Quasi-in-situ observation of fatigue crack growth behavior of friction stir welded 2024-T4 joint

This study presents a quasi-in situ observation of the fatigue crack growth behavior in a friction stir welded 2024-T4 joint. The microstructure and fatigue properties of the joint were investigated using electron backscatter diffraction (EBSD) technique, scanning electron microscopy (SEM), and fatigue crack growth tests. The fatigue crack growth behavior of the joint was examined by conducting fatigue crack growth tests with different notch locations. The results show that the sample with the notch in the stir zone (SZ) exhibited the highest resistance to fatigue crack growth, followed by the notched samples of the Advancing side (AS) and Retreating side (RS) weldments. Microstructural observations showed a homogeneous microstructure with a fine grain size in SZ and it was observed that this fine-grained structure significantly enhanced the material’s resistance to fatigue crack growth. The experimental results were further analyzed using the Paris model to provide a quantitative understanding of the crack growth behavior. The study underlines the impact of microstructural characteristics and notch location on the fatigue performance of the weldment. Overall, the quasi-in situ observations and experimental findings contribute to a comprehensive understanding of the fatigue crack growth behavior in friction stir welded 2024-T4 joints.


Introduction
Friction stir welding (FSW) has emerged as a preferred joining technique in aviation structural manufacturing due to its numerous advantages, including improved mechanical properties and reduced defects [1,2].It has numerous potential applications in the aerospace industry, ranging from aircraft fuselage panels to engine components [1,3].The reliability of FSW joints critical to ensuring the structural integrity of aerospace components.Fatigue crack propagation behavior has become a major concern, as it might jeopardize the performance and lifespan of welded structures [4,5].
Several research studies have been conducted to explore the fatigue behavior of FSW joints in various aluminum alloys [6][7][8][9][10].Huang et al [11] studied the microstructure, hardness, tensile properties, and crack growth behavior of a 12-mm 5083 aluminum-alloy bobbin tool for friction stir welding joint.They noticed decreased mechanical properties after welding, with the heat-affected zone showing lower tensile strength and elongation compared to the base material.Cracks in the weld expanded quickly, especially near the heat-affected zone.Qiao et al [12] compared the fatigue crack growth properties of AA 5754 aluminum alloy welded pipes in friction stir welding (FSW) and gas tungsten arc welding (GTAW).The GTAW specimen exhibited the highest crack growth rate in the welded metal (WM), while the FSW specimen showed the lowest rate.This difference was attributed to coarse grains and high residual stress in the GTAW WM, affecting the plastic region near the crack tip.Vuherer et al [8] studied the fatigue behavior of Friction Stir Welded AA2024-T351 joints at varying welding speeds.The results showed no significant differences in the fatigue behavior between the welding speeds when analyzing the Paris diagram.However, the joint welded at 116 mm min −1 had the maximum fracture toughness and resistance to fracture in the thermal-mechanical affected zone (TMAZ).The fracture toughness of the base material was similar to that of joints welded at speeds of 73 mm min −1 and 150 mm min −1 .Zhang et al [13] studied the impact and fatigue crack growth (FCG) behavior in FSW of 6061-T6 aluminum alloy.The SZ has smaller grain size, hardness, and impact energy compared to the base material.Particle presence hindered crack propagation, while microstructural variables influenced FCG rate.Kumar et al [14] investigated fatigue crack growth in friction stir welded AA7075-T651 aluminum alloy joints.The stress intensity factor range ΔK of the welded joint was lower than the unwelded parent metal, attributed to precipitate dissolution during welding.This suggests reduced fatigue crack growth resistance in the welded joints compared to the parent metal.
However, limited attention has been paid to the observation and correlation of fatigue crack growth behavior with mechanical properties, notably in the case of friction stir welded 2024-T4 joints.To address this research gap, this study aims to provide a systematic investigation of the fatigue crack growth behavior of friction stir welded 2024-T4 joints.Previous research has primarily focused on macroscopic fatigue properties, while microscopic observations of crack growth and its correlation with microstructural variations remain lacking.Thus, in this study, four groups of samples notched at different regions, including the base material (BM), advancing side (AS), retreating side (RS), and stir zone (SZ), will be evaluated to compare fatigue crack growth behavior.Furthermore, advanced characterization techniques including electron backscatter diffraction (EBSD) and scanning electron microscopy (SEM) will be utilized to achieve quasi-in situ observation of fatigue crack growth behavior in friction stir welded 2024-T4 joints.

Friction stir welding for joint preparation
Prior to welding, the aluminum alloy plates were properly cleaned to ensure consistent and high-quality welds.Friction stir welding was then performed to join the plates of 2024-T4 aluminum alloy plates with the dimensions of 300 mm × 100 mm × 2 mm. Figure 1 shows that the plates were joined at a rotation speed of 1200 rpm and a traverse speed of 300 mm min −1 [15].Following weldment preparation, samples for FCG test were prepared from the locations shown in figure 1.

Fatigue crack growth test
Before the FCG test, tensile tests were performed using a UTM5000 testing machine following the ISO 4136:2001 standard.The samples were cut perpendicularly to the welding direction, with the welded joint located at the center of the sample.The tensile tests were conducted at a loading speed of 1 mm min −1 .Then, the hardness profile was tested on a polished sample across the welded joint using a Vickers hardness tester (200 HVS-5).The test was carried out with a load of 100 g and a dwell time of 15 s.
Compact tension (CT) specimens were utilized for conducting fatigue crack growth tests.The MTS-810 machine was employed to assess the rate at which the fatigue crack grew during testing.The testing procedure adhered to the guidelines outlined in the ASTM E647 standard.All the experiments were conducted with a load ratio (R) of 0.1 and a frequency of 10 Hz.The maximum load (P max ) was set to 1000 N, while the minimum load (P min ) was set of 100 N. The fatigue crack growth test was conducted at room temperature, applying cyclic loading with the specified load ratio and frequency.The FCG direction was parallel to welding direction, four groups of compact tension (CT) samples were prepared with the crack tip notched at base material (CT-BM), stir zone (CT-SZ), advancing side (CT-AS) and retreating side (CT-RS), respectively.Crack length was measured continually using the compliance method with a crack mouth clip gauge.Data on the number of cycles and crack length were gathered for further investigation.The obtained data was then processed to calculate the fatigue The results obtained from the FCGR analysis were analyzed and compared to identify any differences between the samples.The influence of the friction stir welding process on the fatigue crack growth behavior of the joints was evaluated.Finally, the results were discussed in terms of their significance for the fatigue performance of friction stir welded 2024-T4 joints.The effects of welding parameters, such as rotation rate and traverse speed, were considered in the interpretation of the results.

EBSD examination
To achieve quasi in situ observation of fatigue crack growth behavior of friction stir welded 2024-T4 joint, EBSD examinations were conducted to characterize the featured zones.The EBSD samples were polished by a mechanical polishing machine and then electro-polished in a solution (30 ml HClO4 and 270 ml CH3CH2OH).EBSD analysis were performed using the OXFORD NORDLYS X-MAX system and HKL-Channel5 software.The step size for the EBSD scan was 0.2 μm at stir zone and 0.3 μm at BM, AS and RS, which working at 20 kV with inclination angle of 70 degrees.

Microstructural characterization of weldment
The microstructural characterization of a friction stir welded 2024 joint is depicted in figure 2. Figure 2(a) illustrates visual insight into the microstructure of the joint.It reveals distinct zones that are formed during the friction stir welding process.The retreating side (RS), represented as zone 'b,' corresponds to the region where the tool moves away from the direction of welding.The stir zone (SZ), denoted as zone 'c,' is the central region where intense plastic deformation and mixing of material occur due to the rotating FSW tool.Lastly, the advancing side (AS), indicated as zone 'd,' is the region where the tool advances into the material.
In figure 2(b), the Stir Zone (SZ or Zone 'c') of the friction stir welded joint is depicted.The SZ is the central region where the most significant microstructural changes occur.These changes are primarily driven by the intense plastic deformation and mixing of material caused by the rotating friction stir welding (FSW) tool.One notable feature observed in the SZ is a finer grain structure compared to the base material.The severe plastic deformation experienced during FSW leads to grain refinement within this zone.Additionally, dynamic recrystallization can occur in the SZ, resulting in the formation of refined equiaxed grains [3].This process contributes to further grain refinement and the generation of a more uniform microstructure.The microstructural changes in the SZ also involve the disruption or breaking of the original grain boundaries.The severe plastic deformation and mixing action of the FSW tool disrupts the grain boundaries to be disrupted, leading to a more homogeneous microstructure within the SZ.Furthermore, the SZ exhibits a more uniform distribution of alloying elements compared to the base material.The intense mixing and shearing during FSW result in a more homogenous distribution of alloying elements throughout the SZ.
In the Retreating Side (RS) zone depicted in figure 2(c), the microstructure shows evidence of deformation, such as elongated grains or flow patterns.Within this region, the microstructure typically displays some degree of deformation.The thermal and mechanical effects of the welding cause the grains to undergo elongation or exhibit flow patterns.This deformation is a consequence of the plastic deformation imposed on the material as the FSW tool retreats during the welding process.The movement of the FSW tool away from the welding direction induces changes in the microstructure of the RS zone, primarily through the elongation and flow of the grains.These changes can be attributed to the thermal and mechanical forces exerted on the material during the welding process [16].
In figure 2(d), the Advancing Side (AS or Zone 'd') of the friction stir welded joint is depicted.Within the AS region, the material is displaced and pushed forward by the friction stir welding (FSW) tool.The microstructure in the AS zone exhibits some degree of deformation and flow patterns, but to a lesser extent compared to the RS zone.The plastic deformation and flow in the AS zone are relatively limited.The grains in this region are typically equiaxed in nature, closely resembling the grain structure of the base metal.Unlike the RS, the AS zone does not display prominent flow patterns resulting from plastic deformation and mixing [17].Instead, the microstructural changes in the AS zone are more subtle.This suggests that the microstructure in the AS zone undergoes fewer alterations compared to the RS and SZ.In summary, the microstructure of the Advancing Side (AS) in figure 3(d) exhibits limited deformation, with equiaxed grains resembling the base metal.The absence of prominent flow patterns further indicates that the microstructural changes in the AS zone are less pronounced compared to the Retreating Side (RS) and Stir Zone (SZ).

Fatigue crack growth behavior with different notch locations
Figure 3 shows the hardness distribution results of the friction stir welded 2024 joint.As can be seen from the figure, although the SZ region has fine grains, its hardness (∼120 HV) is significantly lower than that of the TMAZ (∼140 HV), HAZ (∼135 HV), and BM (∼135 HV) regions.This decrease in hardness is due to the dissolution of strengthening phases in the SZ during the welding process [11][12][13].Figure 4 compares the tensile properties of the base metal and the welded joint.It is clear that continuous yielding phenomenon occurred in both samples, but the mechanical properties changed largely after welding.The strength of the joint is approximately 70% of that of the BM, and the elongation is about 30% of that of the BM.From the microstructure results in figure 2 and the hardness results in figure 3, it is evident that the microstructure of the friction stir welded joint is uneven.Consequently, during tensile deformation, local stress tends to concentrate in the weaker SZ region and the overall specimen elongation decreased.
The fatigue crack growth behavior of friction stir weldments can be influenced by the location of a preexisting notch or crack tip within the weld structure.Different notch locations (see figure 5), such as at the base material (CT-BM), retreating side (CT-RS), advancing side (CT-AS), and stir zone (CT-SZ) can result in varying crack growth characteristics.Figure 5 displays the measured crack lengths plotted against the counted cycles for various cases.All cases utilized a maximum load per cycle of 1000 N. It is observed that, at the same cycle, the CT-BM sample exhibited a greater crack growth length compared to the other samples, indicating a decrease in crack growth resistance.Conversely, the CT-SZ sample demonstrated the highest crack growth resistance, as evidenced by the lower crack length at the same cycle number.
The results show that the fatigue crack growth (FCG) rate increases as the notch cycle (N) increases, which is a common behavior observed in fatigue cracks [18].The base material (CT-BM) has a shorter fatigue life compared to the welding zones (RS, SZ, and AS), indicating that the weld joint has effectively hindered the expansion of fatigue cracks.The smooth and continuous crack growth curve for CT-BM suggests a stable crack propagation behavior.In contrast, the curves for CT-SZ, CT-AS, and CT-RS exhibit noticeable fluctuations, which may be attributed to factors such as local stress concentrations, microstructural variations, or welding conditions specific to each zone.
The fatigue failure process involves various phases such as crack initiation, gradual fracture growth, and eventual collapse.The second phase can be subdivided into stages, but the initial stage of fracture formation in  shear is usually restricted and does not occur in welded joints.Interestingly, in welded joints, the initiation stage is often minimal or nonexistent.Instead, fatigue cracks in welded joints typically originate from pre-existing flaws, which can be seen as already initiated cracks.As a consequence, most of the fatigue life in welded joints is spent on the propagation phase of crack growth.This understanding underscores the importance of considering pre-existing flaws and focusing on crack propagation when assessing the fatigue behavior and lifespan of welded joints [19].
The stress intensity factor (ΔK) is a parameter that characterizes the stress conditions near the crack tip.This concept was introduced by Paris and Erdogan [20] to describe the state of stress adjacent to a crack.Consequently, since fatigue crack propagation (FCP) is influenced by these stresses, the stress intensity concept can also be utilized to quantitatively define crack propagation rates [21].Thus fatigue crack growth behavior was analyzed based on Paris and Erdogan work by utilizing power law i.e.Equations (1) and (2), whereas for C(T) samples, the ΔK is determined using equation (3).
Figure 6 illustrates the relationship between the fatigue crack growth rate and the corresponding stress intensity factor range for different samples.Due to the material's inhomogeneity, the crack growth data exhibited scattering behavior in all cases.The influence of the maximum load on the initial stress intensity factor range, as described in equation (3), was taken into account.The initial stress intensity factors had a minor impact on the crack propagation behavior.Regardless of the material, the crack growth behaviors were nearly identical, with only the steady crack growth region being clearly observed.
In the log-log curves [22], it was observed that the crack growth rate increased linearly with the stress intensity factor range.The steady crack growth behavior could be determined using equation (1).The exponent 'm' represents the slope of the line, while the parameter 'C' could be determined by the intercept of the fitting line.The values of 'm' and 'C' are listed in table 1.The value of 'm' is crucial in determining the fatigue crack resistance [23][24][25].The corresponding modeling equations for selected C(T) samples are shown as equations (5) to (8). ´- ´-

Examination of fractured surfaces
The EBSD Inverse Pole figure (IPF) images provide valuable insights into the grain size and distribution within the samples.In figure 7, the IPF map of the base metal and its microstructure at the tip of the crack after fatigue fracture is presented.The results of the analysis revealed a homogeneous grain size distribution with uniform orientation.This suggests that the base metal exhibited a consistent microstructure throughout, characterized by grains that are similar in size and aligned in a uniform manner.
In the case of the CT-SZ sample shown in figure 8, a significant observation is made regarding the grain size.The microstructure clearly exhibits a fine grain size in the stir zone, indicating that the material in this region consists of smaller and more densely packed grains [3].The reason for the enhanced resistance of the material to fatigue crack growth, as discussed earlier, is due to the presence of a finer microstructure.The fine-grained structure, characterized by smaller grains and a higher density of grain boundaries, serves as obstacles to the propagation of fatigue cracks.
On the other hand, the microstructure of the CT-AS sample shown in figure 9 and the CT-RS sample shown in figure 10 share similar features.In both the CT-AS and CT-RS samples, the cracks propagate from the prefabricated position to the BM and the grain structures display comparable characteristics in terms of grain size.This suggests that the that the fatigue crack growth resistance of the welding zone (SZ, AS, RS) is higher than that of BM.As one may have noticed, the log-log curves of CT-AS and CT-RS in figure 6 almost overlap, which  shows that the crack growth behavior of prefabricated cracks on the AS and RS is mainly related to the overall response of the joint during crack growth and has little to do with the prefabricated position.
The SEM analysis was used to examine the fracture morphology of the tested samples.Figures 11-13 present the surface appearance of fatigue fractures observed under SEM for CT-SZ, CT-AS, and CT-RS samples.The fractures in all samples exhibited visible initiation sites, suggesting that fatigue cracks originated from various locations.In figure 11, the fatigue cracks that cross through the SZ (zone 1 and zone 2 in figure 11) have more dimples, but they are also shallow.Figure 12 shows the fatigue crack propagation fracture morphology that initiate at AS.The crack initiates from zone 1 and extends to zone 2 (in figure 12).Although zone 1 in figure 12 is already located near BM shown in figure 8, there is a partial cleavage fracture area in zone 1.It suggests that the plasticity of this zone is lower than that of BM, which may be affected by the friction stir process of AS.Zone 2 in figure 12 shows significant deep and numerous dimples, which is the typical fracture morphology of BM.The results of the fracture morphology in figure 13 are similar to figure 12.This is consistent with the similar propagation behavior discussed above.
The fatigue fracture initiates at the tip of the milled notch and propagates across the stir zone (SZ) until failure occurs.A notable finding is that the weld area (zone-1) in the CT-SZ sample exhibits an exceptionally fine and homogeneous distribution of grains, also evident in figure 8.In comparison to the other samples, this fine and uniform distribution enhances the material's resistance to fatigue crack growth.The presence of these evenly dispersed, extremely small particulates act as barriers, hindering the growth of fatigue fractures and slowing down the rate of fatigue crack propagation [26].
It is worth mentioning that the fracture behavior of the CT-AS and CT-RS samples is similar.Consequently, their fatigue crack growth rates are comparable.In both cases, the crack initiates at the milled notch's tip and gradually extends towards the base metal (BM).This similarity in crack initiation and propagation behavior can be attributed to the similarities in the local stress conditions and microstructural characteristics present on the advancing and retreating sides of the joint.A comparison with the CT-SZ sample revealed that both CT-RS and CT-AS had numerous fracture initiation sites characterized by ridges, undulations, or waviness on their surfaces.In contrast, CT-SZ exhibited a smooth surface, indicating that CT-SZ had a longer fatigue fracture propagation life compared to the other joints.Ductile mode fracture was observed in all samples.However, the size and shape of the dimples observed in each joint differed, which can be attributed to variations in grain sizes, as depicted in their respective IPF's (Inverse Pole figure).

Conclusions
In conclusion, the quasi-in situ observation of the fatigue crack growth behavior in the friction stir welded 2024-T4 joint provided valuable insights into the microstructural characteristics and fatigue properties of the weldment.Four groups of samples notched at different regions, including the base material (BM), advancing side (AS), retreating side (RS), and stir zone (SZ), have been evaluated to compare fatigue crack growth behavior.EBSD and SEM have been utilized to achieve quasi-in situ observation of fatigue crack growth behavior in friction stir welded 2024-T4 joints.Following are the key findings observed in this research work: • The fine-grained structure in the stir zone significantly enhanced the joint's resistance to fatigue crack growth.
• Notch location played a crucial role in the fatigue crack growth behavior, with the sample notch positioned in the stir zone exhibiting the highest resistance, followed by notched samples from the advancing and retreating side weldments.
• The use of the Paris model enhanced the reliability and accuracy of the experimental results, allowing for a more comprehensive assessment of the joint's fatigue performance.• The joint exhibited improved fatigue crack growth resistance compared to the base metal, indicating the positive effect of the friction stir welding process on the joint's fatigue performance.

Figure 1 .
Figure 1.The macro morphology of FSWed 2024-T4 joint with schematic showing of CT samples.

Figure 2 .
Figure 2. The Inverse Pole figure (IPF) maps of friction stir welded 2024 joint: (a) the morphology, (b) the microstructure of SZ, (c) the microstructure of RS and (d) the microstructure of AS.

Figure 3 .
Figure 3. Hardness distribution of the FSW joint.

Figure 4 .
Figure 4. Strain-stress curves of base metal and FSW joints.

Figure 5 .
Figure 5.The relationship of fatigue crack length a and cycle index N.

Figure 6 .
Figure 6.FCG rate curves described by experimental data and Paris model.

Figure 7 .
Figure 7.The IPF map of the microstructure at the crack of BM after fatigue test.

Figure 8 .
Figure 8.The IPF map of the microstructure at the crack of SZ after fatigue test.

Figure 9 .
Figure 9.The IPF map of the microstructure at the crack of AS after fatigue test.

Figure 10 .
Figure 10.The IPF map of the microstructure at the crack of RS after fatigue test.

Figure 11 .
Figure 11.The morphologies of the crack of SZ after fatigue test.

Figure 12 .
Figure 12.The morphologies of the crack of AS after fatigue test.

Figure 13 .
Figure 13.The morphologies of the crack of RS after fatigue test.
B is the thickness of the C(T) sample, mm; W is the width of the C(T) sample, mm; m, C are the material parameters.

Table 1 .
Extracted values from the Paris model in figure 6.