*2.1. Friction Stir Welding of Similar and Dissimilar Alloys*

Recent studies have revealed that the material positioning during FSW of dissimilar plays an important role towards the strength of the weld. The good weld is produced when the hardest material is positioned on the retreating side while keeping the softer one on the advancing side during welding [7,8]. In an attempt towards analyzing the impact of material positioning during FSW dissimilar alloys, various studies have been performed in this regard. Dilip et al. [9] have performed the FSW of AA2219-T87 and AA5083-321 with the aim of performing the microstructural analysis of the joint. The weaker material (AA2219-T87) was positioned on the advancing side while the stronger material (AA5083-H321) was kept on the retreating side. The microstructural analysis revealed that the joint and the retreating side were dominated by the material which was placed on the advancing side (AA2219-T87). The microhardness value corresponding to the weaker material was observed on the retreating side where most tensile failure occurred.

Friction stir welding of the 3 mm thick AZ31B magnesium alloy and AA5052-H32 was performed by Taiki et al. [10]. The aluminium plate was positioned on the advancing side and magnesium (Mg) plate on the retreating side during welding. There was a variation in welding speed and tool speed. The microstructure analysis revealed that the joint was dominated by the AA5052-H32. It was also noted from the microstructural analysis that the dominating AA5052 had refined grains compared to parent material although the hardness value dropped compared to AA5052-H32 base metal. Hardness distributions of the cross-section revealed that the intermetallic compounds (IMCs) partly existed in the stir zone (SZ). All the samples failed at the center of the joint during tensile tests analysis. This failure location showed that the joint was dominated by the material that was positioned on the advancing side during welding.

Cavaliere & Panella, [11] conducted a study on the effect of tool position on fatigue properties of dissimilar 4 mm thick AA2024 and AA7075 plates joined by FSW. The AA2024 was positioned on the advancing side while AA7075 was kept on the retreating side. The joint attained when the tool was positioned 1 mm off the center (towards AA7075) had a higher hardness value compared to the joint attained when tool was 1.5 mm off the center of the weld. The maximum tensile properties of both joints were lower than the parent materials. Both joints revealed a ductile failure mode characterized by the presence of very fine dimples. The strong effect on fatigue crack growth was attributed to the positive fracture resistance (Kr) value measured on the cross-section of the different welds.

Peng et al. [12] performed friction stir welded on the 5 mm AA5A06-H112 and AA6061-T651 plates. This welding was performed under controlled cooling conditions i.e., forced air cooling (FAC) and natural cooling (NC) conditions. The AA5A06-H112 was positioned on the advancing side while AA6061-T651 was kept on the retreating side. The 0.5 MPa pressure was used to blow the air towards the welding direction which then intersected the surface of materials at angle of 30◦. The microstructural analysis and microhardness test results for the joint produced under FAC were found to be higher compared to those that of the joint produced under NC condition. The tensile results for the joint produced under FAC condition were 10% higher than those produced at NC condition. The joint produced under NC condition had coarser grains compared to the joint produced under FAC condition. Both joints had ductile failure mode but the dimple size for the joint produced under FAC were higher than the joint produced under NC condition.

Shah et al. [13] investigated the influence of the tool eccentricity towards the friction stir welded dissimilar metals joint quality. They discovered that placing the stronger material on the advancing side improves the tensile strength and the percentage elongation of the joint. Their metallurgical analysis revealed that the tool eccentricity also plays a vital role towards the material flow however, there are some limitations when it comes to material mixing. The analysis of the joint formed when two dissimilar alloys are used in friction stir welding normally focuses on the mechanical properties. However, Giraud et al. [14] have gone to the extent of analyzing the compounds that are being formed during the FSW of dissimilar alloys. They have discovered that there are intermetallic compounds (IMCs) that are formed during the FSW of dissimilar alloys. These IMCs have a brittle nature which could lead to greater mechanical weakness.

Khodir and Shibayanagi [15] assessed the joint formed when AA2024-T3 was friction stir welded with AZ31 magnesium alloy. Their study involved the variation of welding speed at a constant rotational speed. The AA2024-T3 was located on the advancing side for all the welding. The microstructural analysis revealed that the increase in welding speed impacted the phase redistribution in the stir zone. The AA2024-T3 was distributed towards the lower regions of the stir zone while the AZ31 dominated the upper regions below the tool shoulder of the stir zone. The microstructural analysis also revealed a consistent formation of laminates structures in the SZ near the advancing side boundary between SZ and thermal affected zone (TMAZ) which were independent from welding speed variation. There were also intermetallic compounds that were formed in the SZ which contributed towards the fluctuation of the hardness distribution.

Rodriguez et al. [16] have friction stir welded AA6061-T6 and AA7050-T7451 with the purpose of assessing the microstructure and mechanical properties of the dissimilar welded joint. Their study involved the variation of rotational speed while keeping the welding speed constant. The AA7050-T7451 was positioned on the advancing side while AA6061–T6 was kept on the retreating side during welding. Tensile analysis revealed that the joints produced at lower speed were weaker than the base metals hence the fracture occurred at the SZ. The joints that were produced at higher rotational speed were stronger than AA6061-T6 base metal hence the fracture occurred consistently towards the AA6061-T6. The variation in fracture location was found to be directly linked with the material mixing at the SZ. The microstructural analysis revealed the ductile mode of failure. Moreover, the energy dispersive X-ray spectroscopy (EDS) results reveal the existence of three distinct layers where layer 1 had a nominal composition of AA6061, layer 2 had a composition of AA7050 and layer 3 had the combination of the two. Similar results were reported by Gou et al. [17] when they performed FSW on dissimilar AA6061-AA7075.

Mofid et al. [18] performed a study on the friction stir welding of the 3-mm thick AZ31C-O magnesium alloy to AA5083 in air and under nitrogen liquid. Their study involved the tracking of the temperature profile during welding and they attained this through the installation of thermocouples. There was a notable decrease in IMCs formation for the joints produced under liquid nitrogen compared to joints produced through air. The X-ray diffraction (XRD) analysis results exhibited the intermetallic phases of Al3Mg2, Al12Mg17 and Al2Mg3. The stir zone of the welds produced under nitrogen atmosphere showed a smoother interface compared to welds produced through air atmosphere. The attained maximum temperature during the welding was 676 K and 651 K respectively during air weld and under water weld.

Friction stir welding of AA2024-T365 and AA5083-H111 was performed by El-Hafez and El-Megharbel [19]. Variation in process parameters and pin profiles were employed with the purpose of analyzing their influence to the microstructure and tensile properties. The stronger material (AA2024-T365) was positioned on the advancing side throughout the welding. The combination of the highest speeds of 1120 rpm and 1400 rpm with 80 mm/min achieved the best strength and joint efficiency of 90% and this was due to sufficient heat being generated. Square pin profile produced higher strength joints compared to triangular and stepped profiles. Locating AA2024 on the advancing side (AS) played a significant role towards joint strength improvement. Cole et al. [20] also reported that the material placed on the advancing side dominates a major portion of the weld zone.

Vivekanandan et al. [21] used vertical milling machine for the friction stir welding of AA6035 and AA8011 with the aim of evaluating the mechanical properties of the dissimilar weld joint. The varying welding speed at a constant rotational speed was employed throughout the welding. The welds produced at the welding speed of 60 mm/min were found to be the best results compared to other speed combinations. This parameter combination produced fine grains at the center of the weld which contributed to the increase in hardness value. The dissimilar friction stir welding of undiluted copper and AA1350 sheet with a thickness of 3 mm was investigated by Li et al. [22]. The AA1350 was placed on the advancing side throughout the welding performance. The microstructural results in the nugget zone showed the vortex-like pattern and lamella structure. There was no formation of IMCs in the nugget zone. The hardness dispersion revealed that the hardness on the copper side was higher than that on the AA1350 side and the hardness at the bottom of the nugget was generally higher than those previously mentioned. The tensile properties of the dissimilar welds were all lower than those of the base metals. A ductile-brittle mixed fracture surface was observed on the dissimilar joints of the tensile tested specimens.

Friction stir welding was applied on the 1.3 mm thick stiffened AA2024-T3 panels with the aim to analyze the crack growth behaviour [23]. The experimental tests were correlated to linear elastic finite element method and dual boundary element method (DBEM). It was found that the DBEM showed better results and accurate as the stress level increased as the crack was approaching the stiffener. A Similar study was conducted by Citarella et al. [24] using a hybrid technique to assess the fatigue performance of multiple cracked friction stir welded AA2024-T3 joints. The crack propagation experimental tests were evaluated using the contour method in order to analyze the distribution of the residual. The metallographic analysis results showed a visible initial defect which led to initial crack for the simulation. The experimental fracture surface confirmed the crack propagation. The numerical crack was comparable to fatigue area shown by the post- mortem fractography.

Sheng et al. [25] used friction stir welding technique to join the AA6005—T4 plates with the purpose of investigating the weldability, microstructure and mechanical properties of the said alloy. The microstructural analysis results showed recrystallized grains in the nugget zone with equiaxed grain sizes of about 2.2 μm. The maximum ultimate tensile strength of about 174 MPa equivalent to 83.8% of the base material was obtained. A microhardness was reduced to 58 HV0.2 by the dissolution of phase β. In another study, an impact of using a bobbin type tool in friction stir welding of AA6082-T6 plates at different rotational speeds was investigated [26]. This investigation involved the variation of tool rotational speed. The tensile strength was found to be increasing linearly with rotational speed. However, it was discovered that the tensile strength reached its maximum when the rotational speed of 800 rpm was employed. The tensile strain of 7.9% was achieved at the same rotational speed. However, the strength and hardness were found to be having an inverse relationship with the increment of heat input at the speed beyond 800 rpm. The fractographic results showed a dimple fracture with white second phase particles of AlFeMnSi.

The 6 mm thick sheets of AA6061 and AA5086 were friction stir welded together to analyze the evolution of microstructure in the stir zone and its influence on tensile properties of the joints [27]. The welding parameters used were the rotational speed of 500 rpm, traverse speed of 35 mm/min and the axial force of 4.9 kN. The tensile properties of the joints correlated with microstructural features and microhardness values. The dissimilar joint exhibited a maximum hardness of 115 HV and a joint efficiency of 56% which was higher than the hardness of the base metals. This was attributed to the defect-free stir zone formation and grain size strengthening. Table 1 below give a tabulated review of the above literature. The idea behind the incorporation of this table is to show the typical positioning of the materials during welding of dissimilar materials and alloys. Table 1 also shows the mostly used tool material and tool profile in performing welding of dissimilar alloys/materials.

**Table 1.** Friction stir welding of dissimilar materials/alloys (RS—Retreating side, AS—Advancing side, SZ—Stir zone, TRS—Tool rotational speed, WS—Welding speed, El—Elongation, YS—Yield strength, JE—Joint efficiency, NS—Not specified.).

