International Journal of Fatigue 174 (2023) 107712

Contents lists available at ScienceDirect

International Journal of Fatigue

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

On improved fatigue properties of aluminum alloy 5086 weld joints

Rajneesh Jaisawal a, Vidit Gaur a, *, Shahnawaz Ahmed b
a
Department of Mechanical and Industrial Engineering, Indian Institute of Technology Roorkee, Uttarakhand 247667, India
b
Defence Metallurgical Research Laboratory, DRDO, Hyderabad 500058, India

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

Keywords: In this work, we aimed to investigate the fatigue properties of AA5086-H321 and its weld joints. The AA5086-
Fatigue H321 plates were welded in a butt configuration using the friction stir welding. Load-controlled axial sinusoidal
Welding cyclic tests were conducted in the ambient conditions, at a frequency 20 Hz, and at different load ratios i.e., Rσ =
Mean-Stress
− 1, 0.1, and 0.5. The weld joints, in general, are known to have reduced properties. However, in this study, the
Intergranular
Aluminum
base alloy and its weld joint had similar hardness and ultimate tensile strength, but with marginally reduced
elongation (∼ 4.4%). The endurance limits of the weld joints were either higher (∼ 28% and ∼ 11% at R = − 1
and 0.1, respectively) or same (at R = 0.5) as compared to that of the base alloy. The difference in fatigue lives of
base alloy and its weld joint kept diminishing with increasing R-ratio. Microscopic examination of fracture
surfaces in weld joints revealed three types of crack initiations: voids, inclusions, and tool-marks, and two types
of crack propagation mechanisms: intergranular and transgranular. However, in base alloys, only defect-free
surface initiation with transgranular crack growth was observed. Mean stress, as expected, had a detrimental
effect on fatigue properties, with no definitive control on the type of crack initiation or its propagation mech­
anism. An increase in the fatigue lives of weld joints was attributed to the crack growth retardation during the
intergranular fatigue crack propagation due to frequent defection of crack path, which was absent in the base
alloys. Most of the intergranular facets were observed near the surfaces (top or bottom) of the weld nugget zone
and disappeared when the crack front reaches its interface boundary with the heat affected zone. All these as­
pects of fatigue damage in microstructurally heterogeneous weld joints and the plausible explanations for the
observed mechanical fatigue properties are presented in this study.

1. Introduction applications.
In general, fatigue lives and endurance limits of weld joints, irre­
Aluminum 5086-H321 is a non-heat treatable alloy with potential spective of the welding technique, are lower as compared to the base
applications in marine, automobile, and aerospace industries. These alloys. This detrimental effect has been attributed to the associated weld
alloys have high specific strength, corrosion resistance, and good defects like voids, inclusions, and micro-profiled geometrical irregular­
weldability. Such alloys, acquire their strength from solid solution ities [4–6]. The location and size of these defects can also affect the
strengthening mechanisms, due to Mg-atoms, and strain hardening, due fatigue lives of a component and often serves as the potential crack
to rolling operation. To fabricate large structures like hull, deck, etc. in initiation sites [7]. However, such defects are minimal in the friction stir
shipbuilding industries, the joining of these alloys plays an important welding process and thus, their joints have better properties as
role in their design. Conventional fusion-based welding techniques are compared to the conventional ones [8]. Several studies have been per­
generally used for the purpose but have low weld efficiency and are formed on the investigation of mechanical and fatigue behavior in
often associated with various defects & poor joint efficiencies [1], which friction stir weld joints of 5xxx series Al alloys [9–11]. Wang et al. [12]
can degrade the fatigue properties of the weld joints. In this research studied the tensile and fatigue properties of friction stir weld joints of
work, we aimed to explore the suitability of the eco-friendly solid-state AA5083-H19 alloy and reported a maximum joint efficiency of 95 %
Friction Stir Welding (FSW) [2] technique, which is already proven to with a slight reduction in the fatigue lives, with defect-free surface crack
have reduced defects and better joint efficiency than the conventional initiation, but with same endurance limit. Dickerson et. al [13] reported
ones [3], specifically for AA5086-H321 alloy and its structural a loss in tensile and fatigue performance as the cracks initiated from the

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

https://doi.org/10.1016/j.ijfatigue.2023.107712
Received 30 March 2023; Received in revised form 24 April 2023; Accepted 7 May 2023
Available online 21 May 2023
0142-1123/© 2023 Elsevier Ltd. All rights reserved.
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

root flaws present in the friction stir welded AA5083-O and AA5083- Samples for microstructure and property characterization were then
H321 alloys. The effect of these flaws up to certain size remained un­ prepared from these plates, as discussed next.
identifiable. James et al. [14] investigated the fatigue response of
AA5083-H321 alloy in the as-weld (using FSW), polished conditions and 2.2. Characterization of microstructure
observed that most of the cracks initiated from the surface micro-
notches generated by the tool along with a few evidences of voids Cubical samples of 10 mm size were extracted from the different
linking on the surface in as weld conditions. Few other studies are also zones of the welded plates i.e. Base Metal (BM), Nugget Zone (NZ),
available in literature on different aluminum alloys but many of them Thermo-Mechanically Affected Zone (TMAZ), and Heat-Affected Zone
dealt with conventional welding techniques [15]. However, the data (HAZ). The surfaces of these specimens were polished mechanically by
available in the literature related to friction stir welding of AA5086- successive grades of emery paper (ranging from P400 to P3000), fol­
H321 is very limited and those available, did not explore their fatigue lowed by colloidal silica polishing (~ 0.02 μm) and cleaning in
properties much. Thus, available database cannot provide a better un­ ultrasonic-bath of alcohol for 2 mins. The test samples were then electro-
derstanding of correlation between the fatigue crack initiation incurring polished in an electrolyte of methanol, and perchloric acid in the ratio of
from the welding flaws and the fracture location in different zones of 1:4. The electropolishing was performed at 12 V for 30 s and within the
weld joints. This research work is an attempt to contribute to the liter­ temperature range of − 1 0C and − 5 0C. The measurement of grain size
ature gap. was done in line with the procedures stated in ASTM E112-13 [16]. The
In this work, we aimed to examine the mechanical properties and texture properties of the developed joint were characterized with the
fatigue lives of AA5086-H321 and its FSW joints. The plates of Electron Back-Scattered Diffraction (EBSD) technique, performed using
aluminum 5086-H321 alloy were welded in a butt configuration. The a scanning electron microscope (SEM). A scanning area of 200 μm × 200
test specimens of dog-bone shaped were then cut along the weld’s μm was selected for the EBSD analysis and the raw data was post-
transverse direction. The monotonic tensile and load-controlled axial processed using a MATLABTM based open-source MTEX toolbox [17].
cyclic tests were conducted on these specimens at different load ratios
( − 1, 0.1, and 0.5) in ambient conditions and using a 20 Hz sinusoidal 2.3. Characterization of mechanical properties
waveform. Unlike our anticipation, an improvement in fatigue proper­
ties of AA5086-H321 weld joints was observed, irrespective of the Different mechanical properties of the weld joints like hardness,
applied mean stress (R-ratio). The observations have been explained and tensile, fatigue etc. were characterized. Vickers hardness (HV0.1) was
argued using the various aspects of fatigue crack initiation and crack measured on cubical specimens using Mitutoyo HM200 micro-hardness
growth damage mechanisms, specifically in the weld joints, and are machine as per the ASTM E384-17 [18] under ambient laboratory air
discussed in this manuscript. conditions. For each test, a force equivalent to 100 g was applied for a
dwell period of 10 s on a suitably polished surface of the specimen.
2. Procedures Several equally-spaced measurements were carried out along the depth
and across the weld cross-section, 1 mm and 0.5 mm respectively, to
2.1. Welding generate a complete hardness map of the weld joint.
To characterize the monotonic tensile properties, standard sub-sized
In this study, 5 mm thick sheets of aluminum 5086-H321 having the flat dog-bone shaped specimens as per ASTM E8/E8M [19] (Fig. 1c)
chemical composition as shown in Table 1 were used. From these larger were used. The test samples were cut transverse to the rolling direction
sheets, small plate samples were then cut using the mechanical band- using Wire-cut EDM while ensuring the presence of weld joint (WJ)
saw and shaper machine, of size 100 mm × 50 mm × 5 mm, followed within the gage length. The samples were then cleaned using ethanol to
by mechanical polishing of surfaces using emery paper (P400 grade) to discard any oil or metal residues and were tested in ambient conditions.
remove any irregularity along the edges and degreasing using an acetone Strain-controlled monotonic loading tests were conducted on a 300 kN
solution. All the plate samples were welded at an ambient temperature capacity tensile testing machine, using a 12 mm gage length exten­
(24 ± 2◦ C) and at a relative humidity level of ~ 50 %. The FSW was someter. A uniform strain rate of 0.1 %/sec was applied and a set of ten
performed on a modified vertical milling machine, of 15 HP capacity, (10) nos. of tests were done for a particular test condition.
with sample plates placed in butt configuration and along the rolling Similar specimens, as shown in Fig. 1d, were used to investigate the
direction (see Fig. 1a). A cylindrical tool with conical-threaded pin (see cyclic response of the material and their weld joints. Load-controlled
Fig. 1b), made of X40CrMoV5 tool steel, was used. The geometrical uniaxial cyclic tests were performed using the MTS servo-hydraulic fa­
dimensions of the threaded tool used to perform FSW are shown in tigue testing machine, as per the ASTM E466-07 [20]. A sinusoidal-type
Fig. 1b. Several attempts were made to optimize the weld process and waveform was used to simulate the cyclic loadings. The tests were
the optimized parameters of the FSW process used in this study were: a performed at a frequency of 20 Hz and at different load-ratios/stress-
tool transverse speed of 90 mm/min while rotating at a rotation speed of ratios (R = σmin /σmax = -1, 0.1, and 0.5). Three (03) tests for each
664 rpm and maintaining a tilt angle of 1.50. The term ‘optimized’ with loading condition were repeated and the tests were continued till 107
respect to FSW process represents a friction stir weld joint with the cycles, denoted as ran-out, or till final fracture, whichever reached first.
highest joint efficiency in terms of ultimate tensile strength, defined as For a particular test condition, the endurance limit was designated as the
follows: load value at which the specimen sustained 107 load cycles. The post-
fracture analysis was then performed for all the broken specimens
σut(WJ)
Joint efficiency (%), η = × 100 were to investigate the associated damage mechanisms.
σ ut(BM)
After welding, the top surface of welded plates was machined, ~
0.15 mm, with the help of a shaper machine to discard any splashes from
welding, but the shoulder/tool marks region remained unaffected.

Table 1
Chemical composition (% by wt.) of as received AA5086-H321 alloy.
Element Mg Mn Si Zn Cr Fe Cu Ti Al

Wt. (%) 4.30 0.51 0.18 0.15 0.08 0.08 0.03 0.01 Remaining

2
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 1. A schematic representation of (a) welding setup, along with geometrical dimension of (b) weld tool, (c) tensile test specimens, and (d) fatigue test specimens.

3. Results average size of 43.78 ± 14.81μm, were observed in the base metal (BM).
The grains were highly elongated in the rolled direction (~40 – 120 μm)
3.1. Microstructure as compared to the transverse direction (~20 μm). The microstructure in
HAZ was relatively coarsened owing to the temperature rise (400 ◦ C to
A microstructural inhomogeneity in the grains’ morphology was 480 ◦ C [2]) resulting from the tool-work piece interaction. The average
observed among the different zones in the weld joint i.e. BM, NZ, TMAZ, grain size in the HAZ was 28.18 ± 15.48μm. The TMAZ, with an average
and HAZ. The coarse, elongated, and pancake shaped grains, with an grain size of 22.65 ± 18.94μm, revealed partially deformed and

Fig. 2. (a) A schematic depicting the microstructurally-different zones in a FSW joint, along with the typical microstructures associated with (b) BM, (c) NZ, (d)
TMAZ, and (e) HAZ. BM: Base Metal, NZ: Nugget Zone, HAZ: Heat Affected Zone, TMAZ: Thermo-Mechanically Affected Zone, FSW: Friction Stir Welded, AS:
Advancing Side, RS: Retreating Side, ND: Normal Direction, TD: Transverse Direction, WD: Welding Direction, RD: Rolling Direction.

3
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

elongated grains owing to the tool shearing effect, while the weld nugget Table 2
zone (NZ) comprised of fine equiaxed grains, average size of Micro-hardness values for the different zones welded AA5086-H321 joints. AS:
15.36 ± 7.32μm, due to dynamic recrystallization resulting from the Advancing Side, RS: Retreating Side, BM: Base Metal, NZ: Nugget Zone, HAZ:
severe plastic deformation process of FSW [3]. A schematic of weld joint Heat Affected Zone, TMAZ: Thermo-Mechanically Affected Zone.
depicting all these microstructurally different zones along with their Zone HAZ-AS TMAZ- NZ TMAZ- HAZ-RS BM
characteristic microstructural features is shown in Fig. 2. AS RS

Average 80.68 81.58 ± 83.14 81.98 ± 81.17 87.10

value ± 2.39 2.53 ± 3.11 3.29 ± 3.18 ± 2.21
3.2. Hardness and tensile properties

The hardness profile obtained for the weld joint, on its cross- mostly characterized by the mixed B + C type serrations, with type B
sectional plane, is shown in Fig. 3 and the average values are tabu­ mostly near the necking domain.
lated in Table 2. It can be deduced that no appreciable variation in the
average hardness values was observed (=82.09 ± 2.43 HV0.1) in the
3.3. Fatigue properties
transverse, and its perpendicular direction (i.e. depth), but was slightly
lower than that of the base metal (=87.10 ± 2.21 HV0.1). Fig. 4 shows
Fig. 5 shows the stress-life (S–N) curve, the stress amplitude (σa) vs
the average stress–strain curve for both the material conditions i.e., the
the number of load cycles (Nf), at different load ratios (R) for BM and
base metal (BM) AA5086 Al-Mg alloy and its weld joint (WJ). The
WJ, both. Table 4 and Table 5 tabulates the experimentally obtained
average values of mechanical properties are reported in Table 3. Unlike
fatigue data for different material conditions i.e. the weld joint (WJ) and
our expectations, the strength of the BM and its WJ was nearly same i.e.
the base metal (BM). All the fatigue life data (only broken specimens)
∼ 100% joint efficiency, but with a marginal reduction in ductility
was fitted using a simple power law, designated as Basquin fits,
(∼ 4.4%). All the tensile specimens of weld joints were failed from the
expressed as follows [24]:
heat affected zone (HAZ), attributed to its low hardness (see Table 2)
due to the partial annealing effect [21]. ′
σ a = σ f (2Nf )b (1)
Noticeably, both BM and WJ exhibit an additional feature of unstable
intermittent relaxation–oscillation type flow response i.e., a serrated Here σf and b represent the fatigue strength coefficient and the
′

flow behavior, after a particular strain of 0.35%, and 1%, respectively. Basquin exponent, respectively. The fatigue lives depreciated gradually
It indicates the presence of dynamic strain aging (DSA) manifested as with increasing stress amplitude at particular R-ratio, for both BM and
temporary obstacle to mobile dislocations by fast diffusing solute atoms WJ. The mean stress revealed a depreciating effect on fatigue lives and
during the monotonic loading due to the PLC effect [22]. Serration, in endurance limits, for both the material conditions i.e. decreased as R-
general, can be of five types [23]: type A (periodic, locking serrations), ratio was increased. An appreciable drop was observed in fatigue lives
type B (oscillations about the mean level, but with small amplitude), and endurance limit while going from R = − 1 towards 0.1, then from
type C (oscillations below the mean level), type D (plateau type), and R = 0.1 towards 0.5. Like mechanical properties, the fatigue life data of
type E (similar to type A but at high strain rates with little or no work WJ was either higher or similar to that of the BM. However, this dif­
hardening). The serrations in BM can be categorized as type A, followed ference in fatigue lives and endurance limit between BM and WJ tend to
by type B at large strain on onset of necking. However, the WJ was diminish at high R-ratios. The observed endurance limits at different R-

Fig. 3. The measured profile of Vickers hardness (HV0.1) across the weld cross-section. AS: Advancing Side, RS: Retreating Side, BM: Base Metal, NZ: Nugget Zone,
HAZ: Heat Affected Zone, TMAZ: Thermo-Mechanically Affected Zone.

4
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 4. A stress–strain curve of target base alloy 5086-H321, and its weld joint. The inset graph is the magnified view of the observed serration behavior. BM: Base
Metal, WJ: Welded Joint, RD: Rolling Direction, ND: Normal Direction, WD: Welding Direction, TD: Transverse Direction, ∊: ˙ Strain rate, T: Ambient Temperature.

Table 3
Mechanical properties of AA5086-H321 alloy and its weld joint. YS0.2%: Yield Strength at 0.2% offset, UTS: Ultimate Tensile Strength, EL: Elongation, E: Young’s
modulus, and σe : Endurance Limit.
YS0.2% (MPa) E (GPa) UTS (MPa) EL (%) σe (MPa)
R ¼ -1 R ¼ 0.1 R ¼ 0.5

BM 160.75 ± 8.33 65.79 ± 3.45 314.91 ± 7.88 29.60 ± 1.92 70 45 35

WJ 149.32 ± 5.98 57.04 ± 3.45 315.96 ± 6.69 25.19 ± 3.35 90 50 35

ratios, i.e. − 1, 0.1, and 0.5, for the base metal and its weld joint were 4. Discussion
[70, 45, and 35 MPa] and [90, 50, and 35 MPa], respectively. The
endurance limit of the weld joint was enhanced by ∼ 28% at R = − 1 as 4.1. Crack initiation mechanisms
compared to that of its base metal. All the fatigue data was re-plotted in
the Goodman diagram for the base metal and its weld joint, both, as Microscopic (SEM) examination of broken test coupons revealed two
shown in Fig. 6. The overall plot has been divided into three different different types of crack initiations: defect-free surface initiation and
domains: Safe region, Fatigue failure region, and the ductile failure re­ defect-induced surface initiation. In base metal, all the fatigue cracks
gion. The best prediction of endurance limits was satisfied by the lower were initiated from surface and were defect-free. However, in the case of
envelope formed by the Soderberg line of the weld joint, as shown in weld joint, both defect-free and defect-induced surface crack initiations
Fig. 6. The relations for these models are as follows: were observed. Table 4 enlists the type of observed damage mechanisms
Modified Goodman Line: for all the test samples of weld joints.
σa σm Fig. 7a shows an example of crack initiation for a specimen loaded at
+
σe σu
=1 σ a = 60 MPa, R = 0.1 and having a life of 945,164 cycles with defect-
free surface initiation. This was a favorable mode in the specimens
Soderberg Line: where the crack location was in the BM region, either on the top or the
σa σm bottom side of the welded plates, and preferentially in the smoother
+ =1
σe σy surface region i.e., free from any surface irregularities induced during
welding process. An evidence of primary crack path deflection at the
Where σa and σ m represent the applied stress amplitude and mean grain boundary, specifically near to the crack initiation site, along with a
stress (mid-range stress), respectively. σ e is defined as the endurance non-propagating secondary crack crossing the plausible slip bands
limit corresponding to R-ratio, R = − 1 and σy denotes the yield stress. To within a grain is illustrated in Fig. 7b. Such type of crack initiation was
have a better insight about the associated damage mechanisms and characterized by the crystallographic facets along with evident slip lines
probable explanations for the observed results, fracture surfaces were and rough (zig-zag) surface morphology, as shown in Fig. 7c. Such crack
examined and are discussed next. morphologies are, in general, shear driven i.e. stage-I type, and have also
been reported in MIG welded Al-Mg alloy (Al-5083) [25]. Similar type of
crystallographic features near to the crack initiation sites were also

5
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 5. Stress-life curve (S–N curve) at different load ratios (R = -1, 0.1, and 0.5) for both BM and WJ. BM: Base Metal, WJ: Welded Joint, RD: Rolling Direction,
ND: Normal Direction, WD: Welding Direction, TD: Transverse Direction, T: Ambient Temperature, R: Load Ratio.

observed for the higher R-ratios. A fracture surface of a specimen loaded edge of the plates while handling them during fixture (i.e., mechani­
at σ a = 45 MPa, R = 0.5 and having a life of 1,109,284 cycles, showing cally) or can be due to the wear of tool’s pin (Fe-based) [33]. These
such features is delineated in Fig. 7d with a primary crack propagating inclusions, mostly, initiate crack by debonding from the aluminum
along the preferred crystallographic planes but was observed only up to matrix thereby resulting a cavity and accelerating the damage process
a certain distance from the initiation site (∼ 1 mm). [34]. They were frequently observed in clusters near to the crack initi­
The observed defect-induced crack initiation can further be classified ation sites, which is known to be more detrimental to the fatigue lives
into three types: Void-induced initiation, Inclusion-induced initiation, and [35]. A critical observation of data provided in Table 4 reveals that the
Tool marks-induced initiation. All these defects tend to accelerate the inclusion-induced initiation had lower impact on fatigue lives as
crack initiation process due to the highly localized stress-fields around compared to voids induced initiation.
them [26,27], and are discussed next. Tool marks-induced initiation. Fig. 8e and 8f shows this type of crack
Void-induced initiation. Fig. 8a and 8b illustrates this type of crack initiation mode. These tool marks were formed on the top surface of
initiation mode. Two different profiles of voids were observed: irregular welded plates having the characteristics features of micro-saw tooth
and nearly elliptical shaped. All the types of voids were located at or profile, when projected on the cross-sectional plane, as shown on a
near to the surface. Irregular shaped voids were mostly observed on top fracture surface in Fig. 8e and 8f. This external flaw, also known as
surface of the welded plates i.e. in the tool’s shoulder contact region, and surface irregularities, is due to the shoulder of the rotating tool and
can be attributed to the insufficient plunging action of the tool. While, exhibits higher surface roughness of 3.10 ± 0.21 µm as compared to the
the ellipsoidal shaped voids were preferably observed at the bottom side other regions (0.50 ± 0.07 µm). A few specimens broke from the BM/
of welded plates, near to NZ, and is often termed as “root flaws” formed HAZ interface at the top i.e. shoulder contact side (see Table 4), as this
due to insufficient flow of material [28,29] The average size of such interface is the transition point for this change in surface roughness [9].
defects was in the range of 60–150 µm, comparable to or bigger than the Like voids and inclusions, these geometrical irregularities facilitate the
grain size. Such defects are often known to reduce the crack initiation early crack initiation due to the highly localized stress concentration
period significantly and to induce scatter in fatigue lives [30]. factor [36]. Mean stress tends to favor such type of crack initiation (see
Inclusion-induced initiation. Fig. 8c and 8d shows this type of crack Table 4), probably due to sufficiently large plastic flow near to the peak-
initiation mode. Two types of crack-initiating inclusions were observed: valleys of surface profile during the initial ramp loading.
Metallics (Fe-rich) and Non-metallics (O-rich). Most of the crack- Fig. 9 and Fig. 10 shows the observed two different special cases of
initiating inclusions were non-metallics while only one case of multiple crack-initiation sites, both at relatively smaller fatigue lives and
metallic inclusion was found. The EDAX analysis performed on each of loaded at different R-ratio (0.5 and 0.1). Fig. 9 shows the fracture surface
them is also shown in Fig. 8c and 8d. Like voids, almost all of the crack- of a specimen loaded at σa = 68 MPa, R = 0.5 and having a life of 74,374
initiating inclusions were located on the surface or sub-surface region. cycles. All the cracks initiated from the voids of size ranges between 40
The presence of O-rich non-metallic inclusion can be attributed to the and 60 µm, located in the heat affected zone (HAZ) and closer to the
high affinity of aluminum towards ambient oxygen thereby forming a nugget zone (NZ) on the bottom surface side of the welded plates. Post-
passive oxide layer, later on fragmented into pieces by the rotating FSW analysis revealed that these voids were formed on the advancing side of
tool [31,32]. The presence of Fe-rich metallic inclusion, only in one case the weld joint and thus can be attributed to the low velocity of material
and of size ~40 µm, can be due to the external sources in the abutting flow in the back advancing edge of the tool pin [37]. Fig. 10 shows the

6
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Table 4 Table 5
High cycle fatigue test data of welded AA5086-H321. σa : Stress amplitude, Nf : High cycle fatigue test data of base AA5086-H321. σa : Stress amplitude, Nf :
Fatigue life, R: Load ratio, P-CIS: Primary Crack initiation site, CGR: Crack Fatigue life, R: Load ratio.
growth regimes, TS: Top Surface, BS: Bottom Surface, SS: Side Surface, AS: No. σa (MPa) Nf (Cycle) R
Advancing Side, RS: Retreating Side, IG: Intergranular, TG: Transgranular, NZ:
Nugget Zone, HAZ: Heat Affected Zone, TMAZ: Thermo-Mechanically Affected 1 115 152,727 − 1
2 115 195,510 − 1
Zone.
3 115 159,232 − 1
No. σa (MPa) Nf R P-CIS CGR Fracture 4 100 213,638 − 1
features feature Location 5 100 352,713 − 1
6 100 341,510 − 1
1 130 100,205 − 1 Tool marks IG NZ (TS)
7 85 850,424 − 1
2 130 84,737 − 1 Void IG NZ (TS)
8 85 455,900 − 1
3 130 81,747 − 1 Tool marks IG NZ (TS)
9 85 438,843 − 1
4 120 215,424 − 1 Inclusion TG AS-HAZ
10 70 7,467,533 − 1
(BS)
11 70 10,000,000 − 1
5 120 262,108 − 1 Tool marks IG RS-HAZ
12 70 10,000,000 − 1
(TS)
13 90 82,431 0.1
6 120 214,179 − 1 Free-surface TG BM (TS)
14 90 58,015 0.1
7 110 209,118 − 1 Void IG NZ (TS)
15 90 66,103 0.1
8 110 588,103 − 1 Free-surface TG AS-HAZ
16 75 162,943 0.1
(BS)
17 75 194,148 0.1
9 110 430,492 − 1 Free-surface TG RS-HAZ
18 75 160,461 0.1
(BS)
19 60 565,990 0.1
10 100 263,052 − 1 Void IG RS-HAZ
20 60 990,390 0.1
(TS)
21 60 734,660 0.1
11 100 556,350 − 1 Free-Surface TG RS-BM (BS)
22 45 10,000,000 0.1
12 100 315,899 − 1 Inclusion TG BM/RS-
23 45 10,000,000 0.1
HAZ (BS)
24 45 10,000,000 0.1
13 90 1,001,888 − 1 Free-Surface TG AS-BM (TS)
25 68 117,146 0.5
14 90 10,000,000 − 1 — — —
26 68 112,368 0.5
15 90 10,000,000 − 1 — — —
27 68 139,567 0.5
16 90 53,722 0.1 Free-surface IG NZ (BS)
28 60 196,849 0.5
17 90 39,507 0.1 Mechanical IG RS-HAZ
29 60 189,738 0.5
dent (BS)
30 60 156,357 0.5
18 90 38,599 0.1 Inclusion TG BM (BS)
31 45 1,978,001 0.5
19 75 195,015 0.1 Tool marks IG AS-HAZ
32 45 382,448 0.5
(TS)
33 45 10,000,000 0.5
20 75 146,045 0.1 Inclusion IG AS-HAZ
34 35 10,000,000 0.5
(TS)
35 35 10,000,000 0.5
21 75 113,377 0.1 Inclusion IG AS-HAZ
36 35 10,000,000 0.5
(TS)
22 60 945,164 0.1 Free-Surface TG RS-BM (BS)
23 60 712,272 0.1 Free-Surface IG NZ/RS-
TMAZ (SS) initiation, probably from the surface irregularities [38], with one major
24 60 958,754 0.1 Void TG RS-HAZ crack and other secondary cracks. Interestingly, intergranular facets
(BS) were found around all the crack initiation sites, respectively, and are
25 50 2,760,299 0.1 Inclusion IG NZ (SS) discussed next in detail.
26 50 10,000,000 0.1 — — —
27 50 10,000,000 0.1 — — —
28 68 52,720 0.5 Free-surface IG NZ (BS) 4.2. Crack propagation mechanisms
29 68 74,347 0.5 Void IG AS-HAZ
(SS)
30 68 93,335 0.5 Free-Surface IG RS-HAZ Fractographic examination of fractured surfaces revealed two
(BS) different types of fatigue crack growth mechanisms: intergranular facia-
31 60 110,401 0.5 Inclusion TG BM/AS- controlled (see Fig. 11a) and transgranular facia-controlled (see
HAZ (BS)
Fig. 11b). A clear presence of slip bands along the grain boundaries
32 60 291,140 0.5 Tool marks IG NZ (TS)
33 60 144,934 0.5 Tool marks IG NZ (TS) during the intergranular fracture was observed, along with the frequent
34 52 581,805 0.5 Free-surface TG AS-HAZ evidence of intergranular crack branching, and is shown in Fig. 11c and
(BS) 11d. These slip lines at the grain boundaries appeared to have a terrace
35 52 305,992 0.5 Tool marks IG AS-HAZ topography and are formed to suppress the interfacial energy up to
(TS)
36 52 335,494 0.5 Tool marks IG NZ (TS)
certain extent [39]. This can provide a certain degree of relaxation
37 45 1,109,284 0.5 Inclusion TG AS-HAZ during the de-cohesion along the grain boundary during the crack
(BS) growth. The fracture surfaces of specimens subjected to fully reversed
38 45 573,266 0.5 Tool marks IG AS-HAZ loadings (R = − 1), in both WJ and BM, were completely mated and
(TS)
thus, no useful information could be extracted except the presence of
39 45 855,777 0.5 Inclusion TG AS-BM (BS)
40 35 4,607,255 0.5 Void TG RS-HAZ some typical tire tracks [40] and tear ridges [41], as shown in Fig. 11e
(BS) and 11f. Transgranular type mechanisms was mainly predominant in
41 35 10,000,000 0.5 — — — failures of base alloys while in the case of weld joints, both types of
42 35 10,000,000 0.5 — — — mechanisms were observed. The presence of intergranular cracks on the
fracture surfaces appeared to be controlled by the microstructural het­
fracture surface of another specimen loaded at σ a = 90 MPa, R = 0.1 and erogeneity across the weld joint (NZ, HAZ, and BM), while no such
having a life of 53,722 cycles. The specimen broke from the NZ of the heterogeneity was anticipated nor found in the case of base alloys.
weld joint and no voids or inclusions could be located around any of the The weld joint specimens having crack initiation either in the HAZ or
crack initiation sites. It is a classic case of multiple surface crack at the BM/HAZ interface, predominantly towards the BM region, as well
as the base metal specimens exhibit the transgranular crack growth

7
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 6. Goodman plot of fatigue data points for both base metal as well as its weld joint along with the modified Goodman line and Soderberg line. The red color star
is for base metal and green circular dot is for its weld joint. Hollow symbols represent the ran-out specimens. (For interpretation of the references to color in this
figure legend, the reader is referred to the web version of this article.)

Fig. 7. SEM micrographs (a) showing the crack initiation from free-surface along with (b) the non-propagating secondary crack, and (c) crystallographic facets at
R = 0.1, σa = 60 MPa, and Nf = 945, 164 cycles and (d) at R = 0.5, σa = 45 MPa, and Nf = 1, 109, 284 cycles .

8
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 8. The observed different types of fatigue crack initiation modes.

mechanism. The fracture surfaces were also characterized by the input from the tool-shoulder, along with rapid cooling rate thereby
chevron or river-like patterns (see Fig. 11b). The region of HAZ, BM, and resulting in relatively finer and equiaxed grains within the NZ. The
their interface is characterized by relatively larger grains and can favor presence of such grain morphology within the NZ can favor the propa­
such type of crack growth behavior. This can be attributed to the pref­ gation of cracks via intergranular mechanism [43]. Intergranular facets
erential cyclic strain localization within the larger grains and along the around the crack initiation sites were also observed in the cases where
favorable crystallographic planes, slip bands, during the cyclic loadings crack initiated from the bottom side of welded plates within the NZ. This
[42]. can also be attributed to the presence of relatively finer grains within the
On the other hand, the weld nugget zone (NZ), characterized by the NZ, near to the backing plate in contact with the work piece, due to the
finer grains, was found to be the favorable region for the observed relatively lower peak temperature and short thermal cycle, in-turn
intergranular crack growth which later on changed to the transgranular resulting in faster heat dissipation during the welding [2].
one upon reaching the HAZ region, characterized by relatively larger
grains. All the specimens showing the intergranular facets were failed in 4.3. Fatigue lives and endurance limit
the NZ, with the cracks initiating from the top side of the welded plates i.
e. the side of tool-tip contact. This side of the weld joint can suffer Weld joints of any alloy, in general, are known to have reduced fa­
relatively higher degree of deformation at the tool-tip than the heat tigue properties as compared to that of their base alloys. However, in the

9
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 9. An example of observed fatigue damage mechanism with multiple crack initiation sites at R = 0.5, σa = 68 MPa, and Nf = 74, 374 cycles. BM: Base Metal,
HAZ: Heat Affected zone, TMAZ: Thermo-Mechanically Affected Zone, NZ: Nugget Zone, RD: Rolling Direction, ND: Normal Direction, WD: Welding Direction, TD:
Transverse Direction, AS: Advancing Side, RS: Retreating Side.

present study, similar to hardness and tensile properties, the weld joints AA5083-H32, which reduced the effective crack driving force i.e. the
of AA5086 were found to have either higher or similar fatigue proper­ effective ΔK, and thus increased the fatigue lives. However, in the pre­
ties, irrespective of the applied R-ratio, as compared to the parent metal. sent study most of the specimens with intergranular fracture were sub­
The difference was higher at low R-ratios i.e. at R = − 1, which was jected to high mean stresses (R ≥ 0.1). So, compressive residual stresses,
continuously diminishing with its increasing value i.e. at R ≥ 0.1. The even if present, are most likely to get relaxed during the ramp loading
reason for such a behavior can be attributed to the type of observed itself and thus cannot account for enhanced fatigue lives in the presence
crack growth mechanism in the respective scenarios i.e. transgranular of intergranular facets. Moreover, in our previous study on same mate­
and intergranular. Intergranular fatigue crack growth is, in general, rial but with different temper condition (AA5086-H116) and different
slower than the transgranular one, however, mixed opinion does exist in tool type (non-threaded tip), the residual stresses measured on the cross-
the literature [44,45]. The work by Jata et al. [44] concluded a faster sectional plane of the weld joint were mostly tensile [9].
intergranular crack growth in the friction stir welded AA7050-T7451, However, in many other studies the beneficial effect of intergranular
however, the work by Kim et al. [45] reported an opposite effect in fracture on fatigue lives has been attributed to the frequent grain
friction stir welded AA5083-H32. Kim et al. [45] reported a retardation boundary interactions during the propagation stage [46]. The localized
effect during the intergranular faceted crack growth and attributed it to microstructural heterogeneity present across the weld joint serve as a
the compressive residual stresses in the NZ of friction stir weld joints of barrier to crack propagation owing to a more frequent crack path

10
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 10. An example of fatigue damage mechanism with multiple crack initiation sites for the specimen tested at R = 0.1, σa = 90 MPa, and Nf = 53,722 cycles . BM:
Base Metal, HAZ: Heat Affected Zone, TMAZ: Thermo-Mechanically Affected Zone, NZ: Nugget Zone, RD: Rolling Direction, ND: Normal Direction, WD: Welding
Direction, TD: Transverse Direction, AS: Advancing Side, RS: Retreating Side, CIR: Crack initiation regimes, CGR: Crack growth regimes, FFR: Fast fracture regimes.

deflection at the grain boundaries, see Fig. 7b. A crack initiating from behavior of the AA5086-H321 and its friction stir welded joints were
the NZ in the weld joint, characterized by fine and equiaxed grains, investigated. The uniaxial fatigue tests were performed at different R-
propagates preferentially along the grain boundaries before entering ratios ( − 1, 0.1, and 0.5) using a sinusoidal-type loading waveform and a
into the HAZ, which is a combination of big and small grains: a typical test frequency of 20 Hz. The following important conclusions can be
bi-modal structure (see Fig. 2d). This tortuous crack path retards the drawn based on the results observed, presented and discussed in this
crack growth rate in the weld joints due to repeated grain boundary study:
interaction, thereby improving the fatigue lives. Similar observations on
crystallographic faceted crack growth have also been reported in duplex • Unlike expected, no difference in the hardness or the tensile prop­
stainless steel [47] and Cr-Mo steel [48] but, however, attributed to erties of the base alloy and its weld joint was observed. However, a
different reasons. This argument can very well explain the observed marginal reduction in ductility (∼ 4.4%) was noted in the weld joint.
enhancement in fatigue properties of weld joints as compared to that of • Both fatigue lives and endurance limits of the weld joints were either
their base alloys. Another reason for the retarded crack growth rate or higher or similar to that of their base alloys, irrespective of the
the increased fatigue lives during the intergranular fracture can be the applied R-ratio.
roughness-induced crack closure effects [49,50]. The tortuous crack • The difference in the fatigue lives of base alloy and its weld joints was
surface formed during the intergranular fracture certainly contributes to higher at low mean stresses (R = − 1) and tends to diminish with
this effect and can also explain the observed fatigue behavior. increasing R-ratios (R ≥ 0.1).
• Fractographic examination of weld joints revealed three types of
5. Concluding remarks crack initiations: voids, inclusions, and tool-marks, and two types of
crack propagation mechanisms: intergranular and transgranular.
In the present study, the mechanical properties and the fatigue

11
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 11. Different types of fatigue crack growth mechanisms observed in AA5086 and its weld joint.

• An enhancement in fatigue lives of weld joint was attributed to the crack growth observed in the present study are also depicted along with
retarded intergranular crack growth rate owing to the frequent grain the applied loading direction.
boundary interaction and roughness-induced crack closure, which
was absent in the base alloys. Data availability
• Mean stress, as expected, has detrimental effect of fatigue lives and
endurance limits, with no definitive control on type of crack initia­ The raw/processed data required to reproduce these findings cannot
tion or propagation. be shared at this time as the data also forms part of an ongoing study.

Finally, based on the different types of crack initiation or propaga­

CRediT authorship contribution statement
tion mechanisms observed in this study, a schematic summary delin­
eating the different types of fatigue damage mechanisms observed in the
Rajneesh Jaisawal: Data curation, Formal analysis, Conceptuali­
friction stir welded AA5086-H321 along with the failure locations with zation, Methodology, Investigation, Validation, Writing – original draft.
respect to the surface is shown in Fig. 12. All the microstructurally-
Vidit Gaur: Data curation, Formal analysis, Conceptualization, Meth­
different zones across the weld joints have also been shown in the odology, Investigation, Validation, Writing – review & editing, Super­
figure. The two specific regions i.e. top and bottom side, of intergranular
vision, Funding acquisition, Project administration. Shahnawaz

12
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

Fig. 12. A schematic representation observed damage mechanisms in the weld joint of AA5086-H321.

Ahmed: Project administration, Writing – review & editing. [9] Choudhary S, Gaur V. Materials Science & Engineering A Enhanced fatigue
properties of AA5086 friction stir weld joints by Mater Sci Eng A 2023;869:144778.
https://doi.org/10.1016/j.msea.2023.144778.
Declaration of Competing Interest [10] Hassan HA. Study the weld Region by Friction Stir Welding of AA5086-H32 by
Using a. Int J Appl Eng Res 2018;13:4966–74.
The authors declare the following financial interests/personal re­ [11] Goyal A, Garg RK. Establishing Mathematical Relationships to Study Tensile
Behavior of Friction Stir Welded AA5086-H32 Aluminium Alloy Joints. SILICON
lationships which may be considered as potential competing interests: 2019;11:51–65. https://doi.org/10.1007/s12633-018-9858-4.
Vidit Gaur reports financial support was provided by Defence Research [12] Wang BB, Xue P, Xiao BL, Wang WG, Liu YD, Ma ZY. Achieving equal fatigue
and Development Organisation. strength to base material in a friction stir welded 5083–H19 aluminium alloy joint.
Sci Technol Weld Join 2020;25:81–8. https://doi.org/10.1080/
13621718.2019.1630571.
Data availability [13] Dickerson TL, Przydatek J. Fatigue of friction stir welds in aluminium alloys that
contain root flaws. Int J Fatigue 2003;25:1399–409. https://doi.org/10.1016/
S0142-1123(03)00060-4.
Data will be made available on request. [14] James MN, Hattingh DG, Bradley GR. Weld tool travel speed effects on fatigue life
of friction stir welds in 5083 aluminium. Int J Fatigue 2003;25:1389–98. https://
Acknowledgements doi.org/10.1016/S0142-1123(03)00061-6.
[15] Kurbet R, Karthikareddy G, Monica J. A Review on Friction Stir Welding over other
Welding Techniques of Aluminium Alloys. Solid State Technol 2021;64:3713–29.
The authors are thankful to the Materials panel of Naval Research [16] INTERNATIONAL A. Standard Test Methods for Determining Average Grain Size.
Board (NRB) – Defense Research and Development Organization Astm E112-10 2010;13:1–27. https://doi.org/10.1520/E0112-13R21.1.4.
[17] Niessen F, Nyyssönen T, Gazder AA, Hielscher R. Parent grain reconstruction from
(DRDO) for funding this project (Grant No. NRB/4003/MAT/PG/477)
partially or fully transformed microstructures in MTEX. J Appl Crystallogr 2022;
and IIT Roorkee for providing all the relevant infrastructural facilities to 55:180–94. https://doi.org/10.1107/S1600576721011560.
facilitate the research work. [18] ASTM. Standard Test Method for Microindentation Hardness of MaterialsKnoop
and Vickers Hardness of Materials 1. Annu B ASTM Stand 2010:1–42. https://doi.
org/10.1520/E0384-17.
References [19] ASTM E8. ASTM E8/E8M standard test methods for tension testing of metallic
materials 1. Annu B ASTM Stand 2010;4:1–27. https://doi.org/10.1520/E0008.
[1] Thompson FC. Welding of Aluminium and its Alloys. Welding of aluminium and its [20] Testing SF, Alignment S, Tensile U, Axial C, Application F, Testing F. Standard
alloys 1956;177(4508):568–9. https://doi.org/10.1038/177568b0. Practice for Conducting Force Controlled Constant Amplitude Axial Fatigue Tests
[2] Mishra RS, Ma ZY. Friction stir welding and processing. Mater Sci Eng R Reports of. Met Mater 2010;1(03):1–6. https://doi.org/10.1520/E0466-07.2.
2005;50:1–78. https://doi.org/10.1016/j.mser.2005.07.001. [21] Garg A, Bhattacharya A. Effect of microstructural variation on strain localization in
[3] Rhodes CG, Mahoney MW, Bingel WH, Spurling RA, Bampton CC. Effects of friction double-sided friction stir welded AA6061-AA7075 joints. Strain 2022;58:1–18.
stir welding on microstructure of 7075 aluminum. Scr Mater 1997;36:69–75. https://doi.org/10.1111/str.12413.
https://doi.org/10.1016/S1359-6462(96)00344-2. [22] Aboulfadl H, Deges J, Choi P, Raabe D. Dynamic strain aging studied at the atomic
[4] He C, Cui SM, Wu YZ, Luo ZF, Wang QY. Fatigue crack initiation mechanism of scale. Acta Mater 2015;86:34–42. https://doi.org/10.1016/j.
aluminium alloy welded joint in gigacycle fatigue. Appl Mech Mater 2013;376: actamat.2014.12.028.
177–80. https://doi.org/10.4028/www.scientific.net/AMM.376.177. [23] Zhemchuzhnikova D, Lebyodkin M, Yuzbekova D, Lebedkina T, Mogucheva A,
[5] Lin S, Deng Y-L, Tang J-G, Deng S-H, Lin H-Q, Ye L-Y, et al. Microstructures and Kaibyshev R. Interrelation between the portevin le-chatelier effect and necking in
fatigue behavior of metal-inert-gas-welded joints for extruded Al-Mg-Si alloy. almg alloys. Int J Plast 2018;110:95–109. https://doi.org/10.1016/j.
Mater Sci Eng A 2019;745:63–73. https://doi.org/10.1016/j.msea.2018.12.080. ijplas.2018.06.012.
[6] Tagawa T, Tahara K, Abe E, Katsuragi Y, Shinoda T, Minami F. Fatigue properties [24] Shukla S, Komarasamy M, Mishra RS. Grain size dependence of fatigue properties
of cast aluminium joints by FSW and MIG welding. Weld Int 2014;28:21–9. https:// of friction stir processed ultrafine-grained Al-5024 alloy. Int J Fatigue 2018;109:
doi.org/10.1080/09507116.2012.715881. 1–9. https://doi.org/10.1016/J.IJFATIGUE.2017.12.007.
[7] Lombard H, Hattingh DG, Steuwer A, James MN. Optimising FSW process [25] Gaur V, Enoki M, Okada T, Yomogida S. A study on fatigue behavior of MIG-
parameters to minimise defects and maximise fatigue life in 5083–H321 aluminium welded Al-Mg alloy with different filler-wire materials under mean stress. Int J
alloy. Eng Fract Mech 2008;75:341–54. https://doi.org/10.1016/j. Fatigue 2018;107:119–29. https://doi.org/10.1016/j.ijfatigue.2017.11.001.
engfracmech.2007.01.026. [26] Ammar HR, Samuel AM, Samuel FH. Effect of casting imperfections on the fatigue
[8] Zhou C, Yang X, Luan G. Fatigue properties of friction stir welds in Al 5083 alloy. life of 319-F and A356–T6 Al-Si casting alloys. Mater Sci Eng A 2008;473:65–75.
Scr Mater 2005;53:1187–91. https://doi.org/10.1016/j.scriptamat.2005.07.016. https://doi.org/10.1016/j.msea.2007.03.112.

13
R. Jaisawal et al. International Journal of Fatigue 174 (2023) 107712

[27] Günther J, Krewerth D, Lippmann T, Leuders S, Tröster T, Weidner A, et al. Fatigue [40] Smirnova LL, Zinin AV. A Study of Cyclic Fracture of Structural Materials Under
life of additively manufactured Ti–6Al–4V in the very high cycle fatigue regime. Int Low-Cycle Overloads. Met Sci Heat Treat 2021;62:759–64. https://doi.org/
J Fatigue 2017;94:236–45. https://doi.org/10.1016/j.ijfatigue.2016.05.018. 10.1007/s11041-021-00634-4.
[28] Bin CH, Yan K, Lin T, Ben CS, Jiang CY, Zhao Y. The investigation of typical [41] Dai Q, Liang Z, Chen G, Meng L, Shi Q. Explore the mechanism of high fatigue
welding defects for 5456 aluminum alloy friction stir welds. Mater Sci Eng A 2006; crack propagation rate in fine microstructure of friction stir welded aluminum
433:64–9. https://doi.org/10.1016/j.msea.2006.06.056. alloy. Mater Sci Eng A 2013;580:184–90. https://doi.org/10.1016/j.
[29] Dada OJ. Fracture mechanics and mechanical behaviour in AA5083-H111 friction msea.2013.05.057.
stir welds. Sci African 2020;8:e00265. [42] Barat K, Ghosh A, Doharey A, Mukherjee S, Karmakar A. Crystallographic
[30] Gaur V, Briffod F, Enoki M. Micro-mechanical investigation of fatigue behavior of evaluation of low cycle fatigue crack growth in a polycrystalline Ni based
Al alloys containing surface/superficial defects. Mater Sci Eng A 2020;775:138958. superalloy. Int J Plast 2022;149:103174. https://doi.org/10.1016/j.
https://doi.org/10.1016/j.msea.2020.138958. ijplas.2021.103174.
[31] Campbell T, Kalia RK, Nakano A, Vashishta P, Ogata S, Rodgers S. Dynamics of [43] Taheri Mousavi SM, Hosseinkhani B, Vieillard C, Chambart M, Kok PJJ,
oxidation of aluminum nanoclusters using variable charge molecular-dynamics Molinari JF. On the influence of transgranular and intergranular failure
simulations on parallel computers. Phys Rev Lett 1999;82:4866–9. https://doi.org/ mechanisms during dynamic loading of silicon nitride. Acta Mater 2014;67:
10.1103/PhysRevLett.82.4866. 239–51. https://doi.org/10.1016/j.actamat.2013.12.032.
[32] Mutombo K, du M. Corrosion Fatigue Behaviour of Aluminium 5083-H111 Welded [44] Jata KV, Sankaran KK, Ruschau JJ. Friction-stir welding effects on microstructure
Using Gas Metal Arc Welding Method. Arc Weld 2011. https://doi.org/10.5772/ and fatigue of aluminum alloy 7050–T7451. Metall Mater Trans A Phys Metall
25991. Mater Sci 2000;31:2181–92. https://doi.org/10.1007/s11661-000-0136-9.
[33] Garg A, Bhattacharya A. Assessing profile damage of uncoated and AlTiN coated [45] Kim S, Lee CG, Kim SJ. Fatigue crack propagation behavior of friction stir welded
FSW tools after successive travel on AA6061-T6 plate. CIRP J Manuf Sci Technol 5083–H32 and 6061–T651 aluminum alloys. Mater Sci Eng A 2008;478:56–64.
2021;35:839–54. https://doi.org/10.1016/j.cirpj.2021.09.009. https://doi.org/10.1016/j.msea.2007.06.008.
[34] Zerbst U, Madia M, Hellmann D. An analytical fracture mechanics model for [46] D’Urso G, Giardini C, Lorenzi S, Pastore T. Fatigue crack growth in the welding
estimation of S-N curves of metallic alloys containing large second phase particles. nugget of FSW joints of a 6060 aluminum alloy. J Mater Process Technol 2014;214:
Eng Fract Mech 2012;82:115–34. https://doi.org/10.1016/j. 2075–84. https://doi.org/10.1016/j.jmatprotec.2014.01.013.
engfracmech.2011.12.001. [47] Tofique MW, Bergström J, Svensson K, Johansson S, Peng RL. ECCI/EBSD and TEM
[35] Meyer S, Diegele E, Brückner-Foit A, Möslang A. Crack interaction modelling. analysis of plastic fatigue damage accumulation responsible for fatigue crack
Fatigue Fract Eng Mater Struct 2000;23:315–23. https://doi.org/10.1046/j.1460- initiation and propagation in VHCF of duplex stainless steels. Int J Fatigue 2017;
2695.2000.00283.x. 100:251–62. https://doi.org/10.1016/j.ijfatigue.2017.03.035.
[36] Schork B, Kucharczyk P, Madia M, Zerbst U, Hensel J, Bernhard J, et al. The effect [48] Gaur V, Doquet V, Persent E, Roguet E. Effect of biaxial cyclic tension on the
of the local and global weld geometry as well as material defects on crack initiation fatigue life and damage mechanisms of Cr-Mo steel. Int J Fatigue 2016;87:124–31.
and fatigue strength. Eng Fract Mech 2018;198:103–22. https://doi.org/10.1016/ https://doi.org/10.1016/j.ijfatigue.2016.01.021.
j.engfracmech.2017.07.001. [49] Zhu ML, Xuan FZ, Tu ST. Observation and modeling of physically short fatigue
[37] Liechty BC, Webb BW. Modeling the frictional boundary condition in friction stir crack closure in terms of in-situ SEM fatigue test. Mater Sci Eng A 2014;618:86–95.
welding. Int J Mach Tools Manuf 2008;48:1474–85. https://doi.org/10.1016/j. https://doi.org/10.1016/j.msea.2014.08.021.
ijmachtools.2008.04.005. [50] Liu SD, Zhu ML, Zhou HB, Wan D, Xuan FZ. Strain visualization of growing short
[38] Madia M, Schork B, Bernhard J, Kaffenberger M. Multiple crack initiation and fatigue cracks in the heat-affected zone of a Ni–Cr–Mo–V steel welded joint:
propagation in weldments under fatigue loading. Procedia Struct Integr 2017;7: Intergranular cracking and crack closure. Int J Press Vessel Pip 2019;178:103992.
423–30. https://doi.org/10.1016/j.prostr.2017.11.108. https://doi.org/10.1016/j.ijpvp.2019.103992.
[39] Lynch S. A review of underlying reasons for intergranular cracking for a variety of
failure modes and materials and examples of case histories. Eng Fail Anal 2019;
100:329–50. https://doi.org/10.1016/j.engfailanal.2019.02.027.

14