Mechanics of Materials
Mechanics of Materials
Mechanics of Materials
Mechanics of Materials
journal homepage: www.elsevier.com/locate/mechmat
A R T I C L E I N F O A B S T R A C T
Keywords: Friction stir welding (FSW), when in butt-joint configuration, was found to produce high-mechanically per
FSW forming joints. However, when in overlap configuration, stress concentration is noticed at the weld edges that
Adhesive bonding significantly reduces the mechanical performance of the joints. Additionally, the welding regions tend to enhance
Corrosion
localized corrosion attack due to microstructural changes. In the present work, a new hybrid joint - that is a
Lap joints
Tension
combination of friction stir welding with a prior step of adhesive bonding of the overlap surface - was investi
gated. Friction stir welded, adhesive bonded and hybrid single lap joints were exposed to exfoliation corrosion
solution - for different exposure times - and mechanically tested to investigate the effect of adhesive bonding on
the mechanical behavior of pre-corroded joints. The hybrid joint (adhesive bonding + friction stir welding)
showed excellent tensile mechanical performance in terms of deformation, exhibiting approximately 100 %
increase when compared against the respective values of adhesive bonded and/or friction stir welded joints. This
was attributed to the elimination of the ‘hook defect’ in the friction stir welded area as well as to the out-of-plane
displacement restriction of the bonded area. The corrosion-induced degradation mechanisms of the hybrid joint
were also investigated and compared against the friction stir welded and the adhesive bonded joints for various
exposure times. The hybrid joints outperform the friction stir welded and adhesive bonded joints for all the
investigated corrosion exposure times; after very long corrosion exposure times (e.g. 48 h), the hybrid joints
maintained their maximum tensile strength by almost 90 %, while the adhesive bonded and the friction stir
welded joints retained 85 % and 59 %, respectively. Almost double fracture deformation was noticed for the
hybrid joints for all the investigated corrosion exposure times. This superiority was attributed to the dual pro
tection of the intermediate adhesive film; initially the sealing effect protects the lap joint at the primary corrosion
stages until debonding takes place and additionally protects the friction stir welded area in-between the lapped
sheets from localized corrosion on the interface between the thermo-mechanically affected and heat affected
zones.
* Corresponding author.
E-mail address: nalexop@aegean.gr (N.D. Alexopoulos).
https://doi.org/10.1016/j.mechmat.2021.104122
Received 11 March 2021; Received in revised form 13 September 2021; Accepted 25 October 2021
Available online 28 October 2021
0167-6636/© 2021 Elsevier Ltd. All rights reserved.
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
aspect, with the major variable to be the fuel consumption; a significant cathodic to base material for AA5083. Formation of galvanic cells be
reduction in the aircraft energy consumption can be achieved through tween the BM and the welding zones due to different electrochemical
weight savings, along with decrease in material waste and potential is enhanced by the heating-induced microstructural trans
manufacturing time. The need to reduce energy consumption and formations. Copper (Cu) dissolution in the aluminium alloy matrix of
manufacturing costs while maintaining or increasing structural integrity AA2024-T351, due to local melting and the subsequent S-type
in aerospace applications is driving the aeronautical manufacturers to (Al2CuMg) phase dissolution, increases the corrosion potential of the
wards the development of lightweight metallic structures through novel matrix resulting in localized corrosion attack according to Liu et al.
manufacturing and joining techniques. Welding, bonding, and extrusion (2005). Generally the Al–Cu alloys (e.g. AA 2024, AA 2014, AA2219)
techniques are continuously decreasing the exploitation of rivets in light show reduced corrosion resistance in the HAZ, as the phases in the grain
weight applications as was mentioned in Lenczowski (2002), such as in boundaries of this zone are more anodic when compared to the other
aircraft structures for material savings because of the absence of other zones of the weld, as reported by Rao and Rao (2004). Xu et al. (2009)
materials like sealants, intermediate rivets, lock bolt rivets, etc., as showed that the top of the nugget zone in AA2219 friction stir welded
described in Alexopoulos et al. (2016) and lead time saving calculations joints are more corrosion-resistant than the bottom surface of the welded
according to Gialos et al. (2018), where up to 67 % time savings can be joints. Proton et al. (2013) found similar results and this was attributed
achieved during manufacturing of subscale components by exploiting to the existence of larger intermetallic phases in these zones. Likewise,
the laser beam welding process. Jariyaboon et al. (2007) showed that the welded zones of AA2024 are
Exploitation of high-strength aluminium alloys in aircraft fuselage more sensitive to corrosion attack than the base material itself.
and wings parts accelerated the necessity to investigate for new welding The face-to-face (butt) joint configuration is not always a feasible
techniques for joining such aerospace structures, since they were joint type due to tolerance issues and difficulties in the achievement of
considered as non-weldable materials for fusion welding due to their zero gap weld as reported by Shultz et al. (2010). Several trials on FSW
poor solidification microstructure and extensive porosity formation ac in the overlap configuration, showed that although the process can be
cording to Mishra et al. (2005). Alternative joining methods have been successfully applied, nevertheless the mechanical properties of the
developed during the last years, such as the welding at solid-state or the welded joint may suffer a significant decrease. Fersini and Pirondi
adhesive bonding. The technology of Friction Stir Welding (FSW) was (2008) investigated the fatigue properties of friction stir welded
found to be capable of welding high strength precipitation-hardened single-lap joints, where they showed that the fatigue efficiency of
alloys (e.g. AA2024 and AA7075), since it can overcome the disadvan single-lap joints with respect to the butt joints is only 18 % at 105 cycles.
tages of the conventional fusion welding processes (e.g. solidification Song et al. (2014), also investigated the mechanical properties of friction
cracking and/or porosity), as well as to create high performance welded stir welded lap joints, where they revealed that a hardness decrease in
joints with lower distortions and residual stresses, e.g. Bussu et al. (2003) HAZ as well as in SZ of the top AA2024 sheet was noticed due to
studied the role of residual stresses on fatigue crack propagation in thermal-induced microstructural transformations. Fersini and Pirondi
friction stir welded AA2024-T351 specimens and showed that the lon (2007) showed that the material flow in friction stir welded overlaps,
gitudinal residual stresses in the plane joint line are lower than the results in an upward tip at the edge of the weld, which in combination
respective measured away from the weld line. John et al. (2003) re with the out-of-plane bending acts as a crack-like defect that is often
ported that friction stir welded 7050-T7451 specimens exhibited low called in the literature as the “hook defect”. Papadopoulos et al. (2013)
residual stresses, thus making it an environmentally friendly joining proposed a multiple-pass welding so as to eliminate this kind of defect
technique, with high quality performance joints. Niu et al. (2017) and consequently to increase the joining efficiency of the overlap joints.
showed that the joint efficiency was higher than 80 % on AA2024 fric Despite of the better results in terms of mechanical performance, the
tion stir welded joints, while Jata et al. (2000) reported that for welding exploitation of a multiple-pass friction stir weld requires higher
AA7050, the joining efficiency was more than 75 %. The excellent me lead-time and increased tool wear, thus resulting into a non-sustainable
chanical properties of the friction stir welded joints have been reported process. In this case, adhesive bonding technique can be combined with
in the research of Genevois et al. (2005), where a quantitative investi the overlap friction stir welding so as to decrease the out-of-plane
gation of the precipitations was performed on AA2024 friction stir bending and possibly to increase the joint strength. Chowdhury et al.
welded joints. Additionally, Zhoua et al. (2006) investigated the fatigue (2013) reported that the prior application of an adhesive layer in be
properties of a friction stir welded AA5056 alloy, where it was shown tween the sheets to be welded during overlap FS spot welding, resulted
that the friction stir welded joints exhibited 6 to 14 times higher fatigue in increased mechanical performance of the joints. Adhesive bonding
life than the MIG welded joints. Nevertheless, these welds were highly has the advantage of exploiting a continuous adhesive layer, while it is
susceptible to corrosion and stress corrosion cracking, although these well known that riveting always exhibits lower mechanical properties
issues may be surpassed through a proper, post-weld heat treatment. The due to the stress concentration around the holes. To this end, smaller
exposure of the welded joints to environments with varying chemical overlap lengths can be exploited with the adhesive bonded joints, than
composition and operating temperature can significantly deteriorate its the traditional riveted joints. This definitely allows for significant weight
mechanical performance by the corrosion exposure. For example, savings in aircraft structures. The use of adhesive bonding as a stand
Jamasri et al. (2011) noticed that the fatigue properties of resistance alone solution in the assembly process of light-weight structures is
spot-welded joints in corrosion solution was essentially lower than the limited by several disadvantages, the most significant being like the
respective performed in air. curing conditions of the adhesive layer. Perhaps a combination of fric
The four different zones of friction stir welded joints, typically tion stir welding and adhesive bonding processes would decrease several
named as stir zone (SZ), thermo-mechanically affected zone (TMAZ), disadvantages so as to increase the sustainability of the process to be
heat-affected zone (HAZ) and base material (BM), exhibit different used in light-weight structures. There is a possibility that the exploita
corrosion behavior due to the microstructural differences caused by tion of the adhesive bonding technology may be capable of improving
heating and plastic deformation during the friction-stir welding process, the corrosion resistance of the joints as well. In fact, Brent (2007) in a
as described by Sullivan et al. (2008). This corrosion potential difference relevant patent proposed the use an adhesive layer in between the sur
was also mentioned by Squillace et al. (2004), where a comparison be faces to be welded in order to protect the welded joints from the
tween FSW and TIG welded joints was performed regarding the micro corrosion exposure. Finally, Brown (2008) found that the exploitation of
structural changes and the corrosion resistance in AA2024-T3 sealants in friction stir spot welds (FSSW) resulted in both, increased
butt-welded joints; enhanced pitting tendency was evident for the base mechanical properties and higher corrosion resistance of the welded
material, while the weld seam and the HAZ showed a passive behavior. joints. The sealant in-between the welded sheets, when appropriately
Similar results were found in Zucchi et al. (2001) where FSW weld was applied with the proper surface treatment, protected the welded joint by
2
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
Table 1
Chemical composition of AA2024-T3 as supplied (% mass).
Si Fe Cu Mn Mg Cr Zn Ti Ti + Zr Others (Total) Al
0.08 0.13 4.5 0.48 1.5 0.03 0.2 0.04 0.04 0.03 Balance
the corrosion exposure. the longitudinal (L) rolling direction according to ASTM E8 (2016)
In the present article, the effect of corrosion exposure on a novel specification (refer to Fig. 1). The machined surfaces (thickness of the
“hybrid” overlap joining method of an Al–Cu–Mg alloy (AA2024- T3) sheet) were polished up to 1200 grit.
was investigated, where the hybrid joint combined adhesive bonding A schematic representation with geometrical dimensions of the
and friction stir welding at the overlap configuration. Tensile mechan configuration of single lap joint specimen is shown in Fig. 2. All the
ical tests were carried out on non-corroded as well as on pre-corroded joints were manufactured according to the procedures and designations
overlap joints. The investigation revealed the degradation mechanisms described in Braga et al. (2019). The overlap surfaces of the specimens
of the hybrid joint with the increasing corrosion exposure time. To this were surface treated through phosphoric acid anodization (PAA) ac
end, the respective tests were also performed on friction stir welded cording to standard ASTM D3933 (2017). The adhesive bonded joints -
joints as well as on adhesive bonded joints, so as to highlight the ad with an overlap width of 40 mm and 0.2 mm thickness of the adhesive
vantages of the hybrid joints in terms of corrosion resistance. layer - were produced in an aluminium mold and then they were inserted
in a hot plate press so as to cure the adhesive for 1 h at 120 ◦ C. The
2. Material and experimental procedure commercially available structural adhesive Araldite 420 from Hunst
man® was used in the present investigation.
The material used in the present investigation was wrought All the welded joints were produced on the FSW system at Hereon
aluminium alloy 2024-T3 in bare condition (non-Alclad), which was Helmholtz-Zentrum in Geesthacht, Hamburg in Germany, refer to
received in sheet form of 1.6 mm nominal thickness. The chemical Fig. 3a. A threaded conical tool with 5 mm diameter was manufactured
composition of AA2024-T3 is presented in Table 1 in weight percentage, and used in the present study (Fig. 3b) and was attached to a grooved
according to the material certificate of the supplier. shoulder of 13 mm diameter (refer to Fig. 3c). The overlap width of the
Initially, standard size tensile specimens - with dimensions 10 mm × FSW joints was equal to 40 mm, while the upper processed area (tool
1.6 mm thickness in the reduced cross-section and total length of 150
mm - were machined from the AA2024-T3 non-Alclad sheet parallel to
Fig. 2. Schematic representation of single lap joint specimen configuration top view (upper – tool side) and side view (down).
3
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
side) can be seen in Fig. 2. The welding process window was chosen then they were subjected to tensile testing in an Instron 100 kN testing
based on some already published data, e.g. Papadopoulos et al. (2013), frame with crosshead displacement rate equal to 0.7 mm/min. Fig. 4
as well as on welding parameter studies performed by Braga et al. shows images of several joints mounted at the jigs of the testing machine
(2019). before and after the execution of the testing protocol. Essential tilting
Sanding, degreasing and 3 M® AC130 surface pre-treatment was was noticed at the end of the tests, even though tabs at the ends of the
followed for the adhesive bonded-friction stir welded joints (hybrid specimens were attached as shown in Fig. 2 and according to the rele
joints), as it is commonly used for aeronautical repair and is a Sol-Gel vant testing standard ASTM D5656 (2017). This tilting effect was
anodization replacement. It was used in this experimental procedure - noticed in several other relevant articles on overlap joints, e.g. in Braga
instead of the PAA surface treatment - for capacity issues, since larger et al. (2019) and will be discussed explicitly in the following.
dimensions’ joints were needed for the welding process that could not fit Finally, the out of plane displacement in the overlap joints was
in the treatment bath (use of anodization procedures were impeded). measured through 3D Digital Image Correlation (DIC) analysis on
Welding was performed after closing down the joint - when the top sheet selected joints of high exposure times, e.g. 24 h and 96 h in this case.
is set above the adhesive layer already adhered to the bottom sheet - and These extremely high exposure times were selected in order to overcome
with the adhesive being in a non-cured condition. As was the case for the the sealing protection of the adhesive layer in the first stages of corro
other two investigated joint types, also the hybrid joints (AB + FSW) had sion. A Vic-3D DIC system was used in the present study, manufactured
an overlap width of 40 mm. by Correlated Solutions©. Two 4.1 MPixel CMOSIS cameras were used
Initially, the effect of corrosion exposure on the tensile mechanical that were both equipped with Schneider-Kreuznach 16 mm focal dis
properties of the bare material (BM) was investigated before assessing tance lenses. The top surface of the overlap joint with dimensions of 25
the corrosion behavior of the joints. All the surfaces to be corroded were mm wide x 40 mm height was the measuring surface area of the cameras.
cleaned with alcohol according to ASTM G1 (2017) standard. Masking of Before testing, the surface corrosion products were removed with a
the tensile specimens’ area to be inserted in the testing machine’s grips metal wire brush, as this layer is composed of corroded accumulated
was performed with the use of appropriate insulating PVC tape. Addi products which are brittle and do not exhibit deformation along with the
tionally, masking of the side-surfaces (thickness area) was performed specimen. Afterwards, the specimens were painted with a thin-layered
(refer to Fig. 1) to simulate real corrosion exposure conditions of the white base coating and sprayed randomly with black spot speckle in
specimens (since the cross-section is normally not exposed to corrosion appropriate size, so as to ensure that the features within the obtained
in real aircraft structures). This was mandatory, since the corrosion ef image pair are uniquely identifiable. The stereo camera system was then
fect was found to be much more aggressive on the side-surfaces of the calibrated using a purpose made calibration plate to determine the
specimens - where the corrosion products create wedging stresses within relative position of cameras. At least 5 reference images were taken and
the elongated grains, e.g. Posada et al. (1997) - having a contribution then acquisition was made at 2 Hz frequency.
percentage to the total ductility decrease of more than 60 % according to
Charalampidou et al. (2021). Severer corrosion attack on the specimen’s 3. Results and discussion
side-surfaces was also revealed for AA2024-T3 in Pantelakis et al.
(2016). The relative surfaces of the joints were also masked in a similar The experimental results are presented and discussed in detail to
manner. reveal the corrosion-induced degradation mechanisms of the investi
Exposure to the exfoliation corrosion solution (here after called gated joints.
EXCO) was performed for various exposure times according to ASTM
G34 (2018) standard, for both base material (tensile specimens) as well 3.1. Mechanical behavior of non-corroded joints
as the three different investigated cases of the single lap joints. The
corrosive solution was consisted of (a) sodium chloride (4.0 M), (b) A comparison between the tensile mechanical behavior of the three
potassium nitrate (0.5 M) and (c) nitric acid (0.1 M) that were all diluted different cases of reference (non-corroded) lap joints, e.g. adhesive
in 1 l of distilled water. The volume of corrosion solution was calculated bonded (AB), hybrid (AB + FSW) and friction stir welded (FSW) joints is
to be analogous to the exposure area of the specimens and it was pre attempted in this section and the relevant typical load - displacement
cisely 20 ml/cm2. The corrosion solution as well as the corrosion
exposure tests were performed at constant room temperature.
The pre-corroded specimens were immediately cleaned from the
corrosion surface products according to ASTM G1 (2017) standard and
Fig. 4. Images of the hybrid (AB + FSWed) specimens (reference – without Fig. 5. Typical load - displacement curves of the three different investigated
corrosion) on the testing machine of the upper side (tool side) on the left, (b) joints, namely (i) adhesive bonded (AB), (ii) adhesive bonded and then friction
lower side (root side) on the middle and typical fracture of the joint with stir welded (AB + FSWed) and finally (iii) friction stir welded (FSWed) of
inclination of the sheets on the right. AA2024-T3 single lap joints without any corrosion exposure.
4
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
curves are graphically presented in Fig. 5. It is obvious that the adhesive deformation values. The tilting degree was quantified with image
bonded joints exhibited high load to fracture values of approximately analysis and will be discussed in the following section. From the me
15.5 kN and 4.5 mm displacement to fracture. The possibility to FS weld chanical point of view, this tilting effect disrupts the uniform shear in
the joint after the adhesive bonding application (hybrid joint), results to the overlap area, and besides shear, tension occurs at both edges of the
increased mechanical properties in terms of load capacity and fracture overlap area. The combined loading at the edges locally downgrades the
deformation of the joint. Application of friction stir welding after ad joint - as it is well known that adhesives do not work well under tension
hesive bonding inhibits a slight increase of approximately 6 % in the (local fracture) - and debonding takes place. Thus, the integrity of the
mechanical strength of the joint (maximum load to fracture), while a joint for high mechanical loading levels lies on the capability of the
duplication of the total deformation capability (+ 100 %) of the hybrid bonded area at the mid overlap region to withstand the shear loadings as
joint is noticed. However, the friction stir welded joint exhibits lower well as debonding coming from the edges of the overlap area. The whole
load values and lower displacement to fracture by approximately 0.70 concept of the predominant described mechanisms for adhesive bonded
mm than the respective of the adhesive bonded joints. However, the joints, resulted to 15.6 kN and 4.69 mm average values of maximum
values of the peak load and displacement to fracture for the friction stir load and displacement to fracture, respectively. These observations are
welded joints are significantly higher than the respective for riveted well in accordance with the literature results, e.g. Pitta et al. (2018),
joints of AA2024-T3 specimens, referred to Pitta et al. (2018). For where the failure mode for pure adhesive bonded metal-to-metal joints
instance, AA2024-T3 (metal to metal configuration) riveted lap joints of AA2024-T3 was characterized adhesive failure - with all the adhesive
failed at 5.46 kN peak load and 1.3 mm displacement to fracture, while attached to the substrate – as well as to secondary bending in the joint’s
the displacement of 1.3 mm is due to the rivets shear observed in sub edges. Bonded joints showed smoother load transfer due to the adhesive
strate or doublers. Additionally, pure adhesive bonded joints withstand spread all over the overlap area.
higher mechanical load than pure riveted joints, with the metal-to-metal The loading mechanism in the friction stir welded joint seems to be
adhesive bonded joint of AA2024-T3 exhibiting an average strength of quite different than the respective of the adhesive bonded joint. Fig. 6
23 kN with displacement to fracture of 1.6 mm. With a factor of safety shows the loading mechanism in a sketch where it is obvious that only
1.5, the calculated limit load of pure adhesive bonded metal-to-metal the welded area works under shear and definitely not the entire overlap.
joint is 15.3 kN for safe operation that is in accordance with the re To this end, tilting also occurs with mechanical stretching of the joint
sults of 15.5 kN of the present study. Regarding the hybrid joints of where the aluminium sheets are suffering from tension at the edges of
metal-to-metal configuration, showed the highest average strength of all the TMAZ region (shown with orange circles in the sketch) while the stir
studied joint configurations with approximately 28.3 kN peak load and zone (SZ) is loaded under shear (see also provided load line). Conse
5 mm displacement to fracture. quently, lower provided stiffness than the respective of the adhesive
The load transfer mechanism for the three investigated cases of joints bonded joints as well as lower maximum loading capability of approx
is graphically showed in Fig. 6. For the case of adhesive bonded joint, the imately 6 % (14.5 kN) and displacement to fracture of approximately 10
resin (adhesive) layer is perhaps the major promoter of joint strength in % (4.08 mm) is noticed. Similar loading mechanism is noticed for the
such joints. It is well known that the adhesives work very well under riveted lap joints where the transfer load - comprised of a bearing force
shear where they present high shear strength, but low tensile strength. resulting from the bearing pressure exerted by the rivet shank on the
Marked up in the schematic representation of the single lap joint is the hole surface and a friction force induced by friction between the mating
identical load line for the load transfer of the joint. In the first stages of sheets - is restricted to the riveted area and not to the entire overlap, e.g.
deformation, the entire overlap joint works well under shear. Uniform Skorupa et al. (2014). Additionally, failure mode in metal-to-metal
shear field is evident for low mechanical loading levels and in all over riveted lap joints of AA2024-T3 was characterized by pure rivet shear,
the bonded area that results to slightly higher stiffness of the joint when and failure happened due to complete failure of rivets at the bottom
compared to the respective friction stir welded joint. Nevertheless, due substrate, referred to Pitta et al. (2018).
to eccentricity at the load line, tilting of the joint is noticed with higher Finally, the loading mechanism of the hybrid joint can be seen in the
bottom sketch of Fig. 6. The joining mechanism seems to be a mixture of
the above presented cases for adhesive bonded and friction stir welded,
respectively. Two different synergetic mechanisms were identified for
the hybrid joints: (a) the bonded areas at the edges of the overlap region
marked by dark green color and (b) the welded area (of trapezoid shape
with gray color) at the middle of the overlap region. The bonded areas
contribute to anti-tilting mechanism, as globally the sheets at the
overlap area are not free to deform as previously described. This allows
for higher stiffness of the hybrid joint when compared against the fric
tion stir welded joint. To this end, the loading mechanism is very similar
to the adhesive bonded joint at the low loading levels, where shear
throughout the entire overlap area is evident. With increasing loading
level, tilting appears to some extent and consequently the overlap edges
are loaded in shear and in tension for high tilting degrees. Thus, local
fracture occurs at the overlap edges as well as debonding of the adhesive
layer. Nevertheless, fracture of the joint is restricted due to the presence
of the welded area in the middle of the overlap region that allows for
withstanding higher values of mechanical load and especially higher
displacement to fracture values. With regards to the last parameters, the
hybrid joint allows for almost double (+ 100 %) values of displacement
to fracture Lfr that is evident from the mix-mode loading transfer
mechanism. The first mechanism is evident until debonding of the
overlap area edges takes place where the second mechanism takes over
with the shear of the welded area. The two mechanisms work in serial
Fig. 6. Loading mechanism for the three different investigated cases of (the first one in low loading levels and the second one in high loading
joint specimens. levels) by enhancing the deformation capability of the hybrid joint. In
5
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
addition, the exploitation of the adhesive layer eliminates the welding exposure times (e.g. for 2 h), where the pitting corrosion mechanism is
defects on the friction stir welded specimens, e.g. the “hook defect” that is limited according to the photographs in Fig. 7a; this degradation can be
essentially noticed in Fersini and Pirondi (2007). The same failure attributed to the well-known mechanism of hydrogen embrittlement in
mechanism was noticed in Pitta et al. (2018), where fracture occurred in this material, as investigated by Kamoutsi et al. (2006). More than 50 %
two stages comprised of initial adhesive failure in low loading levels and decrease in elongation at fracture was noticed for higher corrosion
finally rivet shear. Hybrid joints were found to perform better than exposure times, e.g. Fig. 7b, where exfoliation corrosion takes place as
riveted and pure bonded due to higher stiffness. can be seen in the photographs attached to Fig. 7a. On the contrary, the
tensile mechanical properties of (a) yield stress and (b) ultimate tensile
strength were not essentially decreased by corrosion exposure. The
3.2. Corrosion exposure on non-welded material remaining percentage of both mechanical properties was approximately
80 % even after the highest investigated exposure time of 96 h. This
The effect of corrosion exposure time on the mechanical behavior decrease is mainly attributed to the decrease of the effective thickness of
(engineering stress σ – engineering strain ε) of AA2024-T3 base material the specimens that results from the accelerated growth of
is presented in Fig. 7a. Photographs of the surfaces of pre-corroded corrosion-induced surface micro-cracks according to Alexopoulos et al.
specimens are also included in Fig. 7a in order to show the corrosion- (2012). However, a considerable reduction in yield stress is observed
induced degradation mechanism. It is obvious that both axial nominal from 6 to 12 h of corrosion exposure time where a change in corrosion
stress and axial nominal strain of the specimens were negatively affected damage takes place, e.g. from pitting to exfoliation corrosion. This is also
by corrosion exposure, with the highest decrease to be evident for the evident from the photographs in Fig. 7a where it is obvious that for high
axial nominal strain values with increasing exposure time to the EXCO exposure times (≥ 12 h), the pitting density and size tend to increase and
solution. Additionally, the exposure of AA2024-T3 specimens to the two cases of corrosion damage can be noticed: a) formation of
corrosive environment results in surface deterioration - as can be seen in corrosion-induced surface micro-cracks and b) accumulated corrosion
the photographs included to Fig. 7a - due to nucleation of corrosion- products on the large surfaces.
induced surface pits at the first stages of corrosion attack that are
evolving to micro-cracks with corrosion evolution and lead to exfolia
tion for higher exposure times. Essential ductility degradation of 3.3. Corrosion morphology on the macrostructure of the joints
approximately 35 % (refer to Fig. 7b) was noticed even for short
Photographs from the fractured specimens of the three investigated
welded joints can be seen in Fig. 8. On the left row, the respective photos
of the fracture surface (bonded area) of the adhesive bonded joints can
be noticed. As already described, the fracture area is characterized by
the debonded area at the edges of the overlap region and the main area
(in the middle) that was the last area attaining mechanical load before
fracture. On the right row, the friction stir welded specimens can be seen
without any corrosion exposure for the case of the first line. The top (tool
side) and bottom surface of the joint shows that fracture initiated at the
interface of stir zone (SZ) and thermomechanically affected zone
(TMAZ) at the bottom surface and ended within the shoulder area at the
top surface of the joint. This is possible due to the “hook defect” that was
noticed in Fersini and Pirondi (2007) and for similar works. The fracture
mechanism for the hybrid joint is presented in the middle row of Fig. 8.
This joint fails at the interface between TMAZ and heat affected zone
Fig. 8. Images from the pre-corroded surfaces of the investigated lap joints of
Fig. 7. (a) Typical engineering stress - strain curves of AA2024-T3 tensile AA2024-T3 sheets for Adhesive Bonded (left column), Adhesive bonded +
specimens (non-welded) for various exposure times to EXCO solution and (b) FSWed (middle column) and FSWed (right column) cases and for various
normalized corrosion-induced decrease of the mechanical properties. exposure times to the EXCO solution.
6
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
(HAZ), and the fracture mechanism is slightly different than the welded area. The corroded surfaces were heavily exfoliated that essen
respective of the friction stir welded joints. This phenomenon can be tially decreases the effective thickness of the material, as mentioned in
attributed to the better mixing and stirring of the adhesive layer within Alexopoulos and Papanikos (2008). To this end, when the mechanical
the stir zone that eliminates the appearance of the “hook defect”. stress exceeds the local ultimate tensile strength of the corroded
Regarding fracture path in riveted lap joints of AA2024-T3, the fatigue cross-section of the welded sheet, fracture is initiated from the
cracks were found to initiate in one of the end rivet rows on the facing corrosion-induced cracks and consequently the specimen fails at the end
surface of the loaded sheet, e.g. Skorupa et al. (2014), while cracking of HAZ/BM region. This mechanism was also noticed for the same
around the hole forming an eyebrow pattern was occurred in 15 % of the corrosion exposure in the hybrid joint, thus implying that the 48 h
fractures for AA2024-T3, e.g. Stanley et al. (2016). Failure due to exposure is extremely harsh for this thickness (1.6 mm) of the alloy.
bearing and net tension in non-corroded samples of an orbital riveted
steel/aluminium joint was also noticed in Bella et al. (2018). The cracks 3.4. Corrosion exposure on the AB joints
were found to develop from the hole along an orthogonal direction than
the load axis. The mechanical performance of the adhesive bonded specimens is
The corrosion attack after 6 h of exposure can be seen in the second not significantly affected by corrosion exposure as can be seen in Fig. 9.
line of Fig. 8 and for the three different joints, respectively. More uni Nevertheless, a considerable decrease in the displacement to fracture Lfr
form corrosion attack with the form of pits in the whole exposed surface was observed even for the short exposure times, e.g. 2 h, which can be
is evident on the adhesive bonded joints. The corrosion attack on the attributed to the mechanism of hydrogen embrittlement as for the case
friction stir welded joint is more localized at the processed zone (SZ and of the base material. On the contrary, the maximum load Fmax of the
TMAZ) for the short exposure times of 6 h and this was also noticed in joints - that represents the joint strength in this case - remains almost
Squillace et al. (2004). For the case of hybrid joint, corrosion (mainly unaffected by corrosion exposure time, since no considerable decrease
oxides) is evident primarily at the SZ and secondarily at HAZ while was noticed even for the maximum corrosion exposure time of 72 h. This
fracture occurs preferably at the HAZ/TMAZ interface. phenomenon is probably attributed to the prevention of the corrosive
For higher exposure times of 24 h, the corrosion pattern seems to be solution to penetrate the aluminium alloy surface in the bonding area of
the same but more intensive than the respective lower exposure times. the joints where the adhesive layer is present. As was already discussed
Higher pitting formation and exfoliation are noticed in the adhesive in a previous section, the resin (adhesive) layer is the major promoter of
bonded joints at the unprotected area (without adhesive) of the large joint strength in such joints and is highly corrosion resistant; however,
surfaces of the joint. Regarding friction stir welded joints, the corrosion the outer aluminium surfaces are severely corroded for high exposure
products are more intense in the SZ and TMAZ zones while pitting times (≥ 24 h) that leads to a slight mechanical properties degradation
corrosion occurs in the base material (BM) as well. For higher exposure due to the reduction of the effective thickness.
times, fracture follows the TMAZ/HAZ interface as it is possible to be the Fig. 10 presents photographs of the fractured pre-corroded surfaces
weakest ligament of the welded joint. As these regions are more sus of AA2024-T3 adhesive bonded specimens for different exfoliation
ceptible to corrosion attack, e.g. in Rao and Rao (2004), it is evident that corrosion exposure times. The specimen’s top surfaces exposed to the
pitting and subsequent micro-cracking occurs leading to accelerated corrosion solution (left side) as well as the side-surfaces that were
crack growth under quasi-static loading. The same mechanism can be covered (right side) are shown in the figure. The degree of tilting be
noticed for the hybrid joint where the localized corrosion products are tween the touching surfaces of the joints was calculated from the side-
the nucleation sites for crack growth initiation and subsequent fracture. surfaces images and it was used as a deformation measure in this case.
Finally, for the heavily corroded adhesive bonded joints after 48 h of Surface corrosion attack starts with the form of small pits preferably
exposure, higher exfoliation is occurred at the large surfaces of nucleated at the overlap edges for short exposure times, e.g. 2 h in
aluminium sheets. The fracture analysis revealed that corrosion also Fig. 10a. More uniform corrosion was noticed on the top surfaces of the
penetrated the joints at the ends of the overlap area, as was also noticed specimens after 6 h of exposure, while general exfoliation on the
in Mandel and Kruger (2015), thus accelerating debonding of the joint. corroded surfaces can be noticed for higher exposure times. Fracture on
As for the friction stir welded joints after 48 h of exposure, a different these specimens starts from the edges of the bonded area that are loaded
fracture mechanism occurs, where the fracture location is away from the in tension for short exposure times (it was revealed from the deforma
tion of the adhesive layer) with high tilting of the specimens to be
evident, e.g. 6◦ inclination between the joints. Corrosion evolution seems
to not essentially affect the tilting capability of adhesive bonded joints
since the inclination is almost the same even after the highest exposure
time.
The test results of the hybrid joints are presented in Fig. 11, where no
essential decrease of the maximum load Fmax was noticed, even for high
corrosion exposure times. This behavior can be explained by means of
higher corrosion resistance of the adhesive layer since the adhesive is
one of the major promoters of joint strength in the hybrid joints.
Regarding displacement to fracture, a considerable corrosion-induced
decrease was observed even for the short exposure times due to plasti
cization of the adhesive layer, e.g. by Ni et al. (2020) and the decrease of
the bonding area width. However, a considerable increase in the
displacement to fracture was observed after the application of the FSW
process - when compared against the adhesive bonded specimens -
especially for the short exposure times regime (≤ 24 h) - that was ex
pected due to the synergistic effects of the joint mechanisms.
Regarding fractured surfaces of pre-corroded hybrid joints shown in
Fig. 12, it was noticed that the edges of the joint specimens interface
start to gradually separate and fracture takes place on the TMAZ region
Fig. 9. Typical load - displacement curves of AA2024-T3 AB joints pre- of the processed side for the reference specimens, e.g. Fig. 12a. Corrosion
exposured to the EXCO solution for various exposure times. starts preferably at the HAZ and moves on to the interface of SZ/TMAZ
7
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
Fig. 10. Images of the fractured surfaces of pre-corroded AA2024-T3 ABed joints for (a) & (b) 2 h (c) & (d) 6 h, (e) & (f) 24 h and (g) & (h) 48 h exposure to the
EXCO solution.
8
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
Fig. 12. Images of the fractured surfaces of pre-corroded AA2024-T3 AB + FSW joints for (a) & (b) 0 h (c) & (d) 6 h, (e) & (f) 24 h and (g) & (h) 48 h exposure to the
EXCO solution.
The strain nephograms extracted from the DIC analysis are presented
in Fig. 16. The nephograms can present an instinctual evolution of axial
strain on the monitored surface of the overlap and might give an insight
of the surface-crack nucleation sites. It can be observed that the region at
the edge of the overlap exhibited the higher stress concentration and
consequently the highest deformation value for all the investigated cases
of specimens examined through DIC. No significant difference in
deformation distribution between friction stir welded and hybrid spec
imens was noticed for specimens exposed to 24 h corrosion. However,
higher deformation values were observed for friction stir welded spec
imens when compared against the hybrid specimens. Regarding speci
mens pre-corroded for 96 h, it can be noticed that the area with the
lowest deformation values is more extended for friction stir welded
specimens and deformation is more concentrated in the edges of the
overlap while for FS weld bonded specimens the deformation level in
creases even from an area away from the overlap edge. As for the case of
specimens corroded for 24 h, higher deformation values were observed
for friction stir welded specimens with the highest value to reach 2.08
mm in contrary to 0.99 mm for hybrid specimens.
The 3D DIC measurements also allowed to measure the out of plane
Fig. 13. Typical load - displacement curves of AA2024-T3 FSW joints pre-
exposured to the EXCO solution for various exposure times.
bending of the single lap joints (SLJs); the load versus out of plane
displacement were plotted in Fig. 17. It was found that in friction stir
Fig. 14. Fractured surfaces of pre-corroded AA2024-T3 FSWed joints for (a) & (b) 0 h (reference), (c) & (d) 6 h, (e) & (f) 24 h and (g) & (h) 48 h exposure to the
EXCO solution.
9
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
the corrosion evolution does not significantly affect the Fmax since the
remaining percentage of the initial Fmax value is approximately 90 %
even after 24 h of corrosion exposure where a plateau is reached. This
slight Fmax degradation is attributed to the water and NaCl saturation of
the adhesive layer at the early stages of corrosion that prevent further
corrosion penetration to the aluminium surface in the bonding area;
however, pitting corrosion on the outer surfaces of the joints results in
corrosion-induced micro-cracks evolution into the aluminium alloy and
therefore results to a decrease of the “effective” thickness of the speci
mens. Hybrid joint is superior to the other two cases regarding corrosion
resistance up to 48 h while for higher exposure times the remaining
percentage of the initial Fmax value almost reached the values of the
adhesive bonded joints. For instance, after 48 h of corrosion exposure
time the Fmax value of friction stir welded specimens was decreased
almost to half value, while for the FS weld-bonded specimens the
remaining percentage of the initial property is approximately 90 %. It
seems that the synergetic effects of adhesive bonding and friction stir
welding are taking place in the case of hybrid joints by enhancing their
corrosion behavior. A plateau value of the Fmax was reached after 48 h of
Fig. 15. Degree of tilting for the different investigated joints, namely FSWed corrosion exposure and further corrosion exposure does not decrease the
(black line), ABed (blue line) and hybrid (AB + FSW) (red line) for various property significantly for all the investigated cases.
exposure times to the EXCO solution. (For interpretation of the references to
color in this figure legend, the reader is referred to the Web version of
this article.)
Fig. 16. DIC measurements of AA2024-T3 for (a) FSWed + 24 h EXCO, (b) AB + FSWed + 24 h EXCO, (c) FSWed + 96 h EXCO and (d) AB + FSWed + 96 h EXCO.
10
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
Fig. 18. Maximum load Fmax of the different investigated joints for various
exposure times to the EXCO solution.
11
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
4. Conclusions
12
C.-M. Charalampidou et al. Mechanics of Materials 164 (2022) 104122
strength decrease rate in this stage is a result of NaCl saturation of the Gialos, A.A., Zeimpekis, V., Alexopoulos, N.D., Kashaev, N., Riekehr, S., Karanika, A.,
2018. Investigating the impact of sustainability in the production of aeronautical
adhesive layer. Stage III: fast mechanical properties degradation.
subscale components. J. Clean. Prod. 176, 785–799.
Corrosion penetration in the adhesive layer/aluminium alloy inter Jamasri, Ilman, M.N., Soekrisno, R., Triyono, 2011. Corrosion fatigue behavior of
face lead to localized pitting nucleation in the HAZ and subsequent resistance spot welded dissimilar metal welds between carbon steel and austenitic
micro-cracking formation in the bonded area. stainless steel with different thickness. Procedia Eng. 10, 649–654.
Jariyaboon, M., Davenport, A.J., Ambat, R., Connolly, B.J., Williams, S.W., Price, D.A.,
2007. The effect of welding parameters on the corrosion behaviour of friction stir
Declaration of competing interest welded AA2024-T351. Corrosion Sci. 49 (2), 877–909.
Jata, K.V., Sankaran, K.K., Ruschau, J., 2000. Friction-stir welding effects on
microstructure and fatigue of aluminum alloy 7050-T7451. Metall. Mater. Trans. 31,
The authors declare that they have no known competing financial 2181–2192.
interests or personal relationships that could have appeared to influence John, R., Jata, K.V., Sadananda, K., 2003. Residual stress effects on near threshold
the work reported in this paper. fatigue crack growth in friction stir welded aerospace alloys. Int. J. Fatig. 25,
939–948.
Kamoutsi, H., Haidemenopoulos, G., Bontzoglou, V., Pantelakis, S., 2006. Corrosion-
References induced hydrogen embrittlement in aluminium alloy 2024. Corrosion Sci. 48,
1209–1224.
Alexopoulos, N.D., Papanikos, P., 2008. Experimental and theoretical studies of Lenczowski, B., 2002. New lightweight alloys for welded aircraft structure. ICAS 56.
corrosion-induced mechanical properties degradation of aircraft 2024 aluminum Liu, Z., Chong, P.H., Butt, A.N., Skeldon, P., Thompson, G.E., 2005. Corrosion
alloy. Mater. Sci. Eng. 498, 248–257. mechanism of laser-melted AA2014 and AA2024 alloys. Appl. Surf. Sci. 247,
Alexopoulos, N.D., Dalakouras, C.J., Skarvelis, P., Kourkoulis, S.K., 2012. Accelerated 294–299.
corrosion exposure in ultra thin sheets of 2024 aircraft aluminium alloy for GLARE Mandel, M., Kruger, L., 2015. Long-term corrosion studies of a CFRP/EN AW - 660 - T6
applications. Corrosion Sci. 55, 289–300. self-piercing rivet joint and a Steel/ENaw-6060-T6 blind rivet joint. Mater. Today
Alexopoulos, N.D., Gialos, A.A., Zeimpekis, V., Velonaki, Z., Kashaev, N., Riekehr, S., Proceed. 2, 131–140.
Karanika, A., 2016. Laser beam welded structures for a regional aircraft: weight, cost Mishra, R.S., Ma, Z.Y., 2005. Friction stir welding and processing. Mater. Sci. Eng. R 50,
and carbon footprint savings. J. Manuf. Syst. 39, 38–52. 1–78.
Andrea, N.M., Bouali, A., Maawad, E., Staron, P., dos Santos, J.F., Zheludkevich, M.L., Montgomery, J., 2007. Aircraft Primary Structure and Material. National Institute of
Amancio-Filhod, S.T., 2019. Corrosion behavior of metal - composite hybrid joints: Aerospace, Workshop on Revolutionary Aircraft for Quiet Communities, Hampton,
influence of precipitation state and bonding zones. Corrosion Sci. 158, 108075. U.S.A.
ASTM D3933-98, 2017. Standard Guide for Preparation of Aluminum Surfaces for Ni, J., Min, J., Wan, H., Lin, J., Wang, S., Wan, O., 2020. Effect of adhesive type on
Structural Adhesives Bonding (Phosphoric Acid Anodizing). ASTM International, mechanical properties of galvanized steel/SMC adhesive - bonded joints. J. Adhes.
West Conshohocken, PA,. Adhes. 97, 102482.
ASTM D5656-10, 2017. Standard Test Method for Thick-Adherend Metal Lap-Shear Niu, P., Wenya, L., Xiawei, Y., 2017. Global and local constitutive behaviors of friction
Joints for Determination of the Stress-Strain Behavior of Adhesives in Shear by stir welded AA2024joints. J. Mater. Sci. Technol. 33, 987–990.
Tension Loading. ASTM International, West Conshohocken, PA,. Pantelakis, S.G., Setsika, D., Chamos, A., Zervaki, A., 2016. Corrosion damage evolution
ASTM E8/E8M -16ael, 2016. Standard Test Methods for Tension Testing of Metallic of the aircraft aluminum alloy 2024-T3. Int. J. Struct. Integr. 7, 25–46.
Materials. ASTM International, West Conshohocken, PA. Papadopoulos, M., Tavares, S., Pacchione, M., Pantelakis, S., 2013. Mechanical
ASTM G1-03 (2017)e1, 2017. Standard Practice for Preparing, Cleaning and Evaluating behaviour of AA 2024 friction stir overlap welds. Int. J. Struct. Integr. 4, 108–120.
Corrosion Test Specimens. ASTM International, West Conshohocken, PA. Pitta, S., Carles, V., de la, M., Roure, F., Crespo, D., Rojas, J.I., 2018. On the static
ASTM G34-01, 2018. Standard Test Method for Exfoliation Corrosion Susceptibility in 2 strength of aluminium and carbon fibre aircraft lap joint repairs. Compos. Struct.
xxx and 7 xxx Series Aluminum Alloys (EXCO Test). ASTM International, West 201, 276–290.
Conshohocken, PA. Posada, M., Murr, L.E., Niou, C.S., Roberson, D., Little, D., Arrowood, A., Debra, G.,
Bella, G. Di, Borsellino, C., Calabrese, L., Proverbio, E., 2018. Durability of orbital riveted 1997. Exfoliation and related microstructures in 2024 aluminum body skins on aging
steel/aluminium joints in salt spray environment. J. Manuf. Process. 35, 254–260. aircraft. Mater. Char. 38, 259–272.
Braga, D.F., Maciel, R., Bergmann, L., da Silva, L.F., Infante, V., dos Santos, J.F., Proton, V., Alexis, J., Andrieu, E., Delfosse, J., Lafont, M., Blanc, C., 2013.
Moreira, P.M., 2019. Fatigue performance of hybrid overlap friction stir welding and Characterisation and understanding of the corrosion behaviour of the nugget in a
adhesive bonding of an Al-Mg-Cu alloy. Fatig. Fract. Eng. Mater. Struct. 42, 2050 aluminium alloy Friction Stir Welding joint. Corrosion Sci. 73, 130–142.
1262–1270. Rao, S.K., Rao, P.K., 2004. Pitting corrosion of heat-treatable aluminium alloys and
Brent, C., 2007. Welded Joints with Polymer Sealant. U.S. Patent Office no. 7225966 B2. welds: a review. Trans. Indian Inst. Met. 57, 593–610.
Brown, J.M., 2008. The Effects of Sealants and Surface Treatments on the Faying Surface Shultz, E.F., Cole, E.G., Smith, C.B., Zinn, M.R., Ferrier, N.J., Pfefferkorn, F.E., 2010.
of Swept Friction Stir Spot Welds. Department of Aerospace Engineering, College of Effect of compliance and travel angle on friction stir welding with gaps. J. Manuf.
Engineering. Wichita State University. Sci. Eng. 132 (4), 041010 -041019.
Bussu, G., Irving, P., 2003. The role of residual stress and heat affected zone properties on Skorupa, A., Skorupa, M., Machniewicz, T., Korbel, A., 2014. Fatigue crack location and
fatigue crack propagation in friction stir welded 2024-T351. Int. J. Fatig. 25, 77–88. fatigue life for riveted lap joints in aircraft fuselage. Int. J. Fatig. 58, 209–217.
Charalampidou, C., Dietzel, W., Zheludkevich, M., Kourkoulis, S.K., Alexopoulos, N.D., Song, Y., Yang, X., Cui, L., Hou, X., Shen, Z., Xu, Y., 2014. Defect features and mechanical
2021. Corrosion-induced mechanical properties degradation of Al-Cu-Li (2198- properties of friction stir lap welded dissimilar AA2024-AA7075 aluminum alloy
T351) aluminium alloy and the role of side-surface cracks. Corros. Sci. 1831, 109330 sheets. Mater. Des. 55, 9–18.
https://doi.org/10.1016/j.corsci.2021.109330. Squillace, A., De Fenzo, A., Giorleo, G., Bellucci, F., 2004. A comparison between FSW
Chowdhury, S., Chen, D., Bhole, S., Cao, X., Wanjara, P., 2013. Lap shear strength and and TIG welding techniques: modifications of microstructure and pitting corrosion
fatigue behavior of friction stir spot welded dissimilar magnesium-to-aluminum resistance in AA 2024-T3 butt joints. J. Mater. Process. Technol. 152, 97–105.
joints with adhesive. Mater. Sci. Eng. 562, 53–60. Stanley, D., Awerbuch, J., Tan, T., Anasori, B., 2016. Reconstruction of fatigue crack
Fersini, D., Pirondi, A., 2007. Fatigue behaviour of Al2024-T3 friction stir welded lap growth in AA2024-T3 and AA2198-T8 fastened lap joints. Theor. Appl. Fract. Mech.
joints. Eng. Fract. Mech. 74, 468–480. 82, 33–50.
Fersini, D., Pirondi, A., 2008. Analysis of the modelling of fatigue failure of friction stir Sullivan, A., Robson, J.D., 2008. Microstructural properties of friction stir welded and
welded aluminum alloy single lap joints. Eng. Fract. Mech. 75 (3–4), 790–803. post-weld heat-treated 7449 aluminium alloy thick plate. Mater. Sci. Eng. 478,
Gao, Y., Xiong, J., Zhang, S., Zuo, Y., 2006. Aging performance of epoxy adhesive in salt 351–360.
water. Adv. Mater. Res. 11–12, 399–402. Xu, W., Liu, J., 2009. Microstructure and pitting corrosion of friction stir welded joints in
Genevois, C., Deschamps, A., Denquin, A., Doisneau-cottignies, B., 2005. Quantitative 2219-O aluminum alloy thick plate. Corrosion Sci. 51, 2743–2751.
investigation of precipitation and mechanical behaviour for AA2024 friction stir Zhoua, C., Yang, X., Luan, G., 2006. Investigation of microstructures and fatigue
welds. Acta Mater. 53, 2447–2458. properties of friction stir welded Al-Mg alloy. Mater. Chem. Phys. 98, 285–290.
Zucchi, F., Trabanelli, G., Grassi, V., 2001. Pitting and stress corrosion cracking
resistance of friction stir welded AA 5083. Mater. Corros. 52, 853–859.
13