depending
on the ageing temperature.
Recent studies for the corrosion of AA2024-T3 showed that the
distribution of the second phase particles has a significant influence
on the alloy corrosion behaviour. If the S-phase particles are prefer-
entially cited at the grain boundaries, then copper depleted regions
are developed adjacent to the boundaries. These are anodic to the
copper rich grain boundaries, and hence, a micro-galvanic corrosion
takes place that may cause corrosion of the particle, or corrosion
of the surrounding matrix [9–14]. Corrosion of the Al2 CuMg phase
with the surrounding matrix was also recently studied in the open
literature [15,16]. Campestrini et al. [17] showed that a potential
difference has been observed between the matrix and the inter-
metallic particles that result in the formation of a severe galvanic
coupling. This weakens the natural oxide film of the aluminium
alloy and explains the change of corrosion from general to localized,
induced by the intermetallic particles. The effect of the precipitate
size, the role of the nanostructure and the grain stored energy on the
pitting potential in Al–Cu–Mg alloys are well documented [18–21].
In addition to the micro-galvanic corrosion associated with the
grain boundary and its neighbouring regions, hydrogen embrit-
tlement has been reported to be a contributing factor to the
corrosion-induced degradation of mechanical properties in 2xxx
aluminium alloys. Hydrogen production is a cathodic reaction dur-
ing the corrosion process; recently in-situ X-ray tomography was
used to capture H2 production at the corrosion surface of AA2024
and the networks of intergranular corrosion extended into the
material [22,23]. Larignon et al. [24] proposed that the grain and
sub grain boundaries are preferential pathways for the diffusion
of hydrogen. Hydrogen embrittlement has been reported to have
an essential effect to the loss of tensile ductility of pre-corroded
AA2024 [25–28]. Charitidou et al. [29] acknowledged that the dif-
fused hydrogen is trapped in four sites in the microstructure of the
alloy. The trapping states T1 and T2, were associated with the inter-
stitial lattice sites [30] and with the semi-coherent interfaces of the
strengthening phases or with the incoherent interphases of disper-
soids and the matrix lattice, respectively. Scamans and Tuck [31]
attributed the trap T3 to the formation of Mg hydride. The trap-
ping site T4 began to release hydrogen at a temperature coinciding
with the dissolution of the strengthening phase, thus implying that
hydrogen is trapped inside/or at the interphase of this second phase
[29]. Scholes et al. [32] also observed that the corrosion in alloy
2024-T3 was initiated at S-phase particles. Saitoh et al. [33] proved
that the above microstructural features of the alloy can be used as
hydrogen traps by employing tritium autoradiography.
In order to investigate the hydrogen embrittlement effect, a
wide variety of methods to introduce hydrogen has been used
over the years, such as cathodic charging [34], exposure to humid
air as well as exposure to a corrosive environment. For the latter
case, it is quite often the degradation mechanism to be a synergy
of stress-concentration by the corrosion-induced surface pits and
subsequent micro-cracking as well as of hydrogen embrittlement.
For example, Alexopoulos and Papanikos [27] performed fracture
toughness mechanical tests on the same alloy and for various expo-
sure times to corrosive environment. They have shown that the
total reduction of almost 30% of the fracture toughness after 96 h
exposure to exfoliation corrosion solution was attributed primary
to the reduction of the alloy effective thickness (22%) and secondary
to the hydrogen embrittlement (8%) mechanism. Hence, in order
to assess the true hydrogen embrittlement effect, short exposure
times should be selected to corrode the specimens so as to prevent
the formation of corrosion-induced surface pits that act as stress
raisers and degrade the ductility of the alloy.
It is evident though, that the tensile mechanical properties of
the material due to the precipitation hardening are a function of
artificial, and hence, of natural ageing. An assessment of different
artificial ageing conditions to the corrosion-induced degradation on
mechanical properties should be attempted. To the authors knowl-
edge, quite a few researchers [34–38] tried to assess the effect
of different artificial ageing conditions to the corrosion potential
of AA2024. Zhang and Frankel [39] also studied the pitting and
intergranular corrosion behaviour of various tempers of AA2024.
Artificial ageing was found to have a strong effect on polarization
curves and localized corrosion morphology of AA2024; it resulted
in more precipitation of intermetallic particles from the matrix and
formation of Cu-depleted matrix phase, which greatly changed the
localised corrosion of AA2024. In the ASM Handbook [40], it is gen-
erally reported that the 2xxx wrought aluminium alloys in the T3
condition have low corrosion resistance when compared to the
“precipitated” T6 and T8 conditions of the alloy. In a most recent
article, Alexopoulos [28] demonstrated that for different artificial
ageing conditions and subsequent exfoliation corrosion exposure
for 24 h exposure time, the aged AA2024 specimens presented
higher corrosion resistance than the T3 condition.
In the present work, the effect of artificial ageing of AA2024-T3
on the tensile mechanical properties degradation due to corrosion
exposure will be investigated. Tensile specimens will be artificially
aged to tempers that correspond to Under-Ageing (UA), Peak-
Ageing (PA) and Over-Ageing (OA) conditions and then will be
subsequently exposed to exfoliation corrosion environment. The
corrosion exposure will be experimentally derived and should be
the least possible so as to avoid the formation of large surface pits,
trying to simulate the hydrogen embrittlement phenomenon. The
purpose of this work is to identify how the resulting microstruc-
ture from artificial ageing affects the pitting corrosion potential,
hydrogen embrittlement and corrosion-induced degradation of the
tensile mechanical properties.
2. Material and specimens
The materials used for the present investigation were AA2024-
T3 wrought aluminium alloy that was received in sheet form with
nominal thicknesses of 3.2 mm and without any surface corro-
sion protection (Alclad). The sheets had geometrical dimensions of
35 × 50 cm, while their thickness is typical for aerostructures. The
weight percentage chemical composition of the alloy is 0.50% Si,
0.50% Fe, 4.35% Cu, 0.64% Mn, 1.50% Mg, 0.10% Cr, 0.25% Zn, 0.15% Ti
and Al rem. Tensile specimens were machined from the longitudi-
nal (L) direction of the material according to ASTM E8 specification
with 12.5 mm × 3.2 mm being the reduced cross-section and 50 mm
being the gauge length of the specimens.
3. Experimental procedure
The experimental procedure flow diagram can be seen in
Fig. 1; tensile specimens were machined, surface-cleaned with
alcohol according to ASTM G1 and then were all artificially aged,
 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424 415

Experimental procedure
All specimens were artificially
Al 2024-T3 aged for different conditions
sheets (bare)
(t = 3.2 mm)
Half of them were
Masking of the
tensile tested
specimens
(reference specimens)
Machining of L
Half of them were

Preparation
tensile specimens
exposed to exfoliation
corrosion solution
Comparison
Tensile tests on the
Surface cleaning
artificially aged
specimens after being
corroded

Mechanical Fractographic
performance analysis

Degradation
mechanisms
Evaluation

Fig. 1. Experimental procedure flow chart.

Table 1
Investigated exfoliation corrosion exposure times and number of tensile specimens of the reference 2024-T3 alloy.

Different exposure times of AA2024-T3 to EXCO solution Total

0.5 h 1h 2h 4h 8h 12 h 24 h

Specimens # 3 3 3 4 4 3 3 23

Table 2
Artificial aging conditions and tensile specimens of the investigated artificial aging conditions.

Temperature Artificial aging of AA2024-T3 # of specimens

0h 0.5 h 1h 2h 4h 6h 9h 15 h 24 h 48 h 63 h 98 h
170 ◦ C 2 2 2 2 2 2 12
190 ◦ C 2 2 2 2 2 2 2 12
210 ◦ C 2 2 2 2 2 2 2 14
Sub-total 40

Artificial aging and subsequent 2 h EXCO corrosion exposure of AA2024-T3

170 ◦ C 3 3 3 3 3 15
190 ◦ C 3 3 3 3 3 3 3 18
210 ◦ C 3 3 3 3 3 3 3 21
Sub-total 57

Tables 1 and 2. Different artificial ageing conditions were per-
formed for the total 120 tensile specimens at three different
temperatures, namely 170, 190 and 210 ◦ C in an electric oven with
±0.1 ◦ C temperature control and for different times. The ageing
temperatures were selected in order to fill the whole tempera-
ture range of the artificial ageing regime of the commercial sheet
AA2024-T3 (Fig. 2) and closely to its upper bound according to the
literature [41]. This high ageing temperature enables to bring the
alloy to the (OA) condition at a very short period of time. Ageing
times were selected to correspond to all ageing conditions, includ-
ing Under-Ageing (UA), Peak-Ageing (PA) and Over-Ageing (OA).
For more details regarding artificial ageing and its effect on tensile
mechanical properties, one can refer to the respective forthcoming
publication [42].
Half of the available specimens were tested in tension according
to ASTM E8, immediately after the artificial ageing heat treatment
in order to assess the effect of artificial ageing on the respective ten-
sile mechanical properties. The other half tensile specimens were
exposed for 2 h to the laboratory exfoliation corrosion environ-
ment (hereafter called EXCO solution) according to specification
ASTM G34. The corrosive solution consisted of the following chem-
icals diluted in 1 l distilled water; sodium chloride (4.0 M NaCl),
potassium nitrate (0.5 M KNO3 ) and nitric acid (0.1 M HNO3 ). More
details can be seen in Refs. [27,43] and in the respective specifi-
cation. The common 2 h exposure time was selected according to
the findings of the present article since, it was concluded that for
longer exposure times surface corrosion products are being formed
(pits) that further degrade the tensile mechanical performance of
AA2024-T3. As the purpose of the present article was to assess the
effect of the hydrogen embrittlement itself, the common 2 h expo-
sure was kept constant for all heat treatments considered. After
the corrosion exposure, the corroded specimens were immediately
cleaned according to ASTM G34 and then subjected to mechanical
testing. Tensile tests were carried out in a servo-hydraulic Instron
100 kN testing machine according to the ASTM E8 specification.
416 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424

600
Solid solution heat treatment 1 hour reference - T3
(performed by the manufacturer) 500
essential
500

Nominal axial stress σ [MPa]

Rapid cooling (quench) stress drop
Heat treatment temperature [ C]

o
0.5 hour
in water with ice < 10 C 400
o

8 hours 2 hours
400 4 hours
Artificial ageing heat treatment
12 hours
300
(varying temperature and duration)
300 24 hours
200
200
T3 condition 100 Aluminum alloy 2024-Τ3
(as received) t = 3.2 mm, direction L
100 Exposure at EXCO solution
0
0 0.00 0.05 0.10 0.15 0.20
0 10 20 30 40 50 60 70 Nominal axial strain ε [-]
Heat treatment duration [hours]
Fig. 3. Typical experimental tensile flow curves of pre-corroded aluminium alloy
Fig. 2. Schematic representation of the temperature and duration of the artificial 2024-T3 (as-received) specimens for different exposure times to exfoliation corro-
ageing heat-treatment of age-hardenable aluminium alloys. sion solution.

Three specimens were tested in each different case to get reliable 20

experiments, t = 3.2 mm
Aluminum alloy 2024-Τ3
average data. t = 3.2 mm, direction L
Exposure at EXCO solution
16
4. Mechanical test results
Elongation at fracture Af [%]
Artificial ageing heat treatment affects all tensile mechanical 12 ΑΑ2024-Τ3, t = 3.2 mm
properties in age-hardenable aluminium alloys, e.g. [5,9,42,43]. Alexopoulos and Papanikos [27]

A small paragraph was added for the convenience of the reader.

8
Yield stress Rp0.2% and ultimate tensile strength Rm are increased ΑΑ 2024-Τ351, t = 1.6 mm
Petroyiannis et al. [25]
with increasing ageing time in the UA condition with a simultane-
ously elongation at fracture Af decrease. In the PA condition, both 4
strength properties have reached a plateaux of maximum strength
with increasing ageing time while ductility a respective minimum.
Finally, at the OA condition, all strength properties are continuously 0
0.01 0.1 1 10 100
decreasing with increasing ageing time while a small restoration of
the ductility is noticed. Exposure time at EXCO solution [hours]

Fig. 4. Evaluated elongation at fracture Af for different corrosion exposure times of

4.1. Assessment of the pitting formation threshold aluminium alloy 2024-T3 and comparison with the literature.

The first experimental study of the present article was to assess

the maximum exfoliation exposure time that degrade the AA2024-
T3 mechanical properties by means of hydrogen diffusion and
subsequent local embrittlement of the specimen. It is well-known
that for higher exposure times to corrosion solution, corrosion-
induced surface pits are formed that act as surface notches; they
have a profound effect on the ductility degradation of the speci-
men as they act as stress raisers. Different corrosion exposure times
were selected to corrode the reference material (without any age-
ing) and their effect on the typical tensile flow curves can be seen in
Fig. 3. It is noticed that even for short exposure times, e.g. half hour
exposure, ductility is decreased. Up to 2 h exposure no essential
stress decrease (vertical drop as pointed in the figure) of the tensile
flow curves is noticed while ductility is continuously decreasing
with increasing exposure corrosion time. For the case of 4 h expo-
sure this stress vertical drop is essential ranging from 40 and up
to 80 MPa. This seems to be the case for even higher exposure
times; after 24 h exposure almost a 15% stress decrease is noticed
when compared with the reference curve. This drop is attributed
to the corrosion-induced surface pitting formation and subsequent
decrease of the effective thickness of the specimen due to the pres-
ence of corrosion-induced surface pits [27]. The authors decided to
select the 2 h exposure as the last investigated exposure time that
does not have an essential vertical stress drop from the flow curve
of the uncorroded material and therefore, the noticed small pitting
formation do not essentially influence the effective thickness of the
material.
The effect of exposure time duration on the tensile mechani-
cal properties has been reported in the literature, e.g. [25–27]. In
the present work, the effect of corrosion exposure time is focused
on the ductility due to its profound impact due to the localised
hydrogen embrittlement phenomenon. Fig. 4 shows the evaluation
results of the present article (3.2 mm thickness) as well as from the
literature for lower thickness [25] and different manufacturer [27].
It can be seen that ductility continuously decreases with increasing
exposure time; an exponential ductility decrease is noticed while
an almost linear approximation can be seen in the semi-log diagram
of Fig. 4.
Typical cross-section metallography photos can be seen in Fig. 5;
actually no generated surface pits can be noticed in Fig. 5a after
only 0.5 h exposure to the corrosive environment. For the case of 1 h
exposure time (Fig. 5b) some pits were generated on both corroded
surfaces. Some surface pits seems to be formed after 2 h exposure
and the pits seems not to grow deeply into the core of the material
but parallel to the surface, Fig. 5c. After 4 h corrosion exposure, the
outer surface seems to be heavily corroded and almost a 40 ␮m cor-
roded layer is formed below the outer specimen surface. This has
an effect on ductility due to the related stress concentration in the
corroded region. In addition, the concept of the formed corrosion
 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424 417

Fig. 5. Cross-section micrographs of pre-corroded aluminium alloy 2024-T3 specimens after (a) 0.5 h, (b) 1 h, (c) 2 h ␬␣␫ (d) 4 h exposure at the exfoliation corrosion solution
(specimen thickness ∼3.2 mm).

layer that decreases the specimen thickness [27] can be embraced (a) 500
and this can be noticed as lower strength properties in the respec-
tive tensile stress–strain curves, e.g. strength (vertical) difference
between the 2 and the 4 h exposure flow curves. To this end, the 450
2 h exposure was selected to corrode the tensile specimens as no
Yield stress Rp0.2% [MPa]

heavily corrosion pits were formed and the tensile stress–strain

curves between the reference and 2 h corrosion exposure are very 400
close with the exception of lower ductility values. This high duc-
tility decrease cannot be explained with the concept of pits that
350
act as stress raisers and therefore, should be explained on the basis
of the hydrogen embrittlement effect; the concept of the present
paper focuses on the susceptibility of various artificial ageing heat Aluminum alloy 2024-Τ3
300 t = 3.2 mm, L direction
o
170 C
treatments to hydrogen embrittlement and therefore, the corrosion Artificial ageing and subsequent o
190 C
exposure of 2 h to EXCO solution is considered as the maximum 2 h exposure at EXCO solution o
210 C
exfoliation corrosion exposure time to achieve this goal. 0
0.01 0.1 1 10 100 1000

4.2. Effect on the mechanical properties Artificial ageing time [hours]

(b) 16
Yield stress and elongation at fracture for different artificial age-
14
ing conditions and subsequent 2 h EXCO exposure time can be seen
in Fig. 6; the present analysis will be focused on the corrosion
Elongation at fracture Af [%]

12
degradation comparison only. Fig. 6a summarizes the conventional
nominal yield stress Rp0.2% test results as average values and respec- 10
tive standard deviation derived from three specimens each. Test
results can be seen for the three investigated artificial ageing tem- 8
peratures. Respective elongation at fracture properties can be seen
6
in Fig. 6b; Af values seems to follow the typical drop on ductility
due to peak-ageing and a small increase at the over-ageing condi- 4
tion [42]. The test results will be compared against the respective Aluminum alloy 2024-Τ3 o
170 C
t = 3.2 mm, L direction o
ones without the corrosion exposure in the next figures. 2 Artificial ageing and subsequent
190 C
o
2 h exposure at EXCO solution 210 C
0
4.2.1. Conventional yield stress 0.01 0.1 1 10 100 1000
Fig. 7 summarizes the conventional yield stress (based on the
Artificial ageing time [hours]
nominal cross-section of the specimens) results for the three dif-
ferent temperatures. Fig. 7a shows the Rp0.2% results of the aged Fig. 6. Evaluated tensile mechanical properties (a) yield stress Rp0.2% and (b) elon-
specimens against the respective aged and corroded specimens for gation at fracture Af for different artificial ageing conditions of aluminium alloy
the 170 ◦ C artificial ageing temperature. Marked in the diagram 2024-T3 and subsequent 2 h exposure at EXCO solution.
are the regions of under- (UA), peak- (PA) and over-ageing (OA)
conditions as a function of ageing time with different colours. The
418 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424

(a) 500 mens with magenta. The available experimental test results were
Aluminum alloy 2024-Τ3 simply connected with the aid of a B-Spline curve (eye-catch) in
480
t = 3.2 mm, direction L
o
order to roughly assess the effect of the two parameters: (a) ageing
Artificial ageing at 170 C and
subsequent 2 h exposure at - 8.6 % time and (b) ageing time and subsequent exposure to the corrosive
460 EXCO solution solution. For higher artificial ageing times, the effect of corrosion
Yield stress Rp0.2% [MPa]

aged exposure seems to result to a slightly higher decrease on yield

440
aged+corroded - 5.3 % stress; the Rp0.2% decrease due to corrosion is almost minimal at
420 the latest stages of under-ageing, it increases marginally up to 8.6%
under- at the peak-ageing and ends at the same order of magnitude (5.3%
ageing
400 decrease) for the case of over-ageing conditions.
over-
ageing Respective test results can be seen in Fig. 7b for the case of 190 ◦ C
380
- 5.6 % peak- ageing temperature. As expected [5,41], the peak yield stress was
360
ageing shifted to lower ageing time (9 h from 48 h) and lower Rp0.2% value
(469 MPa from 487 MPa). It is well-known that with the lower age-
0 ing temperature higher peak strength can be achieved due to the
1E-3 0.01 0.1 1 10 100 1000 well-balanced formation of coherent and subsequent non-coherent
Artificial ageing time [hours] S and S (Al2 CuMg) precipitates, respectively [5]. According to the
(b) 500 literature [44,45] from peak-ageing and after, no more S second
Aluminum alloy 2024-Τ3
peak-
t = 3.2 mm, L direction ageing phase is being precipitated, while growth of the precipitates is
480 o
Artificial ageing at 190 C and over- observed, that leads to larger precipitates and fewer in number.
subsequent 2 h exposure at
460 EXCO solution
ageing Yield stress degradation due to the small 2 h exposure time to the
Yield stress Rp0.2% [MPa]

- 5.7 % corrosion solution is also noticed for all ageing times; for all ageing
440 conditions this decrease seems to be constant and approximately
aged
aged+corroded around 5.7%.
420
under- For the highest investigated ageing temperature, peak stress is
400
ageing even lower than that of the previous ageing temperature, Fig. 7c.
The experimental results of the artificial ageing can be seen with
- 5.8 %
380 blue line and circles, while the aged and subsequent corroded test
- 5.6 %
results with violet. The decrease due to corrosion exposure seems
360 to be around 5% for the under-ageing condition that is marginally
0 lower than the yield stress decrease due to corrosion in the peak-
1E-3 0.01 0.1 1 10 100 1000 ageing condition. In the over-ageing condition, the yield stress
Artificial ageing time [hours] decrease is slightly higher but remains in the same order of mag-
nitude (around 6.7% decrease). Taking into account all investigated
(c) 500 artificial ageing temperatures and times from Fig. 7, it can be con-
under- cluded that yield stress decrease due to corrosion is not sensitive
480 ageing enough to the artificial ageing condition of the material prior to the
460 - 4.3 % corrosion exposure.
Yield stress Rp0.2% [MPa]

440
aged
420
aged+corroded 4.2.2. Elongation at fracture
400
Elongation at fracture values for various artificial ageing times
380
- 5.6 % at 170 ◦ C can be seen in Fig. 8a (red line and squares). As expected,
360 peak- the essential yield stress increase in the peak-ageing condition is
ageing
Aluminum alloy 2024-Τ3 followed by a ductility decrease. To this end, the ductility always
340
t = 3.2 mm, direction L
o
- 6.7 % attains minimum values of approximately 8% elongation at fracture.
320 Artificial ageing at 210 C and
Hence, the increase of approximate 100 MPa in yield stress (more
subsequent 2 h exposure at over-
300 EXCO solution ageing than 26% increase) is redeemed with an essential decrease in duc-
0 tility (from 18.2% to 8.1%) that responds to more than 55% decrease.
1E-3 0.01 0.1 1 10 100 1000
It is well known that in the over-ageing condition (after 60 h arti-
Artificial ageing time [hours] ficial ageing) a partial recovery of ductility is attempted; however,
in the present protocol ductility did not exceed 9% elongation at
Fig. 7. Yield stress comparison between artificially aged against artificially aged
and subsequent corroded for 2 h specimens for (a) 170 ◦ C, (b) 190 ◦ C and (c) 210 ◦ C fracture after 98 h ageing.
artificial ageing temperatures. The respective ductility results of the aged and subsequently
corroded specimens can also be seen in Fig. 8a with magenta colour.
For the case of the T3 condition (as-received) the ductility decrease
non-artificially aged material exhibited 387 MPa yield stress while due to corrosion exposure exceeds 26%. As already discussed in pre-
after 2 h exposure its yield stress was decreased by approximately vious section, no major corrosion-induced surface pits were formed
5.6% (365 MPa). As the logarithmic scale was used to express the that could act as stress raisers and therefore degrade the mechan-
artificial ageing time in hours, it is mathematically impossible to ical properties of the specimens. Hence, this ductility decrease is
illustrate the reference condition T3 without any artificial ageing mainly attributed to the hydrogen embrittlement effect. For higher
(zero hours of artificial ageing); to cope with this problem a very ageing times, e.g. in the peak-ageing condition, lower ductility
small value of 0.015 h was used in all figures to address the refer- decrease is noticed; almost 11% (average) elongation at fracture
ence specimens without any kind of artificial ageing heat treatment. decrease is evident for the peak-aged specimens, clearly show-
The results of the artificially aged specimens were marked with red ing that the fracture mechanism (embrittlement) has changed. For
colour while the artificially aged and subsequent corroded speci- even higher ageing times (over-ageing condition) the corrosion-
 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424 419

(a) ing times the ductility decrease seems to be partially restored at its
20
peak- initial value (around 18%).
ageing For the case of an even higher ageing temperature (Fig. 8c), this
18
kind of bottleneck in the peak-ageing condition is also present.
- 26 %
Elongation at fracture Af [%]

16 However, corrosion-induced ductility decrease is lower than the

under- over-
ageing ageing previous investigated temperatures, i.e. 18% decrease against 11%
14
for the case of 170 ◦ C. The same trend can also be seen for the
12 over-ageing condition, where an increase in the ductility decrease
(restoration to the under-ageing regime values) takes place.
aged Summing up from all investigated ageing conditions, it is obvi-
10
aged+corroded
- 11 % ous that the corrosion mechanism alters with the artificial ageing
8 Aluminum alloy 2024-Τ3
t = 3.2 mm, L direction - 25 % conditions. From the essential 26% ductility decrease in the under-
o
Artificial ageing at 170 C and ageing condition, this decrease takes values around 11% for the
6 subsequent 2 h exposure at
EXCO solution
peak-ageing condition and is partially restored to the initial values
4 (around 22%) for the case of over-ageing condition. The ductility
1E-3 0.01 0.1 1 10 100 1000 decrease in the peak-ageing condition is also varying along with
Artificial ageing time [hours] the ageing temperature; with increasing ageing temperature the
(b) 20 ductility decrease in the peak-ageing regime increases (from 11%
aged at 170 ◦ C to 14% at 190 ◦ C and finally 18% at 210 ◦ C). This should
aged+corroded
18 be correlated with the artificial ageing mechanism itself; the lower
peak- over-
ageing ageing ageing temperatures permit a more balanced/stabilised and well-
Elongation at fracture Af [%]

16 distributed formation of the S precipitates that enables achieving

- 26 %
higher yield stress values, which was indeed the case in the present
14
article.
12
under-
ageing
10
- 18 % 5. Fractography
8 Aluminum alloy 2024-Τ3
t = 3.2 mm, L direction - 14 %
o

6
Artificial ageing at 190 C and The tensile test results revealed that corrosion-induced ductil-
subsequent 2 h exposure at
EXCO solution ity decrease is artificial-ageing sensitive; hence, specimens from
4 under-, peak- and over-ageing conditions were examined with
1E-3 0.01 0.1 1 10 100 1000 the aid of scanning electron microscope (SEM). Fig. 9 shows SEM
Artificial ageing time [hours] images for the under-ageing condition (without any artificial age-
(c) ing) and subsequent 2 h corrosion exposure. Fig. 9a provides a
20
peak- over- general view of the fractured specimen, showing evidence of a
18
ageing ageing ductility-based fracture mechanism due to the 45◦ inclination of
the fracture surface. Not essential surface pitting can be seen in
Aluminum alloy 2024-Τ3
Elongation at fracture Af [%]

16 t = 3.2 mm, L direction both corroded surfaces (large and small-side) and the edge notches
- 26 % o
Artificial ageing at 210 C and were formed by the large accumulative strains nearby the fracture
14 subsequent 2 h exposure at
EXCO solution region. Fig. 9b shows an image of a magnification of the frac-
12 tured area; corrosion products with varying depth can be seen
under- and below them steep surfaces are observed that is evidence of
ageing
10 quasi-cleavage fracture mechanism corresponding to hydrogen
embrittlement. In the smooth surfaces, dimples can be observed
8 - 18 %
- 22 % that were smaller in comparison with those observed in the cen-
aged tre of the tensile specimen. The size of the dimples increased with
6
aged+corroded increasing distance from the external surface of the sample, as
4 also pointed out in Ref. [34]. This could be related to hydrogen
1E-3 0.01 0.1 1 10 100 1000 embrittlement of the strengthening precipitates and/or interface
Artificial ageing time [hours] matrix/precipitates. Fig. 9c shows an image from the large cor-
roded surface, pointing out the de-alloying of some IM particles
Fig. 8. Ductility comparison between artificially aged against artificially aged and (white spots) and a few corrosion-induced surface pits. When cor-
subsequent corroded for 2 h specimens for (a) 170 ◦ C, (b) 190 ◦ C and (c) 210 ◦ C arti- roded in more mild corrosion solution than EXCO [46], AA2024-T3
ficial ageing temperatures.
exhibited substantial subsurface attack in the form of intergranular
corrosion with virtually no grain etchout. Fig. 9d shows a detail of
induced degradation changed again as the ductility decrease was such a pit, showing excessive straining and not micro-cracking coa-
“restored” at the order of magnitude around 25%. lescence. Signs of intergranular attack are also evident at the ends
Fig. 8b shows the respective results for the investigated artifi- of the pits as micro-cracks. An image from the small side surface
cial ageing times at 190 ◦ C. A similar phenomenon is also evident can be seen in Fig. 9e, giving evidence of higher pitting density in
here; the large ductility degradation due to corrosion is decreasing this surface. It is worth noticing that the pits were all concentrated
with increasing ageing time up till the peak-ageing condition. It to the mid thickness of the specimen. In the picture, a number of
reaches a minimum at the peak-ageing regime and is higher (aver- surface macro-cracks can be seen that are the outcome of micro-
age decrease of 14% instead of respective 11%) than the lower ageing cracking coalescence due to straining. Fig. 9f shows a detail of a
temperature. Peak-ageing regime seems to behave as some kind of surface pit, demonstrating the possibility that the corrosion depth
“bottleneck” in terms of ductility decrease, since with higher age- might exceed 50 ␮m.
420 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424

Fig. 9. SEM images of the reference aluminium alloy 2024 T3 specimens after exposure to EXCO solution for 2 h, (a) general view and of the fractured specimen, (b) detail
from the fracture surface, (c) and (d) view and detail from the large side surface, (e) view and detail from the small side surface of the specimen.

Fig. 10 shows the respective images of a peak-aged specimen
(48 h at 170 ◦ C), subsequently corroded and strained till fracture.
Fig. 10a shows a fraction of the fractured tensile specimen; both
corroded surfaces can be clearly seen. Excessive pitting can be
noticed in the large corroded surface, while micro- to macro- crack-
ing in the small side surface is evident. A detail from the fractured
surface can be seen in Fig. 10b; the corrosion-induced attack seems
to be uniform and localized embrittlement below the corrosion
products (at the centre of the picture) can be distinguished. The
fracture mode seems to be the classical void-coalescence due to the
large plastic strains but definitely not exhibiting such high plas-
tic strains as the previous specimen. A closest view of the large
corroded surface as well as a detail of the surface can be seen in
Fig. 10c and d, respectively. It is clear that the corrosion mechanism
changed from the under-aged condition specimen, as the corroded
surface seems to be more exfoliated, while the pits are wide and
shallow. Zhou et al. [21] showed that the intergranular corrosion
mechanism in AA2024-T351 is evident when second phase pre-
cipitates were “absent”. In a follow-up article [47], they showed
that grain boundary attack started beneath the alloy surface and
at locations where large clusters of S particles (GPB zones) were
present. Since the second phase precipitates were being formed
at the aluminium matrix due to the artificial ageing, it is eminent
that the corrosion mechanism to be altered than grain boundary
attack. A surface detail in Fig. 10d shows that surface intergranu-
lar cracks are induced due to the tensile straining, e.g. upper left
region of the photo. Fig. 10e shows the micro cracking network
of the fractured specimen at the small side surfaces of the spec-
imen. Corrosion products are also evident in this surface but not
with the gradient showed at the respective surface of the under-
aged specimen. Micro-cracking coalescence is mainly observed in
this surface; this can be started from centrally-located pits or from
 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424 421

Fig. 10. SEM images of the 48 h at 170 ◦ C (peak-aged) artificially aged aluminium alloy 2024 T3 specimens after exposure to EXCO solution for 2 h, (a) general view and of the
fractured specimen, (b) detail from the fracture surface, (c) and (d) view and detail from the large side surface, (e) view and detail from the small side surface of the specimen.

both edges. Fig. 10e shows a detail of the previous surface, giving
proof of the crack incubation at corrosion pits and coalescence due
to the tensile strain.
Respective SEM images can be seen in Fig. 11 for an over-
aged specimen (63 h at 210 ◦ C), corroded and subsequently tensile
strained till fracture. Fig. 11a shows a general view of both cor-
roded surfaces; fewer corrosion products can be seen on the large
surface while excessive deterioration of the small exposure surface
is evident. Fig. 11b shows a detail of the fracture area close to the
edge; the corrosion products seems to be very uniform and small
regions below the corrosion products exhibit a quasi-cleavage frac-
ture mechanism, e.g. at the bottom left and upper left regions. This
quasi-cleavage fracture mechanism was associated in the literature
[26] with the hydrogen embrittlement mechanism. The rest regions
of the fracture surface exhibit large dimples that are denoting large
plastic deformation before final fracture. Fig. 11c shows a photo of
the large corroded surface; the pits surface seems to be less than the
respective peak aged specimen. A magnification of this surface can
be seen in Fig. 11d, where no cracking network seems to have been
formed due to the induced strain to fracture. Small and large pits
were generated; the larger seems to have a depth around 100 ␮m.
Fig. 11e shows the respective image of the corroded small surface
of the tensile specimen. It can easily be noted that this surface was
uniformly deteriorated, without showing evidence of large pits and
cracks. This deterioration comes from the dissolution of the IM par-
ticles (S precipitates are also included) due to corrosion attack and
some grains were attacked. This dissolution is evidence of the fine
dispersion of the precipitates at the aluminium matrix and seems
to follow certain crystal planes, as also mentioned in Ref. [9]. Hor-
izontal cracking can be seen all over the corroded surface and this
is mainly due to the tensile stretching of the specimen. Fig. 11f
shows how the small surface was deteriorated; the mechanism for
422 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424

Fig. 11. SEM images of the 63 h at 210 ◦ C (over-aged) artificially aged aluminium alloy 2024 T3 specimens after exposure to EXCO solution for 2 h, (a) general view and of the
fractured specimen, (b) detail from the fracture surface, (c) and (d) view and detail from the large side surface, (e) view and detail from the small side surface of the specimen.

this surface seems to be large surface pitting but not large depth of precipitation, S-type phase is fully coherent with the matrix. With
attack. the evolution of the ageing process, more coherent precipitates
are being formed, while at the peak- and over-ageing conditions
the S phases seems to be non-coherent with the matrix. Ralston
6. Discussion
et al. [18] observed that the S-phase precipitates at the T3 or at the
under-ageing condition have thicknesses less than ∼3 nm. At the
In the previous sections, the test results were discussed as a
peak-ageing condition, a most uniform distribution of the S parti-
function of the respective ageing condition and the induced surface
cles can be noticed; a micro-galvanic corrosion takes place that is
damage of the pre-corroded specimen. In this section, the results
believed to cause corrosion of the particles, or corrosion of the sur-
will be discussed on the microstructural level and cross-linked with
rounding matrix (Al-grain) and definitely not that grain boundaries
the precipitation hardening mechanism. Fig. 12 shows the com-
that leads to the uniform corrosion products that was observed in
bined results of the evolution of yield stress with increasing ageing
the corroded specimens. At the over-ageing condition, S precipi-
time at 170 ◦ C; in the double y-axis and marked with blue line are
tates coalescence takes place due to the depleting of the smaller
the ductility decrease results due to the short 2 h exposure to the
in size S particles. The S-phase precipitates have grown in thick-
exfoliation solution. Both curves seem to have a similar trend, i.e.
ness and take values ∼8 nm or greater; such a microstructure was
the Rp0.2% increase is followed by a short decrease in ductility up to
found to be three orders of magnitude more susceptible to pitting
the peak-ageing condition. It is well known that at the first stages of

500
under-
ageing
480
S'
460
Yield stress Rp0.2% [MPa]
(semi-coherent)
440
S''
420
(coherent)
400
380
GPB zones peak-
(coherent) ageing
360
0 -32
0 20 40
Artificial ageing time [hours]
Fig. 12. Yield stress of artificially aged specimens and ductility decrease after
2 h exposure to the corrosion solution and respective stages of the precipitation
strengthening mechanism.
[18], thus explaining the resulting corrosion products presented in
Fig. 11. In Ref. [19], the area fraction of the second phase (combined
S-phase and GPB zones) was determined to be ∼2% and ∼12% for
the under- and over-ageing conditions, respectively.
Hydrogen embrittlement is a possible mechanism for the expla-
nation of the ductility decrease in the absence of any formed
pits. In a series of articles, Haidemenopoulos and co-workers mea-
sured desorbed hydrogen of pre-corroded AA2024 using gas-phase
chromatography for different immersion times [26,48]. Recently,
Kamoutsi et al. [49] showed that hydrogen trapping takes place
and the T4 trap state was associated with the S-phase. The natural
aged specimens exhibited higher desorption rate than the non-
aged specimens, thus implying that the GPB zones and respective
S-phase are associated with the hydrogen trapping. Diffused hydro-
gen inside the material may be trapped at the matrix/strengthening
interphases (T4 trap state). This leads to a local embrittlement of
the interphase of the matrix/precipitates that is responsible for pre-
mature cavity nucleation, plastic growth and finally coalescence
during tensile testing.
Definitely the ductility degradation mechanism at the T3 condi-
tion is the hydrogen embrittlement as no serious pitting occurred
at the corroded surfaces. At this condition, there is a possibility
to have already formed some S-type particles at the grain bound-
aries due to stretching during manufacturing and perhaps natural
ageing of the alloy prior to Laboratory testing. Due to the micro-
galvanic interactions, intergranular corrosion is noticed for large
corrosion exposure times, while their interphase with the sur-
rounding matrix is supposed to act as hydrogen trap. Larignon et al.
[24] also noted the possibility that the grain boundaries themselves
could act as pathways for hydrogen diffusion. Therefore, hydro-
gen traps so close to the grain boundaries might cause premature
fracture through the grains when tensile stressed, thus explaining
the essential loss in ductility (∼26%). With the increase of artifi-
cial ageing at the under-ageing condition, GPB and subsequently
S particles are also being precipitated inside the ␣-matrix (nucle-
ation and growth kinetics for precipitates). This is translated as new
trapping sites that are not obligatory to be on the alloy grain bound-
aries. The trapping capacity of these sites is speculated to be small
as the precipitates are still coherent with the matrix and hydro-
gen can actually take interstitial positions in the crystal lattice. At
the peak-ageing condition, a very good distribution of S precip-
itates, with size ranging from 3 to 8 nm, are being formed inside
the matrix, the latter being semi-coherent with the matrix. Litera-
ture has demonstrated that pitting potential is almost tripled due
to the very-well distributed particles in the grain [18,19]. The last
stage is the over-ageing condition, that few and large particles (due
N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424 423
0 to the coalescence mechanism) are being precipitated in the Al-
over-
ageing matrix. Results from all investigated ageing temperatures showed
Elongation at fracture Af decrease [%]
-4
yield stress that ductility decrease returns to the very high level of ∼26 %, same
-8 as the T3 and subsequent 2 h corrosion exposure levels. The over-
S ageing pitting potential seems to be decreased when qualitatively
-12
(incoherent) compared against the respective SEM images of the peak-aged
-16 specimens. As in this ageing condition, the S-type particles are now
Aluminum alloy 2024-Τ3
t = 3.2 mm, L direction
fully incoherent with the aluminium matrix, it is speculated that
-20
o
Artificial ageing at 170 C the particle/matrix interphase might act as a high hydrogen capac-
-24 ity trap. Hence, in this condition, the hydrogen embrittlement deals
with few and high-capacity traps that quite embrittles the matrix,
-28 despite the very short corrosion exposure time.
ductility decrease
60 80 100 7. Conclusions
The findings of the present work can be summarised briefly:
1. Pitting formation at the T3 condition (without any artificial age-
ing) after 2 h exposure to exfoliation corrosion solution was not
essential and therefore, the almost 26% ductility decrease was
not attributed to the classical notch effect mechanism, but to
the well-established hydrogen embrittlement mechanism.
2. Corrosion-induced yield stress decrease was less sensitive to
the change of microstructure, showing that in all tempers this
decrease is of the same order of magnitude and around 5%.
3. Elongation at fracture decrease due to the corrosion exposure
was correlated with the formation of S-phase precipitates. The
highest ductility decrease was noticed for the T3 condition, the
lowest decrease for the peak-ageing (with formed S and S pre-
cipitates), while it was partially ‘restored’ at the initial level at
the over-ageing condition.
4. This decrease was found to be also ageing temperature sensitive,
as the lowest temperature gave the least ductility decrease, thus
giving proof of the role (nucleation and growth kinetics) of the
precipitates on the corrosion mechanism.
5. Hydrogen embrittlement mechanism and the size/number of the
S-type precipitates were correlated.
6. Discussion on the hydrogen capacity traps of the S phase was
attempted regarding the coherency of the particles with the
matrix as well as their size, number and their location at the
grain boundaries.
Acknowledgements
The authors gratefully acknowledge the financial support of the
European Union (European Social Fund—ESF) and Greek national
funds through the Operational Program “Education and Lifelong
Learning” of the National Strategic Reference Framework (NSRF) –
Research Funding Program: “Archimedes III—Technological Edu-
cational Institute of Piraeus – Experimental and theoretical
investigation of mechanical properties degradation of the aeronau-
tical aluminium alloy 2024 due to corrosion” (MIS 383575).
The authors would also like to thank Wolfgang Dietzel, Carsten
Blawert, Volker Heitmann and Petra Fischer at the Institute of
Materials Research, Corrosion and Surface Technology of Helmholtz
Zentrum Geesthacht, Germany for the microstructure, fracture sur-
face analysis of corroded specimens and fruitful discussions on the
manuscript.
References
[1] A. Boag, A.E. Hughes, N.C. Wilson, A. Torpy, C.M. MacRae, A.M. Glenn, T.H.
Muster, How complex is the microstructure of AA2024-T3? Corros. Sci. 51
(2009) 1565–1568.
[2] R.G. Buchheit, R.P. Grant, P.F. Hlava, B. Mckenzie, G.L. Zender, Local dissolution
phenomena associated with S phase (Al2 CuMg) particles in aluminum alloy
2024-T3, J. Electrochem. Soc. 144 (1997) 2621–2628.
424 N.D. Alexopoulos et al. / Corrosion Science 102 (2016) 413–424
[3] E.A. Starke, J.T. Staley, Application of Modern aluminum alloys to aircraft,
Prog. Aerosp. Sci. 32 (1996) 131–172.
[4] Y.A. Bagaryatsky, Structural changes on aging Al–Cu–Mg alloys, Dokl. Akad.
SSSR 87 (1952) 397–559.
[5] L.F. Mondolfo, Aluminum Alloys—Sructure and Properties, Butterworth,
London, 1976.
[6] S.P. Ringer, K. Hono, I.J. Polmear, T. Sakurai, Nucleation of precipitates in aged
Al–Cu–Mg–(Ag) alloys with high Cu:Mg ratios, Acta Mater. 44 (1996)
1883–1898.
[7] S.P. Ringer, S.K. Caraher, I.J. Polmear, Response to comments on cluster
hardening in an aged Al–Cu–Mg alloy, Scr. Mater. 39 (1998) 1559–1567.
[8] S.C. Wang, M.J. Starink, Two types of S precipitates in Al–Cu–Mg alloys, Acta
Mater. 55 (2007) 933–941.
[9] V. Guillaumin, G. Mankowski, Localized corrosion of 2024-T351 aluminium
alloy in chloride media, Corros. Sci. 41 (1999) 421–438.
[10] G.O. Ilevbare, O. Schneider, R.G. Kelly, J.R. Scully, In-situ confocal laser
scanning microscopy of AA2024-T3 corrosion metrology: 1. localized
corrosion of particles, J. Electrochem. Soc. 151 (2004) B453–B464.
[11] O. Schneider, G.O. Ilevbare, J.R. Scully, R.G. Kelly, In-situ confocal laser
scanning microscopy of AA2024-T3 corrosion metrology: 2. trench formation
around particles, J. Electrochem. Soc. 151 (2004) B465–B472.
[12] L. Lacroix, L. Ressier, C. Blanc, G. Manikowski, Statistical study of the corrosion
behavior of Al2CuMg intermetallics in AA2024-T351, J. Electrochem. Soc. 155
(2008) C8–C15.
[13] L. Lacroix, L. Ressier, C. Blanc, G. Manikowski, Combination of AFM, SKPFM,
and SIMS to study the corrosion behavior of S-phase particles in
AA2024-T351, J. Electrochem. Soc. 155 (2008) C131–C137.
[14] A.E. Hughes, T.H. Muster, A. Boag, A.M. Glenn, C. Luo, X. Zhou, G.E. Thompson,
D. McCulloch, Co-operative corrosion phenomena, Corros. Sci. 52 (2010)
665–668.
[15] T. Hu, H. Shi, T. Wei, F. Liu, S. Fan, E.H. Han, Cerium tartrate as a corrosion
inhibitor for AA 2024-T3, Corros. Sci. 95 (2015) 152–161.
[16] H. Shi, Z. Tian, T. Hua, F. Liu, E.H. Han, M. Taryba, S.V. Lamaka, Simulating
corrosion of Al2CuMg phase by measuring ionic currents, chloride
concentration and pH, Corros. Sci. 88 (2014) 178–186.
[17] P. Campestrini, E.P. van Westing, H.W. van Rooijen, J.H. de Wit, Relation
between microstructural aspects of AA2024 and its corrosion behaviour
investigated using AFM scanning potential technique, Corros. Sci. 42 (2000)
1853–1861.
[18] K.D. Ralston, N. Birbilis, M. Weyland, C.R. Hutchinson, The effect of precipitate
size on the yield strength-pitting corrosion correlation in Al–Cu–Mg alloys,
Acta Mater. 58 (2010) 5941–5948.
[19] K.D. Ralston, N. Birbilis, M.K. Cavanaugh, M. Weyland, B.C. Muddle, R.K.W.
Marceau, Role of nanostructure in pitting of Al–Cu–Mg alloys, Electrochem.
Acta 55 (2010) 7834–7842.
[20] C. Luo, Z. Zhou, G.E. Thompson, A.E. Hughes, Observations of intergranular
corrosion in AA2024-T351: the influence of grain stored energy, Corros. Sci.
61 (2012) 35–44.
[21] X. Zhou, C. Luo, Y. Ma, T. Hashimoto, G.E. Thompson, A.E. Hughes, P. Skeldon,
Grain-stored energy and the propagation of intergranular corrosion in
AA2xxx aluminium alloys, Surf. Interface Anal. 45 (2013) 1543–1547.
[22] S.P. Knight, M. Salagaras, A.M. Wythe, F. De Carlo, A.J. Davenport, A.R.
Trueman, In situ X-ray tomography of intergranular corrosion of 2024 and
7050 aluminium alloys, Corros. Sci. 52 (2010) 3855–3860.
[23] S.P. Knight, M. Salagaras, A.R. Trueman, The study of intergranular corrosion
in aircraft aluminium alloys using X-ray tomography, Corros. Sci. 53 (2011)
727–734.
[24] C. Larignon, J. Alexis, E. Andrieu, L. Lacroix, G. Odemer, C. Blanc, Combined
Kelvin probe force microscopy and secondary ion mass spectrometry for
hydrogen detection in corroded 2024 aluminium alloy, Electrochim. Acta 110
(2013) 484–490.
[25] P.V. Petroyiannis, A.T. Kermanidis, P. Papanikos, Sp.G. Pantelakis,
Corrosion-induced hydrogen embrittlement of 2024 and 6013 aluminum
alloys, Theor. Appl. Fract. Mech. 41 (2004) 173.
[26] H. Kamoutsi, G.N. Haidemenopoulos, V. Bontozoglou, Sp.G. Pantelakis,
Corrosion-induced hydrogen embrittlement in aluminium alloy 2024, Corros.
Sci. 48 (2006) 1209.
[27] N.D. Alexopoulos, P. Papanikos, Experimental and theoretical studies of
corrosion-induced mechanical properties degradation of aircraft 2024
aluminum alloy, Mater. Sci. Eng. A 498 (2008) 248.
[28] N.D. Alexopoulos, On the corrosion-induced mechanical degradation for
different artificial aging conditions of 2024 aluminum alloy, Mater. Sci. Eng. A
520 (2009) 40–48.
[29] E. Charitidou, G. Papapolymerou, G. Haidemenopoulos, N. Hasiotis, V.
Bontozoglou, Characterization of trapped hydrogen in exfoliation corroded
aluminum alloy 2024, Scr. Mater. 41 (1999) 1327–1332.
[30] G. Young, J. Scully, The diffusion and trapping of hydrogen in high purity
aluminum, Acta Mater. 18 (1998) 6337–6349.
[31] G. Scamans, C. Tuck, in: Z. Foroulis (Ed.), Sensitive Fracture of Engineering
Materials, The Metallurgical Society of the AIME, 1979, pp. 464–483.
[32] F. Scholes, S. Furman, A. Hughes, T. Markley, Corrosion in artificial defects. I:
Development of corrosion, Corros. Sci. 48 (2006) 1812–1826.
[33] H. Saitoh, Y. Iijima, K. Hirano, Behaviour of hydrogen in pure aluminium, Al-4
mass% Cu and Al-1 mass% Mg2 Si alloys studied by tritium electron
microautoradiography, J. Mater. Sci. 29 (1994) 5739–5744.
[34] C. Larignon, J. Alexis, E. Andrieu, G. Odemer, C. Blanc, The contribution of
hydrogen to the corrosion of 2024 aluminum alloy exposed to thermal and
environmental cycling in chloride media, Corros. Sci. 69 (2013) 211–220.
[35] C. Blanc, G. Mankowski, Pit propagation rate on the 2024 and 6056 aluminium
alloys, Corros. Sci. 40 (1998) 411–429.
[36] A. Boag, R.J. Taylor, T.H. Muster, N. Goodman, D. McCulloch, C. Ryan, B. Rout,
D. Jamieson, A.E. Hughes, Stable pit formation on 2024-T3 in a NaCl
environment, Corros. Sci. 52 (2010) 90–103.
[37] T. Marlaud, B. Malki, A. Deschamps, B. Baroux, Electrochemical aspects of
exfoliation corrosion of aluminium alloys: the effects of heat treatment,
Corros. Sci. 53 (2011) 1394–1400.
[38] K.S. Ghosh, M.D. Hilal, S. Bose, Corrosion behaviour of 2024 Al–Cu–Mg alloy of
various tempers, Trans. Nonferrous Met. Soc. China 23 (2013) 3215–3227.
[39] W. Zhang, G.S. Frankel, Transitions between pitting and intergranular
corrosion in AA2024, Electrochem. Acta 48 (2003) 1193–1210.
[40] L. Korb, D. Olson (Eds.), ASM Handbook: Corrosion, 13, 9th ed., ASM
International, Metals Park, OH, USA, 1992.
[41] J.E. Hatch, Aluminum: Properties and Physical Metallurgy, ASM International,
1984, pp. 245, January.
[42] N.D. Alexopoulos, Z. Velonaki, C. Stergiou, S.K. Kourkoulis, Artificial ageing
and mechanical properties of aluminium alloy 2024, submitted for
publication, 2015.
[43] N.D. Alexopoulos, C.J. Dalakouras, P. Skarvelis, S.K. Kourkoulis, Accelerated
corrosion exposure in ultra-thin sheets of 2024 aircraft aluminium alloy for
glare applications, Corros. Sci. 55 (2012) 289–300.
[44] J.W. Martin, Precipitation Hardening, 2nd ed., Butterworth-Heinemann,
Oxford, United Kingdom, 1998.
[45] I. Polmear, Light Alloys: Metallurgy of the Light Alloys, 3rd ed., Metallurgy and
Materials Science, Arnold, Great Britain, 1995.
[46] A.M. Glenn, T.H. Muster, C. Luo, X. Zhou, G.E. Thompson, A. Boag, A.E. Hughes,
Corrosion of AA2024-T3 Part III: propagation, Corros. Sci. 53 (2011) 40–50.
[47] X. Zhou, C. Luo, T. Hashimoto, A.E. Hughes, G.E. Thompson, Study of localized
corrosion in AA2024 aluminium alloy using electron tomography, Corros. Sci.
58 (2012) 299–306.
[48] G.N. Haidemenopoulos, N. Hassiotis, G. Papapolymerou, V. Bontozoglou,
Hydrogen absorption into aluminum alloy 2024-t3 during exfoliation and
alternate immersion testing, Corrosion 54 (1998) 73–78.
[49] H. Kamoutsi, G.N. Haidemenopoulos, V. Bontozoglou, P.V. Petroyannis, Sp.G.
Pantelakis, Effect of prior deformation and heat treatment on the
corrosion-induced hydrogen trapping in aluminium alloy 2024, Corros. Sci. 80
(2014) 139–142.