Journal of Manufacturing Processes 64 (2021) 785–792

Contents lists available at ScienceDirect

Journal of Manufacturing Processes

journal homepage: www.elsevier.com/locate/manpro

Comparative evaluation of the effect of the substrate thickness and inherent

process defects on the static and fatigue performance of FSW and
adhesive-bonded overlap-joints in an AA6016 alloy
Enrico Lertora a, *, Davide Campanella b, Marco Pizzorni a, Chiara Mandolfino a, Gianluca Buffa b,
Livan Fratini b
a
Department of Mechanical Engineering, Polytechnic School, University of Genoa, Via All’Opera Pia 15, 16145, Genoa, Italy
b
Department of Chemical, Management, Computer Science, Mechanical Engineering, University of Palermo, Viale delle Scienze, 90128, Palermo, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: This work aims at evaluating the mechanical performance of joints in an AA6016 alloy, manufactured by means
Friction stir welding of two alternative processes: friction stir welding and adhesive bonding. Given its wide use in industrial fields
Adhesive bonding such as automotive and aeronautics, an overlap configuration of joints was selected for the study, and a specific
Aluminum alloy
method based on the overlap length evaluation has been developed to allow comparisons among the cases. Two
substrate thicknesses are considered in order to vary the overall joint stiffness, and the effect of such geometrical
parameter on the mechanical behavior of the joined system is experimentally investigated. The experimental
findings highlight that the stiffness of the joint may emphasize the effect of inherent defects due to process,
severely affecting the mechanical and failure behavior of overlap-configured joints under static- or cyclic-load
conditions.

1. Introduction possible to obtain very similar joints to common spot welds with
completely comparable manufacturing times. Nevertheless, even its
Aluminum alloys have long been at the center of attention of trans­ limitations are similar to those of RSW, being mainly related to the
port design. Thanks to their lightness, these materials allow designers to necessity of proper access from both sides of the joint and, very often,
satisfy design criteria that foresee weight reduction ensuring fuel saving. need of sealing [5].
This synergy also makes it possible to reduce polluting emissions In this scenario, joining technologies for lightweight-structure
allowing environmental protection [1]. manufacturing must be an effective tool to increase joint performance.
However, their metallurgical characteristics, along with consequent This aspect is all the more important in transport applications, in which
distortions due to thermal flow, often make welding difficult to be components are typically subjected to both static and dynamic stresses,
performed via conventional electric-arc technologies [2]. To date, and, therefore, the mechanical behavior of the joints requires an in-
Resistance Spot Welding (RSW) is the predominant process in the depth investigation.
automotive sector for sheet metal joining, thanks to advantages such as Among all the applicable joining technologies, Adhesive Bonding
high operating speeds and ease of automation or robotization. However, (AB) and Friction Stir Welding (FSW) allow manufacturing of
achieving defect-free, structurally sound and reliable welds is not so aluminum-alloy overlap joints in an effective and continuous way. The
easy. The complexity of RSW derives from the need to find a right bal­ former is a highly studied method today thanks to the possibility of
ance between electric current and pressure between the parts to be joining dissimilar materials without bringing about distortions, even
joined, to obtain correct fusion of the weld nugget and limited geometric when the pieces to be joined are small in thickness (such as those
alterations of the surfaces on which the electrodes act [3]. commonly used in the transport field). Furthermore, AB allows sealing
Over time, many studies have been focused on mechanical joining and insulation of the joined parts, with obvious advantages in terms of
methods, such as riveting and clinching [4]. Notably, riveting makes it corrosion resistance of the components. Compared to mechanical

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

https://doi.org/10.1016/j.jmapro.2021.01.043
Received 27 November 2020; Received in revised form 8 January 2021; Accepted 29 January 2021
Available online 23 February 2021
1526-6125/© 2021 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

fasteners or spot welds, AB increases the joint stiffness since it produces 2. Materials and methods
a continuous bond rather than a punctual contact, resulting in more
uniform stress distribution over a wider area, and minimizing the 2.1. Materials
presence of portions on the joint affected by stress concentrations. This
aspect, in principle, remarkably contributes to longer fatigue-life of the In this investigation, an extensive experimental campaign was car­
joints, with respect to those fabricated via RSW or riveting processes. ried out on an AA6016 aluminum-alloy material, usually adopted in
However, several aspects – firstly, surface preparation and environ­ automotive parts construction, as well as for aeronautic or naval parts
mental aging [6–8] – also have to be considered for better estimation of thanks to its high workability, weldability and corrosion resistance. This
the fatigue behavior of the AB joints. Limitations of the bonding process was provided in the form of rolled sheets of dimensions 100 mm × 300
have to be found in the intrinsic toxicity of adhesives and solvents, that mm, having thicknesses of 1.2 or 3 mm. The base material used was a
leads to considerable safety concerns [9]; for this reason, managing the precipitation hardening aluminum alloy containing magnesium and
various phases of the process is not always easy and requires an in-depth silicon as principal alloying elements. In Table 1, the relative chemical
preparation of the personnel involved. composition expressed in weight percentage (wt.%) is reported, and
On the other hand, FSW is a consolidated and widespread technology mechanical properties of the base material, such as yield strength (YS),
for joining aluminum-alloy parts or dissimilar materials. FSW is based ultimate tensile strength (UTS) and Vickers hardness value (HV), are
on friction heating at the facing surfaces of the two parts to be welded, also provided.
resulting in a joint created as a synergy of interface deformations, heat,
and solid-state diffusion. Since welding is carried out at temperatures 2.2. Joining methods
that are lower than the melting point of the material, many problems of
conventional fusion welding techniques are avoided. Indeed, before An ESAB LEGIO 3ST FSW machine (ESAB, Gothenburg, Sweden) was
FSW, many materials were considered as non-weldable or, in case of used to perform welding in conditions of tool position and force control.
dissimilar materials, they could not be combined preserving features Preliminarily experiments were carried out to define the proper
such as good mechanical properties, low distortions, or good surface technological and geometrical parameters of the FSW process. Signifi­
finishing. cant results were obtained using a feed rate of 100 mm/min and 500 rpm
Butt- or overlap-joints are the most common joint configurations in as rotational speed. A force control was applied on the machine, fixing
FSW. In fact, by its very nature, such process lends itself very well to butt the force at 4 kN. The tilt angle of the spindle was set at 2.5◦ , while a
welding, since the tool action is perfect for mixing two materials placed dwell time of 1.5 s was applied to reduce material resistance.
side by side on the same plane. Nevertheless, an overlap configuration is As mentioned above, a notable parameter in the FSW process is
a preferred choice when joint fabrication is possible only operating from represented by the geometry of the tool adopted (Fig.1), the latter being
one of the two sides of the coupling. In this condition, the junction is a chosen on the basis of the thickness of the material to be joined. As
mixed, continuous portion at the interface between the overlapped summarized in Table 2, two different tools were adopted, each specific
sheets. A typical disadvantage of this configuration lies in the presence for the sheet thickness considered. In particular, to perform welding on
of notch effects due to hook defects at the edge of the weld [10]. The the 1.2 mm-thick sheet, the selected pin had conical shape with a 15◦
effect of the process defects on FSW-joint performance was investigated cone angle, 1.2 mm height and 3 mm major diameter, while the shoulder
by Guo et al. [11], who performed fatigue tests on two aluminum series, was 10 mm in diameter. In the other case, in which a sheet 3 mm in
manufacturing both butt- and overlap- FSW joints, in which some de­ thickness was used, a tool having trebled geometrical values was chosen
fects were intentionally inserted. The authors observed that internal accordingly. In this way, the thermal input on the weld was fixed,
defects, such as kissing-bonds, generally reduced the joint fatigue life allowing us to compare the results obtained from the different cases.
more than those on the external surface (such as toe-flash defects due to For the manufacturing of the adhesive-bonded specimens, plates of
rotation of the friction tool). Furthermore, comparing the two configu­ dimensions 25.4 mm × 100 mm were joined. The same thicknesses as
rations, overlap joints resulted in a lower fatigue life than butt-joints1 . those used for FSW were considered.
In recent years, the possible combination of welding and adhesive- The sample bonding surfaces were wiped using an acetone-soaked
bonding has been progressively evaluated, in order to merge the ad­ cloth, subsequently blast with corundum sand to achieve the required
vantages of the two processes, minimizing the respective limitations. surface roughness, and finally wiped again with acetone to remove dust
The most typical combinations concern RSW and structural adhesives, and contaminants. After this operation, the adhesive was applied. A two-
and such hybrid joints have resulted promising to attain higher strength, component thixotropic epoxy adhesive, DP490 produced by 3M™
stiffness and energy absorption, compared to joints manufactured with (Table 3), was used to assemble the joints with single-lap configuration.
the two technologies separately [12,13]. In contrast, the hybrid fabri­ This was prepared by mixing two parts of epoxy resin and one part of
cation of FSW-AB joints is still difficult owing to the weld thermal wave amine-modified curing agent. An adhesive thickness of 0.1 mm was
that affects the adhesive, deteriorating it and compromising the overall adopted. The latter was ensured by using glass spheres finely dispersed
functionality of the joint [14,15]. in the resin during mixing. Since their function is only to control the
To the best of the authors’ knowledge, in no study were FSW and AB resin thickness, only a percentage of 10 % of glass spheres was intro­
systematically compared in order to highlight the conditions that make duced in the adhesive mass, in order not to affect its mechanical char­
adoption of one or the other technology suitable, especially focusing on acteristics significantly. After its application and subsequent
joint geometry. Hence, in this work, the authors propose a comparative overlapping, the adhesive was cured for seven days prior to joint testing.
analysis between FSW and AB technologies, evaluating the static and
dynamic performance of lap joints in an AA6016 aluminum alloy. Sheets 2.3. Static and dynamic testing
of two thicknesses were used, and particular attention was paid to un­
derstanding which of the two processes was preferable on the basis of Both static and dynamic tests were performed using an Instron 8802
their industrial applicability. servo-hydraulic testing machine (Instron, Norwood, MA, USA) equipped
with a 50 kN load cell. The affiliated Bluehill software was used to set
the testing parameters and record the resulting force and displacement
data, along with the number of fatigue cycles at joint failure. All tests
1
In case of butt joints, an increase in fatigue resistance can be more easily were carried out at room temperature (25 ◦ C) and relative humidity of
achieved by acting on the surface finishing of the joint, e.g., via Surface Me­ 40 %.
chanical Rolling Treatment (SMRT) [21] or milling. Tensile testing was performed setting a 1.3-mm/min speed of the

786
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

Table 1
Mechanical properties and chemical composition of the AA6016 base material used.
YS (MPa) UTS (MPa) HV

135 240 75
Chemical composition (wt.%)
Si Cu Fe Mn Mg Cr Zn Ti Other Al
1.2 0.52 0.5 0.2 0.4 0.1 0.2 0.15 0.15 Balance

Fig. 1. FSW tool and detail of the cross-section. The geometrical parameters indicated here are those modified in accordance with the joint thickness.

required a preliminary investigation on the joint overlap in order to

Table 2
determine the optimal conditions under which to compare the two
Geometrical and technological parameters set for the FSW process.
manufacturing technologies.
Parameter Laminate thickness (mm) 1.2 3.0

Shoulder diameter (mm) 10 30 3. Results and discussion

Geometrical Pin height (mm) 1.6 4
Pin diameter (mm) 3 9 3.1. Preliminary static characterization and determination of the overlap
Rotational speed (rpm) 500
Feed rate (mm/min) 100
for the AB joints
Technological
Tilt angle (deg) 2.5
Force (N) 4000 To compare the two processes, five identical FSW specimens (N = 5),
25 mm in width each, were first cut from the original welded plates, and
then tested for tensile strength (Fig. 2a). It not being possible to uniquely
Table 3 determine the effective resistant area of the FSW joint, a normalization
Technical specifications of the 3M™ DP490 adhesive. of the load values obtained from tensile testing was carried out with
Characteristic Base Accelerator
respect to the specimen width (Y = 25 mm), that was the only known
dimension of the joining area.
Epoxy Amine-modified
At first, the 1.2-mm-thickness case was considered for both the FSW
Color Black White
Mix ratio 2 1 and AB joints, being the most critical thickness condition in terms of
Work life 1.5 h @ 23 ◦ C deformation of the substrate during testing. The resulting average value
Handling time 4÷6 h @ 23 ◦ C of Fmax /Y (maximum force per sheet overlap) of the FSW joints was 198
Curing condition 7 days @ 23 ◦ C
N/mm. This value was considered as a benchmark for the following
sizing of the adhesively-bonded joints (Fig. 2b). Indeed, maximum shear
cross-head on the basis of the instructions provided in the EN strength is expressed as the ratio between Fmax and the geometrical
ISO1465standard [16]. The latter was taken as the reference for all the contact area (A = XY). Therefore, the value of width Y was again set
geometrical parameters, except for the overlap length, that was varied to equal to 25 mm, and three sets of AB joints were made by varying the
allow comparisons between the two joining processes investigated. overlap length X among as many values (namely, 5 mm, 10 mm and 15
To assess the fatigue behavior of the FSW specimens, the relative EN mm), in order to identify which value of X provided static-strength
ISO 9664 standard [17] was followed. Fatigue testing was carried out values corresponding to those of the 1.2 mm-thick FSW joints in ten­
under sinusoidal, axial tensile, constant amplitude loading, and stress sile conditions. Also in this case, for each overlap value, five repetitions
ratio was set to R = 0.1. At least four repetitions (N = 4) were performed were performed (N = 5).
for each joint configuration and stress level. The details of load values From the static testing of the 1.2 mm-AB joints (Table 4), the mean
and test conditions will be reported in Section 3, since their definition values of Fmax /Y ratio (and corresponding shear strength) were

787
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

Fig. 2. Geometrical configuration of (a) FSW overlap joint and (b) adhesive bonded joint.

to manufacture AB joints having the same value of Fmax /Y as that of

Table 4
FSW. Nevertheless, such a possibility was discarded and an adhesive
Results of the tensile testing for different substrate thicknesses and variable
overlap of 13 mm was reasonably believed suitable also in this case.
overlap values of the AB joints.
Such a decision was made on the basis of different considerations. From
FSW AB
an experimental standpoint, the use of a different value of overlap would
Thickness (mm)
Fmax /Y (N/mm) Fmax /Y(N/mm) Overlap length, X (mm) introduce a further variable into the system, requiring an increase of the
overlap length (X) of about 41 % to provide an increase of only 29 % in
112.4 ± 6.4 5
1.2 198.0 ± 4.7 175.2 ± 6.1 10 Fmax /Y, with respect to the configuration having X = 13 mm. It is also
211.5 ± 5.8 15 worth noting that, due to the substrate thickness, an increase in the
3 475.0 ± 6.5
337.6 ± 9.8 13 overlap would not, however, be necessary to further increase stiffness of
546.2 ± 13.0 26
the AB joints, since the latter did not exhibit observable deformations
during tensile testing. Finally, from the manufacturing point of view, the
acquired. adoption of X = 22 mm would require an almost double quantity of
As illustrated in Fig. 3a, the trend of the resulting Fmax /Y values resin, affecting any cost-effectiveness evaluations related to adhesive
exhibited by the AB joints was in a direct quasi-linear relationship with bonding.
the overlap increment; hence, it was believed reasonably valid to Based on such considerations, in the subsequent dynamic compari­
determine the proper overlap for the joints via linear interpolation son between the two processes, AB joints were fabricated varying the
among the results. In particular, the value of Fmax /Y sought being be­ thickness of the substrate between 1.2 mm and 3.0 mm, keeping both the
tween 175.2 MPa and 211.5 MPa, the interpolation was performed adhesive thickness (0.1 mm) and overlap length (13 mm) constant.
within the corresponding values of X, namely, 10 mm and 15 mm
respectively. Based on this calculation, the overlap length X corre­ 3.2. Evaluation of the dynamic behavior of the joints
sponding to Fmax /Y = 198 N/mm was 13 mm, and this value was
employed for the fabrication of the AB joints. Based on the tensile results, fatigue testing (FT) was performed
The same experiment was then performed adopting substrates 3.0 adopting two different loading conditions in relation to the substrate
mm in thickness (Fig. 3b). The FSW joints exhibited an average value of thickness adopted. The related four, stress levels are listed in Table 5,
Fmax /Y ratio equal to 475 N/mm. The latter value being remarkably diversified according to the thickness of the sheet metal; the resulting
higher than that of the previous case (due to the increased size of the experimental Wӧhler curves are displayed in Fig. 4.
FSW tool used), the overlap values evaluated and tested here were 13 The first dynamic tests involved the joints made from 1.2-mm-thick
mm and 26 mm, which corresponded to Fmax /Y equal to 337.6 N/mm laminates. The related results are summarized in Table 6.
and 546.2 N/mm, respectively. As in the former case, a linear rela­ In case of FSW joints, regardless of the load level applied and
tionship was assumed between the values. From linear interpolation, it consistently in all joints, fractures grew from the unwelded tips on the
was calculated that an overlap of 22 mm approximatively should be used weld sides at the boundary between the thermo-mechanically altered

Fig. 3. Graphical results of the tensile testing and determination of the overlap length for substrates with thickness of (a) 1.2 mm and (b) 3 mm.

788
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

Table 5 Overall, comparing the behavior of the two joint configurations, it is

Load levels used for the fatigue-life comparisons. to be noted that the 3-mm AB joints exhibited approximately double
Thickness (mm) Load level Fmin (N) Fmax (N) Fm (N) Fa (N) fatigue durability at the higher load levels (01 and 02) with respect to
those made via FSW, and resulted at least one order of magnitude more
01 661 1983 1322 661
durable at the lowest load condition (04). It should be noted that the
02 925 1719 1322 397
1.2
03 1057 1587 1322 265 latter result was achieved without any visible sign of damage of the AB
04 1124 1520 1322 198 joints. Instead, wherever it occurred, failure of the AB joints was mixed
01 1477 4431 2954 1477 in type, consisting of prevalently-cohesive regions after a low number of
02 1772.4 4135.6 2954 1181.6 cycles N (corresponding to the higher fatigue-load level), and progres­
3
03 1920.1 3987 2954 1033.9
04 2067 3840 2954 886
sively presenting more diffuse de-adhesive failures when N increased (as
an effect of the load reduction). Such behavior was not unexpected,
since related to evaluations about the critical strain energy release rate,
zone (TMAZ) and weld nugget (i.e., dynamically recrystallized zone – which is typically lower in Mode-II than in Mode-I [18]. In this regard,
DRZ), notably, nucleating from hook defects generated by stirring the reader is invited to refer to Fig. 5, which illustrates a typical sche­
(mainly on the retreating side). In contrast, the AB joints exhibited matization of a bonded joint subjected to a generic load P, that translates
opposite behavior depending on the fatigue-load condition. Specifically, to different stress components acting on the bonded area. It follows that,
substrate failures (SF) occurred in the former condition (01), that was in case of high-load levels applied, a simultaneous action of shear (τxz )
characterized by plastic deformation of the aluminum substrate. Cracks and cleavage (σz ) stresses occurred, thus, causing the internal defects to
nucleated externally with respect to the bonded area and propagated
transversally, entirely affecting the base material. Hence, the rupture Table 6
was completely due to ductility of the substrate (obviously emphasized Comparison between the fatigue-life (expressed as number of cycles) of FSW and
by the slender thickness), which led to results related to the substrate AB joints having thickness of 1.2 mm.
resistance rather than to that of the adhesive joint. Average number of cycles N
Conversely, when a lower load level (02) was applied, the AB sam­ Thickness (mm) Load level
ples achieved the limit value set at 106 cycles without showing any sign FSW AB
of crack. Such behavior led us to consider of little interest the results 01 2.52 × 104 Substrate failure
obtained in this case, which were therefore not introduced in the Wӧhler 1.2
02 1.44 × 105 > 106
03 5.56 × 105 Not tested
diagram. Nevertheless, the fact that the adhesive area was still intact
04 > 106 Not tested
after every test has to be interpreted as proof of better performance
exhibited by adhesive bonding compared to FSW. Indeed, it should be
noted that each stress condition was remarkably lower than the
maximum one that the DP490 adhesive alone is able to bear before Table 7
Comparison between the fatigue-life (expressed as number of cycles) of FSW and
failing cohesively. In other terms, this led us to conclude that, in cases
AB joints having thickness of 3 mm.
requiring joining of small-thickness sheets, adhesive methods are, in
principle, more preferable than FSW. Average number of cycles N
Thickness (mm) Load level
The second series of dynamic tests was carried out varying the sheet FSW AB
thickness to 3.0 mm. The related results are presented in Table 7. As
01 2.25 × 104 5.60 × 104
previously stated, geometrical parameters such as overlap length X and 02 4.97 × 104 1.36 × 105
3
adhesive thickness were kept constant, again setting their values at 13 03 7.08 × 104 4.32 × 105
mm and 0.1 mm, rispectively. 04 1.04 × 105 > 106

Fig. 4. Comparison of Wӧhler curves of FSW and bonded joints, varying sheet in thickness.

789
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

Fig. 5. Schematization of the possible stress components acting on the bonded area during testing.

grow prevalently within the adhesive layer (Fig. 6a). Matching the interface between TMAZ and DRZ, precisely where the hook defects
fracture surface observation with that of the configuration assumed by were present between the two overlapped sheets (Fig. 7). It should be
the joint during testing, a fracture propagation similar to that described noted that hook defects are recognized as typical imperfections of
by Zhou et al. [19] was observed, with cracks that initiated from both overlap FSW of aluminum alloys, difficult to control or minimize during
edges, growing towards the center of the bonded area up to complete the process [20]. During FT, the loading path is not collinear with the
detachment of the two substrates. Vice versa, a dominant Mode-I (i.e., two substrates, and this condition generates an eccentricity that causes
dominant τxz ) caused the joint to fail prevalently at the resin-substrate the specimen edges to rotate around the weld nugget. It follows that the
interface (Fig. 6b-c). In this case, micro cracks randomly distributed non-stirred overlap regions might become the location of stress con­
within the adhesive layer were also observed. Such behavior was related centrations that are detrimental for the overall fatigue endurance of the
to presence of the glass micro-spheres dispersed in the resin and used as FSW joint. The dynamic testing highlighted how such defects are pre­
thickness control, which might have triggered crack initiations within sent, in principle, independently of the thickness of the base material.
the bonded area. However, as the experiments confirmed, it is equally true that an
Concerning the FSW joints, fractures were again localized at the increased thickness emphasizes their effect, as the joint results in higher

Fig. 6. Crack propagation mechanism, and failure surfaces and related 20×-magnified details of 3-mm AB joints tested for fatigue at (a) 01, (b) 02 and (c) 03
load levels.

790
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

Fig. 7. (a) Macrography of the 3-mm FSW-joint cross-section and (b) typical rupture observed after dynamic testing.

stiffness, which remarkably limits its deformation. Such a constraint area is still a preferred way to reduce the notch-effect at the edges of the
condition determines higher stress concentration localized where the stirring zone, minimizing any stress concentrations and providing for a
weld nugget continuum encounters a solution of continuity (the hook continuum between the overlapped surfaces. In turn, FSW may be
defect). Therefore, nucleating from this position, cracks rapidly grew considered effective in ensuring long-lasting properties to the joints,
within the base material up to the outer surface of the joint, leading to being less sensitive to environmental aging than adhesive bonding. The
the remarkable observed reduction in fatigue endurance of the latter topic is now object of investigation by the authors.
specimen.
Declaration
4. Conclusions
This research did not receive any specific grant from funding
A comparison between Friction Stir Welding (FSW) and Adhesive agencies in the public, commercial, or not-for-profit sectors.
Bonding (AB) was proposed by considering their respective application
to an AA6016 aluminum alloy. An overlap-joining configuration was
adopted and the influence of two different substrate thicknesses was Declaration of Competing Interest
investigated. From a preliminary evaluation of the static behavior of the
joints, it was observed that a comparison of the two processes was The authors declare that they have no known competing financial
possible, for both thin (1.2 mm) and thick (3.0 mm) substrates, adopting interests or personal relationships that could have appeared to influence
an overlap length of 13 mm, i.e., almost equal to that defined by the EN the work reported in this paper.
ISO 1465 standard for the realization of single-lap adhesive joints (12.5
mm). Such value was thus adopted to manufacture the joints for the References
fatigue tests.
The differences observed between the static or dynamic behavior of [1] Kelly JC, Sullivan JL, Burnham A, Elgowainy A. Impacts of vehicle weight
reduction via material substitution on life-cycle greenhouse gas emissions. Environ
FSW and AB joints are significant and mainly related to the inherent Sci Technol 2015;49:12535–42. https://doi.org/10.1021/acs.est.5b03192.
defects of FSW process that have a severe influence on the dynamic of [2] Çam G, İpekoğlu G. Recent developments in joining of aluminum alloys. Int J Adv
the final joint. Specifically, regardless of the thickness adopted, FSW Manuf Technol 2017;91:1851–66. https://doi.org/10.1007/s00170-016-9861-0.
[3] Pouranvari M, Marashi SPH. Critical review of automotive steels spot welding:
joints always performed worse due to significant notch effects caused by process, structure and properties. Sci Technol Weld Join 2013;18:361–403.
hook defects present at portions of interface between the substrates that https://doi.org/10.1179/1362171813Y.0000000120.
were not involved in stirring of the weld nugget. It was noted that joints [4] Li D, Chrysanthou A, Patel I, Williams G. Self-piercing riveting-a review. Int J Adv
Manuf Technol 2017;92:1777–824. https://doi.org/10.1007/s00170-017-0156-x.
having higher thickness are more sensitive to this effect, due to [5] Barnes TA, Pashby IR. Joining techniques for aluminum spaceframes used in
increased stiffness, providing more limited endurance under cyclic load automobiles. Part II - adhesive bonding and mechanical fasteners. J Mater Process
conditions than thinner joints. Technol 2000;99:72–9. https://doi.org/10.1016/S0924-0136(99)00361-1.
[6] Boutar Y, Naïmi S, Mezlini S, Carbas RJC, da Silva LFM, Ben Sik Ali M. Fatigue
Also the AB exhibited different behavior depending on the stiffness of
resistance of an aluminium one-component polyurethane adhesive joint for the
the substrates. Notably, when thinner sheets were used, substrate fail­ automotive industry: effect of surface roughness and adhesive thickness. Int J
ures were obtained at the higher load-level tested, whereas the end-limit Adhes Adhes 2018;83:143–52. https://doi.org/10.1016/j.ijadhadh.2018.02.012.
[7] Boutar Y, Naïmi S, Mezlini S, Ali MBS. Effect of surface treatment on the shear
of the test (i.e., one million cycles) was achieved without failure at the
strength of aluminium adhesive single-lap joints for automotive applications. Int J
load-level immediately lower. Such behavior was interpreted as good Adhes Adhes 2016;67:38–43. https://doi.org/10.1016/j.ijadhadh.2015.12.023.
performance of the AB joint, the contact area being intact in both cases. [8] Costa M, Viana G, da Silva LFM, RDSG Campilho. Effect of humidity on the
Contrariwise, with greater stiffness of the substrates, the adhesive mechanical properties of adhesively bonded aluminium joints. Proc Inst Mech Eng
Part L J Mater Des Appl 2018;232:733–42. https://doi.org/10.1177/
become the weak part of the joint, showing failures that were preva­ 1464420716645263.
lently cohesive or prevalently located at the adhesive-adherend inter­ [9] Adderley CS. Adhesive bonding. Mater Des 1988;9:287–93. https://doi.org/
face, depending on the loading mode. 10.1016/0261-3069(88)90006-4.
[10] Buffa G, Campanile G, Fratini L, Prisco A. Friction stir welding of lap joints:
In view of such comparison, the authors intended this work as a influence of process parameters on the metallurgical and mechanical properties.
necessary benchmark for further developments aimed at combining the Mater Sci Eng A 2009;519:19–26. https://doi.org/10.1016/j.msea.2009.04.046.
two technologies. Indeed, the introduction of adhesives at the joining [11] Guo S, Shah L, Ranjan R, Walbridge S, Gerlich A. Effect of quality control
parameter variations on the fatigue performance of aluminum friction stir welded

791
E. Lertora et al. Journal of Manufacturing Processes 64 (2021) 785–792

joints. Int J Fatigue 2019;118:150–61. https://doi.org/10.1016/j. [17] EN ISO 9664. Test methods for fatigue properties of structural adhesives in Tensile
ijfatigue.2018.09.004. Shear. 1995.
[12] Braga DFO, De Sousa LMC, Infante V, Da Silva LFM. Moreira PMGP. Aluminium [18] Dessureault M, Spelt JK. Observations of fatigue crack initiation and propagation
friction-stir weld-bonded joints. J Adhes 2016;92:665–78. https://doi.org/ in an epoxy adhesive. Int J Adhes Adhes 1997;17:183–95.
10.1080/00218464.2015.1085860. [19] Zhou H, Liu HY, Zhou H, Zhang Y, Gao X, Mai YW. On adhesive properties of nano-
[13] Pizzorni M, Lertora E, Mandolfino C, Gambaro C. Experimental investigation of the silica/epoxy bonded single-lap joints. Mater Des 2016;95:212–8. https://doi.org/
static and fatigue behavior of hybrid ductile adhesive-RSWelded joints in a DP 10.1016/j.matdes.2016.01.055.
1000 steel. Int J Adhes Adhes 2019;95. https://doi.org/10.1016/j. [20] Braga DFO, Maciel R, Bergmann L, da Silva LFM, Infante V, dos Santos JF, et al.
ijadhadh.2019.102400. Fatigue performance of hybrid overlap friction stir welding and adhesive bonding
[14] Petrie EM. Weldbonding-a hybrid method of assembly. Met Finish 2013;111:42–4. of an Al-Mg-Cu alloy. Fatigue Fract Eng Mater Struct 2019;42:1262–70. https://
https://doi.org/10.1016/S0026-0576(13)70163-6. doi.org/10.1111/ffe.12933.
[15] Lertora E, Mandolfino C, Pizzorni M, Gambaro C. Influence of adhesive in FSW: [21] Dong P, Liu Z, Zhai X, Yan Z, Wang W, Liaw PK. Incredible improvement in fatigue
Investigation on fatigue behavior of welded, weld-bonded, and adhesive-bonded resistance of friction stir welded 7075-T651 aluminum alloy via surface
joints in aluminum AA 6082 T6. Materials (Basel) 2019;12. https://doi.org/ mechanical rolling treatment. Int J Fatigue 2019;124:15–25. https://doi.org/
10.3390/ma12081242. 10.1016/j.ijfatigue.2019.02.023.
[16] EN ISO 1465:2009-07. Adhesives - Determination of tensile lap-shear strength of
bonded assemblies n.d. doi:10.7208/chicago/9780226547404.001.0001.

792