Friction Stir Welding of Dissimilar Aluminum Alloy Combinations: State-of-the-Art

Friction stir welding (FSW) has enjoyed great success in joining aluminum alloys. As lightweight structures are designed in higher numbers, it is only natural that FSW is being explored to join dissimilar aluminum alloys. The use of different aluminum alloy combinations in applications offers the combined benefit of cost and performance in the same component. This review focuses on the application of FSW in dissimilar aluminum alloy combinations in order to disseminate research this topic. The review details published works on FSWed dissimilar aluminum alloys. The detailed summary of literature lists welding parameters for the different aluminum alloy combinations. Furthermore, auxiliary welding parameters such as positioning of the alloy, tool rotation speed, welding speed and tool geometry are discussed. Microstructural features together with joint mechanical properties, like hardness and tensile strength measurements, are presented. At the end, new directions for the joining of dissimilar aluminum alloy combinations should guide further research to extend as well as to improve the process, which is expected to raise further interest on the topic.


Introduction
Friction stir welding (FSW) is a solid state welding process which was invented at The Welding Institute (TWI) in UK in 1991 [1]. FSW is regarded as an environmentally friendly and energy efficient joining technique providing one of the best alternatives to fusion welding in order to produce a good combination of microstructure and properties in the joints. FSW has already proved its superiority in joining aluminum (Al) alloys as well as magnesium (Mg) alloys over fusion welding processes because of its solid-state nature. FSW uses a non-consumable rotating tool which has a shoulder and a pin (or more formally probe) at its end which plunges into the base material (BM) and advances in the welding direction [2], as shown in Figure 1. During the process, the shoulder touches the top surface of the BM and the pin moves yielded material around it. As a result of this action, heat is generated by frictional and plastic deformation of the BM by advancing the rotating tool. The shoulder of the tool has a forging action as it restricts the expulsion of plasticized material from the BM, while the pin extrudes material and produces a material flow between the advancing side (AS) and the retreating structural applications. More recently, Ma et al. [3] published a critical review paper on recent developments in FSW of Al alloys. Al alloys have remained the prime selection for structural material in aerospace, shipbuilding and automotive industries for their excellent strength to weight ratio. In order to improve performance while controlling the cost of Al alloys in these industries, there is an increasing demand to weld dissimilar Al joints with FSW. Because of the different physical and chemical properties in dissimilar Al alloy combinations, challenges such as solidification cracking, porosity, formation of intermetallic and so forth, are present. Therefore, the FSW of dissimilar Al alloy combinations has gained attention over the recent years, demonstrating the potential of the process to join these. The present review aims to discuss and analyze the available literature on FSWed dissimilar Al alloy combinations so far.

General Progress in FSW of Dissimilar Al-Al Combinations
There are review papers available on FSW of same Al alloy joints, which discuss various aspects of the process such as tool design, process parameters, heat generation, microstructure and mechanical properties [4][5][6][7][8][9][10][11]. The number of research papers on FSW of dissimilar Al alloy joints published to date is shown in Figure 2 (search on December 15, 2017 found 68 papers from Web of Science). The vast majority of the publications has been in the past 5 years, reaching a peak on 2018. In addition, Magalhães et al. [12] studied research and the extent of industrial application of FSW of similar and dissimilar material joints as shown in Figure 3. The similar material joints of Al alloys are being studied to a far larger extent compared to other alloys and the same trend is observed in the dissimilar material combinations. This trend observed literature clearly identifies the interest on the FSW of dissimilar Al alloy joints, which is expected to increase over the coming years. Figure 1. Schematic of friction stir welding, reproduced from [4], with permission from Elsevier, 2014.

General Progress in FSW of Dissimilar Al-Al Combinations
There are review papers available on FSW of same Al alloy joints, which discuss various aspects of the process such as tool design, process parameters, heat generation, microstructure and mechanical properties [4][5][6][7][8][9][10][11]. The number of research papers on FSW of dissimilar Al alloy joints published to date is shown in Figure 2 (search on 15 December 2017 found 68 papers from Web of Science). The vast majority of the publications has been in the past 5 years, reaching a peak on 2018. In addition, Magalhães et al. [12] studied research and the extent of industrial application of FSW of similar and dissimilar material joints as shown in Figure 3. The similar material joints of Al alloys are being studied to a far larger extent compared to other alloys and the same trend is observed in the dissimilar material combinations. This trend observed literature clearly identifies the interest on the FSW of dissimilar Al alloy joints, which is expected to increase over the coming years.    All papers from the top 10 ranked journals published on FSW, classified as Q1 by Scimago Journal & Country Rank.

Summary of Published Works
In order to identify the key findings on various aspects a summary of existing literature follows (Table 1). For the FSW of dissimilar Al alloy combinations there are the preliminary welding parameters such as the BM placement, the tool rotational speed and welding speed. The placement

Summary of Published Works
In order to identify the key findings on various aspects a summary of existing literature follows (Table 1). For the FSW of dissimilar Al alloy combinations there are the preliminary welding parameters such as the BM placement, the tool rotational speed and welding speed. The placement of the BM affects material flow, while rotational and welding speeds control heat input on both sides of the joint during welding. All of these parameters have been investigated for the different material combinations (see Table 1). In addition, the effects of welding parameters on the mechanical properties that is, the hardness and the joint strength have been investigated. As it can be seen a number of studies have been performed on the effect of the placement of BM (i.e., whether a particular material is placed on the AS or the RS side) on the material flow and the resulting microstructure in the SZ and the mechanical properties of the weld. Other papers have focused on the effect of tool geometry that is, shoulder diameter to pin diameter ratio and pin profile (cylindrical, conical, polygonal) on the microstructure and mechanical properties of the weld.

Positioning of Alloy
The placement of the alloy affects material flow as it strongly influences material stirring and mixing. This can be crucial in the final joint microstructure when the BM combination selected have significant differences in mechanical properties [69,70]. As the material flow during FSW is quite complex on its own, the placement of materials becomes an important parameter in welding, similar to the importance of the rotation and the welding speeds (see Table 1). For example Yan et al. [38] showed this for the Al-Zn-Mg and the Al-Mg-Si combination. There is an interesting material flow resistance behavior at the RS due to the difference in mechanical properties. When the Al-Zn-Mg alloy is placed at the AS, there was limited movement of the Al-Mg-Si alloy material to the AS side because of its stronger ability to flow as shown in Figure 4a. When the Al-Mg-Si was placed at the RS, there was no RS material (Al-Zn-Mg) flow to AS due to the strong resistance to flow by this high strength material as shown in Figure 4b. As it can be seen from Figure 4, the zig-zag line bonding interface formed due to excellent material mixing. The bonding interface may have vortex type in case of poor combination of rotational speed and welding speed and it becomes more prominent for BMs with significant difference in the properties. Niu et al. [71] investigated an AA2024-AA7075 joint and found that the top section of the SZ was composed of the BM of RS, whereas the middle and bottom sections by the BM of AS as shown in Figure 5. Kim et al. [65] also showed that by placing the high strength Al alloy on the AS generates excessive agglomerations and defects due to limited material flow. In essence, the high strength Al should be placed at the RS to minimize the effect of the resistance to material flow.

Positioning of Alloy
The placement of the alloy affects material flow as it strongly influences material stirring and mixing. This can be crucial in the final joint microstructure when the BM combination selected have significant differences in mechanical properties [69,70]. As the material flow during FSW is quite complex on its own, the placement of materials becomes an important parameter in welding, similar to the importance of the rotation and the welding speeds (see Table 1). For example Yan et al. [38] showed this for the Al-Zn-Mg and the Al-Mg-Si combination. There is an interesting material flow resistance behavior at the RS due to the difference in mechanical properties. When the Al-Zn-Mg alloy is placed at the AS, there was limited movement of the Al-Mg-Si alloy material to the AS side because of its stronger ability to flow as shown in Figure 4a. When the Al-Mg-Si was placed at the RS, there was no RS material (Al-Zn-Mg) flow to AS due to the strong resistance to flow by this high strength material as shown in Figure 4b. As it can be seen from Figure 4, the zig-zag line bonding interface formed due to excellent material mixing. The bonding interface may have vortex type in case of poor combination of rotational speed and welding speed and it becomes more prominent for BMs with significant difference in the properties. Niu et al. [71] investigated an AA2024-AA7075 joint and found that the top section of the SZ was composed of the BM of RS, whereas the middle and bottom sections by the BM of AS as shown in Figure 5. Kim et al. [65] also showed that by placing the high strength Al alloy on the AS generates excessive agglomerations and defects due to limited material flow. In essence, the high strength Al should be placed at the RS to minimize the effect of the resistance to material flow.  In the case of the lap joint, the BM placement affects the material flow and leads to the generation of the ubiquitous hook defect. Now the material movement is in an upward direction that is, from the bottom sheet to the top sheet, creating hook defects of various sizes. As expected, in addition to the rotation and welding speed, the placement of the BM affects the hook size as well [53,[72][73][74]. As it can be seen from Figure 6, the hook height is larger at the RS when the AA2024 is placed at the top, while it decreases when the AA7075 is placed as a top plate. In the case of the lap joint, the BM placement affects the material flow and leads to the generation of the ubiquitous hook defect. Now the material movement is in an upward direction that is, from the bottom sheet to the top sheet, creating hook defects of various sizes. As expected, in addition to the rotation and welding speed, the placement of the BM affects the hook size as well [53,[72][73][74]. As it can be seen from Figure 6, the hook height is larger at the RS when the AA2024 is placed at the top, while it decreases when the AA7075 is placed as a top plate.

Tool Rotation and Welding Speeds
Tool rotation and welding speeds control heat generation or heat input as they relate to the material plastic flow during FSW. The tool rotation speed affects the intensity of plastic deformation and through this affects material mixing. Kalemba-Rec et al. [16] showed a proportional relationship between material mixing and tool rotation speed for a dissimilar AA7075-AA5083 joint. However, very large rotation speeds lead to numerous imperfections such as poor surface (flash), voids, porosity, tunneling or formation of wormholes because of the excessive heat input [75][76][77], as shown  In the case of the lap joint, the BM placement affects the material flow and leads to the generatio the ubiquitous hook defect. Now the material movement is in an upward direction that is, from bottom sheet to the top sheet, creating hook defects of various sizes. As expected, in addition t rotation and welding speed, the placement of the BM affects the hook size as well [53,[72][73][74]. A an be seen from Figure 6, the hook height is larger at the RS when the AA2024 is placed at the top ile it decreases when the AA7075 is placed as a top plate.

. Tool Rotation and Welding Speeds
Tool rotation and welding speeds control heat generation or heat input as they relate to th terial plastic flow during FSW. The tool rotation speed affects the intensity of plastic deformatio d through this affects material mixing. Kalemba-Rec et al. [16] showed a proportional relationshi tween material mixing and tool rotation speed for a dissimilar AA7075-AA5083 joint. Howeve ry large rotation speeds lead to numerous imperfections such as poor surface (flash), void

Tool Rotation and Welding Speeds
Tool rotation and welding speeds control heat generation or heat input as they relate to the material plastic flow during FSW. The tool rotation speed affects the intensity of plastic deformation and through this affects material mixing. Kalemba-Rec et al. [16] showed a proportional relationship between material mixing and tool rotation speed for a dissimilar AA7075-AA5083 joint. However, very large rotation speeds lead to numerous imperfections such as poor surface (flash), voids, porosity, tunneling or formation of wormholes because of the excessive heat input [75][76][77], as shown in Figure 7. Low welding speeds increase heat input and are associated with defects like tunneling [55,58,75,78,79]. It is therefore necessary to select the appropriate combination of tool rotation and welding speed for a defect free joint with a good metallurgical bond and mechanical properties. As it can be seen in Table 1, quite a lot of papers have focused on the optimization of these parameters for different combinations of Al alloys [23,29,32,33,35,36,41,42,50,54,55,58,64,80]. welding speed for a defect free joint with a good metallurgical bond and mechanical properties. As it can be seen in Table 1

Tool geometry
The geometry of the shoulder and the pin profile govern heat generation and material flow during welding [81]. The shoulder contributes to a large extent to heat input due to its size. The common shoulder profiles employed are the flat, the concave and the convex. Additional features on the pin such as a spiral or a groove improve frictional behavior as well as material flow. Palanivel et al. [18] reported on the effect of shoulder profiles on the AA5083-AA6351 combination by using three different shoulder features, the partial impeller (PI), the full impeller (FI) and the flat grove (FS) as shown in Figure 8. The full impeller shoulder tool produced the optimum mechanical strength due to the enhanced material flow it produced. The pin profile greatly affects material stirring and mixing. Cylindrical or conical pin profiles which may have features like threads or threads with flats have been used for dissimilar Al alloy combinations as shown in Figure 8. When used without threads a smaller surface is provided to the material, while the threaded and flat features on it increase the contact area while threads guide material flow around the pin in a rotational as well as a translation direction [14,16,47,82]. The polygonal pin profiles produce pulses in the flow during material stirring and mixing, leading to material adhering to the pin [83][84][85][86]. This pulsating effect hinders material flow significantly in the case of dissimilar Al alloy combinations. It is therefore recommended to use a cylindrical or a conical pin profile with various features in the dissimilar Al alloy joints for good material flow to produce sound joints.  [16], with permission from the authors.

Tool Geometry
The geometry of the shoulder and the pin profile govern heat generation and material flow during welding [81]. The shoulder contributes to a large extent to heat input due to its size. The common shoulder profiles employed are the flat, the concave and the convex. Additional features on the pin such as a spiral or a groove improve frictional behavior as well as material flow. Palanivel et al. [18] reported on the effect of shoulder profiles on the AA5083-AA6351 combination by using three different shoulder features, the partial impeller (PI), the full impeller (FI) and the flat grove (FS) as shown in Figure 8. The full impeller shoulder tool produced the optimum mechanical strength due to the enhanced material flow it produced. The pin profile greatly affects material stirring and mixing. Cylindrical or conical pin profiles which may have features like threads or threads with flats have been used for dissimilar Al alloy combinations as shown in Figure 8. When used without threads a smaller surface is provided to the material, while the threaded and flat features on it increase the contact area while threads guide material flow around the pin in a rotational as well as a translation direction [14,16,47,82]. The polygonal pin profiles produce pulses in the flow during material stirring and mixing, leading to material adhering to the pin [83][84][85][86]. This pulsating effect hinders material flow significantly in the case of dissimilar Al alloy combinations. It is therefore recommended to use a cylindrical or a conical pin profile with various features in the dissimilar Al alloy joints for good material flow to produce sound joints.

Microstructure Evolution
The typical microstructure of a FSW joint consists of three distinct zones that is, HAZ, TMAZ and SZ [87,88]. These zones form depending on the thermal and mechanical deformation that the tool induces during welding. The SZ undergoes extensive grain refinement, producing fine grain microstructures, while the TMAZ has an elongated grain structure [89,90]. The microstructure evolution depends on the welding parameters (as discussed in the previous section), as the material movement or flow plays a more important role in the case of dissimilar material combinations compared to same material joints. The appropriate selection of all process parameters results in excellent material mixing on the both sides (AS and RS) of the joint and produces a sound weld. Recently, a comprehensive EBSD investigation for the AA5083-AA2024 joint was reported by Niu et al. [91], as shown in Figure 9. As it can be seen from the EBSD orientation maps (Figure 9 a-d), tilted and elongated grains in the TMAZ and fine grains d in the SZ developed due to dynamic

Microstructure Evolution
The typical microstructure of a FSW joint consists of three distinct zones that is, HAZ, TMAZ and SZ [87,88]. These zones form depending on the thermal and mechanical deformation that the tool induces during welding. The SZ undergoes extensive grain refinement, producing fine grain microstructures, while the TMAZ has an elongated grain structure [89,90]. The microstructure evolution depends on the welding parameters (as discussed in the previous section), as the material movement or flow plays a more important role in the case of dissimilar material combinations compared to same material joints. The appropriate selection of all process parameters results in excellent material mixing on the both sides (AS and RS) of the joint and produces a sound weld. Recently, a comprehensive EBSD investigation for the AA5083-AA2024 joint was reported by Niu et al. [91], as shown in Figure 9. As it can be seen from the EBSD orientation maps (Figure 9a-d), tilted and elongated grains in the TMAZ and fine grains d in the SZ developed due to dynamic recrystallization. Grain boundary orientations also varied in all three zones as shown in Figure 9(e-h). A higher fraction of large (>10 • ) angular grain boundaries was present in the SZ, while more of low (2-10 • ) angular grain boundaries were present in HAZ. Also, a more intense texture in the SZ was formed compared to other zones.
Metals 2019, 9, x FOR PEER REVIEW 13 of 21 recrystallization. Grain boundary orientations also varied in all three zones as shown in Figure 19 (eh). A higher fraction of large (> 10°) angular grain boundaries was present in the SZ, while more of low (2-10°) angular grain boundaries were present in HAZ. Also, a more intense texture in the SZ was formed compared to other zones.

Hardness
The hardness of the FSW joint is related to the joint strength and its deformation behavior, especially in the case of dissimilar material combinations. The hardness distributions of various different Al alloy combinations are shown in Figure 10. The common highly asymmetrical hardness distribution along the cross-section of dissimilar material joints is due to the different microstructural zones (SZ, TMAZ, HAZ) which develop due to the thermo-mechanical history during welding. Since the maximum temperature is reached at the SZ, precipitates or strengthening particles dissolve partially or completely decreasing hardness in SZ. Whereas the lowest hardness values are found in the HAZ due to the coarsening of precipitates or over aging. Therefore, the HAZ always remains the most common zone or site where failure occurs during tensile deformation. It is also worth noting that SZ has higher hardness values compared to the BM (which may be of low strength) because of the combined effect of grain refinement and the effect of both of the BMs in the SZ. However, it is not always true due to different initial conditions of heat-treatable alloy combinations. Recently, Niu et al. [13] reported an interesting hardness behavior of joints prior to and following fracture, by quantifying hardening with the ratio of HVf/HVw, where HVf and HVw are the microhardness of the fractured and the as-welded joints, respectively. This ratio was over one in the SZ, TMAZ and HAZ, which confirmed the strain hardened behavior of the joints as shown in Figure 11. In summary,

Hardness
The hardness of the FSW joint is related to the joint strength and its deformation behavior, especially in the case of dissimilar material combinations. The hardness distributions of various different Al alloy combinations are shown in Figure 10. The common highly asymmetrical hardness distribution along the cross-section of dissimilar material joints is due to the different microstructural zones (SZ, TMAZ, HAZ) which develop due to the thermo-mechanical history during welding. Since the maximum temperature is reached at the SZ, precipitates or strengthening particles dissolve partially or completely decreasing hardness in SZ. Whereas the lowest hardness values are found in the HAZ due to the coarsening of precipitates or over aging. Therefore, the HAZ always remains the most common zone or site where failure occurs during tensile deformation. It is also worth noting that SZ has higher hardness values compared to the BM (which may be of low strength) because of the combined effect of grain refinement and the effect of both of the BMs in the SZ. However, it is not always true due to different initial conditions of heat-treatable alloy combinations. Recently, Niu et al. [13] reported an interesting hardness behavior of joints prior to and following fracture, by quantifying hardening with the ratio of HV f /HV w , where HV f and HV w are the microhardness of the fractured and the as-welded joints, respectively. This ratio was over one in the SZ, TMAZ and HAZ, which confirmed the strain hardened behavior of the joints as shown in Figure 11. In summary, hardness distribution in the dissimilar material joints is closely associated with mechanical behavior such as strain hardening and the fracture origin. (c) AA2024-AA5083 [91] (d) AA2219-AA5083 [60] (e) AA6061-AA7075 [46] (f) AA6061-AA7075 T6 [51]

Tensile Strength
The number of published papers investigating welding the 5xxx-6xxx series alloys to identify the effect of process parameters (especially the tool rotation speed and welding speed) on the joint strength is shown in detail in Table 1. The joint strength increases with the rotation speed due to the enhanced material mixing effect [18,54,57,62]. The tool rotation speed intensifies plastic deformation and welding speed controls the thermal cycle, residual stresses and rate of production. So, it is (a) AA6061-AA1050 [42]; with permission from the authors. (b) AA2024-AA6061, reproduced from [22], with permission from Elsevier, 2108; (c) AA2024-AA5083 reproduced from [91], with permission from Elsevier, 2019; (d) AA2219-AA5083 reproduced from [60], with permission from Elsevier, 2012; (e) AA6061-AA7075 reproduced from [46], with permission from Elsevier, 2015; (f) AA6061-AA7075 T6], reproduced from [51], with permission from Springer, 2014. generation between the tool and the BM. The conical threaded pin was identified as the best possible configuration for the AA 6061-AA5086 joint due to the production of a uniformly distributed precipitates and the distinct generation of the onion rings as material was mixed appropriately in the SZ, as reported by Ilangovan et al. [47]. In summary, the tensile strength of the dissimilar FSWed Al joints relies on the microstructure evolution during FSW, which in turn depends on the heat input as governed by the welding parameters (as discussed in Section 4).

Tensile Strength
The number of published papers investigating welding the 5xxx-6xxx series alloys to identify the effect of process parameters (especially the tool rotation speed and welding speed) on the joint strength is shown in detail in Table 1. The joint strength increases with the rotation speed due to the enhanced material mixing effect [18,54,57,62]. The tool rotation speed intensifies plastic deformation and welding speed controls the thermal cycle, residual stresses and rate of production. So, it is essential to select the appropriate combination of these speeds for weld quality or joint strength. Bijanrostami et al. [33] investigated the AA6061-AA7075 joint to identify that maximum joint strength is achieved with a combination of moderate rotation and low welding speed. When high heat input conditions are used (i.e., high rotation and low welding speeds) large grains and lower dislocation densities develop in the SZ. On other side when low heat input condition are selected (i.e., low rotation and high welding speeds) defects are generated. So, grain size strengthening and low dislocation densities are necessary for joint strength. However, the maximum joint strength of an A356-AA6061 joint was achieved with low rotation and welding speed by Ghosh et al. [58,64]. Evidence of fine grain size, fine distribution of Si particles and reduced residual stresses in the SZ were found for low rotation and welding speeds. Together with rotation and welding speeds, the effect of tool geometry like the pin profile or features [14,18,47,48,62], pin shapes [34,57] and shoulder diameter to pin diameter ratio [37,60] on joint strength have been investigated. The pin profile or feature controls material flow and in effect material mixing at the joint interface, the pin shape affects SZ size as well as material movement and the shoulder to pin diameter ratio controls frictional heat generation between the tool and the BM. The conical threaded pin was identified as the best possible configuration for the AA 6061-AA5086 joint due to the production of a uniformly distributed precipitates and the distinct generation of the onion rings as material was mixed appropriately in the SZ, as reported by Ilangovan et al. [47]. In summary, the tensile strength of the dissimilar FSWed Al joints relies on the microstructure evolution during FSW, which in turn depends on the heat input as governed by the welding parameters (as discussed in Section 4).

Summary and Outlook
With regards to the research published and the appropriate future work to be performed in the FSW of dissimilar Al alloy combinations, the following comments can be proposed:

Al Alloy Combinations
Almost all of the investigations conducted concerned BM in the as-rolled condition that is, 2xxx-5xxx, 2xxx-6xxx, 2xxx-7xxx, 5xxx-6xxx, 5xxx-7xxx Al series. It would be interesting to explore dissimilar Al alloy combinations in as-cast conditions and as a combination between as-cast and as-rolled conditions, depending on the application.

Base Metal Placement
Limited number of papers on the effect of placement is available and still remains inconclusive. Base material placement becomes an issue in the cases where there are significant differences in mechanical properties of the BMs as in the 6xxx-7xxx and the 5xxx-7xxx combinations.

Tool Offset
There is a very limited number of welding parameters optimization studies to study tool offset. It needs further comprehensive evaluation using microstructure characterization to understand the material flow in the SZ.

Bobbing Tool and Stationary shoulder Tool
The bobbin tool [92] and the stationary shoulder tool are considered as a strategic variant of FSW, which have distinct benefits over the conventional FSW tool. Stationary shoulder tool offers low heat input during welding and processing [93][94][95] and would benefit Al alloy dissimilar joints [96].

Corrosion and Fatigue Behavior
Finally, corrosion and fatigue behavior studies of various combinations of dissimilar Al alloy joints would be beneficial to expand its industrial use.

Funding:
The authors would like to thank for financial support National Natural Science Foundation of China (51574196, U1637601). We would also like to express our gratitude to the Guest Editor, Antonello Astarita (University of Naples Federico II, Italy) for the invitation to contribute this article to METALS.

Conflicts of Interest:
The authors declare no conflict of interest.