The Inﬂuence of Multiple Pass Submerged Friction Stir Processing on the Microstructure and Mechanical Properties of the FSWed AA6082-AA8011 Joints

: The AA6082–AA8011 friction stir-welded joints were subjected to submerged multiple pass friction stir processing to evaluate the microstructure and mechanical properties of the joints. A maximum of four submerged friction stir processed passes were used in this study. All the specimens were extracted from three di ﬀ erent joint positions (start, middle and end). The tests conducted included microstructural analysis, tensile tests, hardness and fracture surface morphology of the post-tensile specimens, were performed using a scanning electron microscope (SEM). There was no particular trend in the microstructure and mechanical properties when looking at the specimen positioning in all the passes. The minimum mean grain sizes were reﬁned from 3.54 to 1.49 µ m and the standard deviation from 5.43 to 1.87 µ m. The ultimate tensile strength was improved from 84.96 to 94.77 MPa. The four-pass SFSPed specimens were found to have more ductile properties compared to the one-pass SFSPed one. The hardness of the stir zones in all the passes was found to be higher compared to the AA8011 base material but lower than the AA6082 one. The maximum stir zone hardness of 75 HV was observed on the one-pass SFSP joints.


Introduction
Friction stir processing (FSP) is an advanced microstructural altering process based on the friction stir welding process [1,2]. The grain refinement is a result of the dynamic recrystallization that occurs during the process. The FSP technique does not only refine the microstructural grain size but also modifies the texture of the processed zone [3,4]. FSP has been successfully used in many applications including the fabrication of surface composites of aluminium substrates [5,6], for superplastic high strain rate [7,8], metal matrix composites [9,10] and cast aluminium alloys [11,12]. FSP as an enticing emerging technology continues to grow, however, due to the high temperature the material experiences during FSP, grain growth was found to occur. Submerged friction stir processing (SFSP) came as a solution to minimize grain growth in order to achieve more grain refinement [13]. The SFSP technique works similarly to the normal or rather traditional FSP, the only difference takes place under controlled immersed environments [14,15].
Several studies have investigated the effect of submerged friction stir processing on aluminium alloys. Feng et al. [16] evaluated the impact of submerged friction stir processing on the microstructure of the AA2219-T6 plate. The processing conditions included the traverse speed of 200 mm/min and varying rotational speed. The results showed that the hardness of the stir zone was found to decrease with an increase in rotational speed. The decrease was substantiated to be due to softening of the Table 1. Chemical compositions [28,29].  shown in Figure 1c were applied to the FSW joint. The same tool and parameters used for friction stir welding were also used for friction stir processing. Table 1. Chemical compositions [28,29].      The friction stir-processed plates (FSPed) were cut for different tests using waterjet cutting technology. The tests conducted were the tensile test, microhardness, microstructural analysis and fatigue test. The Hounsfield 50 K tensile testing machine was used to perform the tensile tests of Metals 2020, 10, 1429 4 of 14 the specimens. The tensile specimen geometry and operating parameters were adapted from the ASTM-E8M-04 standard. Figure 2a depicts the tensile test specimen used in this study. The fractured tensile specimens were further analysed for the nature of fracture and morphology using a Tescan MIRA3 RISE scanning electron microscope (TESCAN Orsay Holding, Kohoutovice, Czech Republic). The InnovaTest Falcon 500 hardness testing machine (INNOVATEST Europe BV Manufacturing, Maastricht, The Netherlands) was used to perform the Vickers hardness test following the ASTM E384-11 standard. Figure 2b presents the specimen used for microhardness testing. The objective 10× and objective 20× for specimen focusing were used during setup. A 0.5 kg load and 1 mm interval was applied from the centre to either side of the specimen (advancing and retreating). A single line pattern was used to perform the hardness tests. The friction stir-processed plates (FSPed) were cut for different tests using waterjet cutting technology. The tests conducted were the tensile test, microhardness, microstructural analysis and fatigue test. The Hounsfield 50 K tensile testing machine was used to perform the tensile tests of the specimens. The tensile specimen geometry and operating parameters were adapted from the ASTM-E8M-04 standard. Figure 2a depicts the tensile test specimen used in this study. The fractured tensile specimens were further analysed for the nature of fracture and morphology using a Tescan MIRA3 RISE scanning electron microscope (TESCAN Orsay Holding, Kohoutovice, Czech Republic). The InnovaTest Falcon 500 hardness testing machine (INNOVATEST Europe BV Manufacturing, Maastricht, The Netherlands) was used to perform the Vickers hardness test following the ASTM E384-11 standard. Figure 2b presents the specimen used for microhardness testing. The objective 10× and objective 20× for specimen focusing were used during setup. A 0.5 kg load and 1 mm interval was applied from the centre to either side of the specimen (advancing and retreating). A single line pattern was used to perform the hardness tests. The microstructural analysis was performed using the Motic AE2000 metallurgical microscope (Motic Europe S.L.U., Barcelona, Spain). The specimens were mounted, ground, polished and etched using the modified Keller's and Weck's agents. The modified Keller's reagent chemical composition was 10 mL nitric acid (HNO3), 1.5 mL hydrochloric acid (HCl), 1.0 mL hydrofluoric acid (HF) and The microstructural analysis was performed using the Motic AE2000 metallurgical microscope (Motic Europe S.L.U., Barcelona, Spain). The specimens were mounted, ground, polished and etched using the modified Keller's and Weck's agents. The modified Keller's reagent chemical composition was 10 mL nitric acid (HNO 3 ), 1.5 mL hydrochloric acid (HCl), 1.0 mL hydrofluoric acid (HF) and 87.5 mL distilled water (H 2 O) and the Weck's reagents composition was 1 g sodium hydroxide (NaOH), 4 g potassium permanganate (KMnO 4 ) and 100 mL distilled water (H 2 O). The ASTM E112-12 standard was used to determine average grain size through the use of ImageJ software. Figure 2c shows the specimen positioning used in this study where "S" represents specimens extracted from the start, "M" is used for specimens extracted from the middle, and "E" for specimens extracted towards the end of the processed area. Figure 3 shows the microstructural grain sizes of the multiple passes of submerged friction stir-processed FSWed joints. Table 4 presents the grain sizes measured concerning Figure 3. The 1-pass friction stir-processed FSWed joints had a mean grain size range of 13.17 to 16.51 µm, minimum grain size range of 3.29 to 3.63 µm and a standard deviation range of 4.32 to 6.42 µm. Similar results were reported in the literature [17]. The 2-pass friction stir-processed FSWed joints had a mean grain size range of 7.03 to 7.51 µm, standard deviation range of 2.12 to 3.56 µm and minimum grain size of 3.04 to 4.12 µm. The second pass had a mean grain size range of 5.61 to 6.36 µm, a standard deviation range of 2.03 to 2.61 µm and minimum grain size range of 2.54 to 3.36 µm. The fourth pass had a mean grain size range of 5.03 to 5.38 µm, a standard deviation range of 1.65 to 2.18 µm and a minimum grain size ranges of 1.20 to 1.86 µm. The minimum grain size, mean grain size and the standard deviation decreased with an increase in the number of SFSP passes. The mechanism behind the grain refinement is based on the high plastic deformation and repeated dynamic re-crystalization that occurred during the SFSP process. Furthermore, the post-grain growth during the SFSP is also prevented by the removal of excess frictional heat as a result of rapid water cooling.

Microstructural Analysis
Metals 2020, 10, x FOR PEER REVIEW 5 of 14 87.5 mL distilled water (H2O) and the Weck's reagents composition was 1 g sodium hydroxide (NaOH), 4g potassium permanganate (KMnO4) and 100 mL distilled water (H2O). The ASTM E112-12 standard was used to determine average grain size through the use of ImageJ software. Figure 2c shows the specimen positioning used in this study where "S" represents specimens extracted from the start, "M" is used for specimens extracted from the middle, and "E" for specimens extracted towards the end of the processed area. Figure 3 shows the microstructural grain sizes of the multiple passes of submerged friction stirprocessed FSWed joints. Table 4 presents the grain sizes measured concerning Figure 3. The 1-pass friction stir-processed FSWed joints had a mean grain size range of 13.17 to 16.51 µm, minimum grain size range of 3.29 to 3.63 µm and a standard deviation range of 4.32 to 6.42 µm. Similar results were reported in the literature [17]. The 2-pass friction stir-processed FSWed joints had a mean grain size range of 7.03 to 7.51 µm, standard deviation range of 2.12 to 3.56 µm and minimum grain size of 3.04 to 4.12 µm. The second pass had a mean grain size range of 5.61 to 6.36 µm, a standard deviation range of 2.03 to 2.61 µm and minimum grain size range of 2.54 to 3.36 µm. The fourth pass had a mean grain size range of 5.03 to 5.38 µm, a standard deviation range of 1.65 to 2.18 µm and a minimum grain size ranges of 1.20 to 1.86 µm. The minimum grain size, mean grain size and the standard deviation decreased with an increase in the number of SFSP passes. The mechanism behind the grain refinement is based on the high plastic deformation and repeated dynamic re-crystalization that occurred during the SFSP process. Furthermore, the post-grain growth during the SFSP is also prevented by the removal of excess frictional heat as a result of rapid water cooling.   The microstructural grains during the SFSP process reduced the grain growth and the migration rate of grain boundaries, which then led to fine equiaxed grain structure in the stir zone of the joint [31][32][33]. Additionally, Lou et al. [34] proved that during the multi-pass SFSP, the processed surface undergoes an enhanced cooling rate caused by water, resulting in significantly refined grain sizes. Regarding the specimen positioning, the measured grain sizes had no specific trend. The average grain sizes are depicted in Figure 4. The average minimum grain sizes for the 1-pass SFSPed joint was 3.56 µm, 3.53 µm for the 2-pass, 2.25 µm for the 3-pass and 1.49 µm for the 4-pass joint. The fourth  The microstructural grains during the SFSP process reduced the grain growth and the migration rate of grain boundaries, which then led to fine equiaxed grain structure in the stir zone of the joint [31][32][33]. Additionally, Lou et al. [34] proved that during the multi-pass SFSP, the processed surface undergoes an enhanced cooling rate caused by water, resulting in significantly refined grain sizes. Regarding the specimen positioning, the measured grain sizes had no specific trend. The average grain sizes are depicted in Figure 4. The average minimum grain sizes for the 1-pass SFSPed joint was 3.56 µm, 3.53 µm for the 2-pass, 2.25 µm for the 3-pass and 1.49 µm for the 4-pass joint. The fourth SFSPed Metals 2020, 10, 1429 7 of 14 pass resulted in a very fine homogeneous grain structure. Similar work including the application of multiple pass SFSP was reported in the literature [34][35][36]. SFSPed pass resulted in a very fine homogeneous grain structure. Similar work including the application of multiple pass SFSP was reported in the literature [34][35][36].  Figure 5 depicts the tensile stress and strain curves for the multiple submerged friction stirprocessed FSWed joints. It is evident that the UTS increased with an increase in the number of SFSP passes. The tensile strain also followed the same trend except for the 2-pass SFSP, which had lower strain compared to the rest of the passes. This was due to the joint having a tunnel defect. Looking at specimen positioning as depicted in Figure 6a, there was no particular trend noticed for the UTS in all the passes, but for the percentage elongation, the 2-pass SFSP one increased towards the specimen extracted toward the end of the joint. Figure 6b shows the average UTS and percentage elongation. Both the average UTS and the percentage elongation increased as the number of passes increased. However, the percentage elongation of the 2-pass strain and percentage elongation was found to be lower compared to the rest of the passes due to the tunnel defect that was observed. The increase in the tensile strength and percentage elongation correlates with the microstructural grain sizes. Similar studies where the application of submerged multiple pass SFSP increased UTS were reported in the literature [23,[35][36][37].   Figure 5 depicts the tensile stress and strain curves for the multiple submerged friction stir-processed FSWed joints. It is evident that the UTS increased with an increase in the number of SFSP passes. The tensile strain also followed the same trend except for the 2-pass SFSP, which had lower strain compared to the rest of the passes. This was due to the joint having a tunnel defect. Looking at specimen positioning as depicted in Figure 6a, there was no particular trend noticed for the UTS in all the passes, but for the percentage elongation, the 2-pass SFSP one increased towards the specimen extracted toward the end of the joint. Figure 6b shows the average UTS and percentage elongation. Both the average UTS and the percentage elongation increased as the number of passes increased. However, the percentage elongation of the 2-pass strain and percentage elongation was found to be lower compared to the rest of the passes due to the tunnel defect that was observed. The increase in the tensile strength and percentage elongation correlates with the microstructural grain sizes. Similar studies where the application of submerged multiple pass SFSP increased UTS were reported in the literature [23,[35][36][37]. SFSPed pass resulted in a very fine homogeneous grain structure. Similar work including the application of multiple pass SFSP was reported in the literature [34][35][36].  Figure 5 depicts the tensile stress and strain curves for the multiple submerged friction stirprocessed FSWed joints. It is evident that the UTS increased with an increase in the number of SFSP passes. The tensile strain also followed the same trend except for the 2-pass SFSP, which had lower strain compared to the rest of the passes. This was due to the joint having a tunnel defect. Looking at specimen positioning as depicted in Figure 6a, there was no particular trend noticed for the UTS in all the passes, but for the percentage elongation, the 2-pass SFSP one increased towards the specimen extracted toward the end of the joint. Figure 6b shows the average UTS and percentage elongation. Both the average UTS and the percentage elongation increased as the number of passes increased. However, the percentage elongation of the 2-pass strain and percentage elongation was found to be lower compared to the rest of the passes due to the tunnel defect that was observed. The increase in the tensile strength and percentage elongation correlates with the microstructural grain sizes. Similar studies where the application of submerged multiple pass SFSP increased UTS were reported in the literature [23,[35][36][37].    Figure 7 depicts the SEM fracture surface morphologies of the multiple pass SFSPed tensile specimens. All the specimens showed ductile failure with different dimple arrangements. The four-pass showed finer equiaxed dimples compared to the other SFSP passes. The SEM morphologies correlated with the grain sizes of the respective SFSP passes. The morphologies were also in agreement with the percentage elongation results of the specimens [38][39][40].

Tensile Properties
Metals 2020, 10, x FOR PEER REVIEW 9 of 14 Figure 7 depicts the SEM fracture surface morphologies of the multiple pass SFSPed tensile specimens. All the specimens showed ductile failure with different dimple arrangements. The fourpass showed finer equiaxed dimples compared to the other SFSP passes. The SEM morphologies correlated with the grain sizes of the respective SFSP passes. The morphologies were also in agreement with the percentage elongation results of the specimens [38][39][40].   Figure 8 shows the hardness profiles for the multiple pass SFSPed joints. Studying the figure, the specimen extracted in the middle of the joint had higher hardness compared to the specimens extracted towards the start and the end of the joints. Similar observations were noted in the literature [17]. The hardness of the heat-affected zone (HAZ) on the AA6082 side in 1-pass, 3-pass and 4-pass was found to be higher compared to the thermo-mechanical affected zone (TMAZ) and stir zone, while the 2-pass TMAZ hardness was higher compared to the HAZ and stir zone. However, TMAZ, HAZ and stir zone harnesses were lower in all the passes compared to the base material of AA6082. This is due to the AA6082 being precipitate hardened alloy [17,[41][42][43]. The hardness profiles for all the passes declined towards the stir zone and further to the AA8011 side. However, the degree of declining differed according to the number of SFSP passes. Additionally, the hardness of the TMAZ and HAZ of the AA8011 side in all the passes was found to be lower compared to the stir zone. The stir zone hardness was found to be higher compared to the base material AA8011 hardness. Additionally, looking at the hardness of the stir zone to the AA8011 base material of the 3-pass SFSP for the specimen extracted towards the end of the joint, it can be seen that it was notably lower due to the defects on the position as a result of being cut too close to the SFSP tool exit hole. This is due to the submerged conditions preventing the material softening and coarsening of the grains [17,[23][24][25]. The 1-pass hardness of the stir zone maximum hardness was found to be 75 HV, 65 HV for the 2-pass, 65 HV for the 3-pass and 67 HV for the 4-pass specimen.

Hardness
Metals 2020, 10, x FOR PEER REVIEW 10 of 14 Figure 8 shows the hardness profiles for the multiple pass SFSPed joints. Studying the figure, the specimen extracted in the middle of the joint had higher hardness compared to the specimens extracted towards the start and the end of the joints. Similar observations were noted in the literature [17]. The hardness of the heat-affected zone (HAZ) on the AA6082 side in 1-pass, 3-pass and 4-pass was found to be higher compared to the thermo-mechanical affected zone (TMAZ) and stir zone, while the 2-pass TMAZ hardness was higher compared to the HAZ and stir zone. However, TMAZ, HAZ and stir zone harnesses were lower in all the passes compared to the base material of AA6082. This is due to the AA6082 being precipitate hardened alloy [17,[41][42][43]. The hardness profiles for all the passes declined towards the stir zone and further to the AA8011 side. However, the degree of declining differed according to the number of SFSP passes. Additionally, the hardness of the TMAZ and HAZ of the AA8011 side in all the passes was found to be lower compared to the stir zone. The stir zone hardness was found to be higher compared to the base material AA8011 hardness. Additionally, looking at the hardness of the stir zone to the AA8011 base material of the 3-pass SFSP for the specimen extracted towards the end of the joint, it can be seen that it was notably lower due to the defects on the position as a result of being cut too close to the SFSP tool exit hole. This is due to the submerged conditions preventing the material softening and coarsening of the grains [17,[23][24][25]. The 1-pass hardness of the stir zone maximum hardness was found to be 75 HV, 65 HV for the 2pass, 65 HV for the 3-pass and 67 HV for the 4-pass specimen.

Conclusions
The effects of multiple passes of the submerged friction stir-processed friction stir-welded AA6082-AA8011 dissimilar joints on the microstructure and mechanical properties were successfully investigated. Based on the results obtained, the following conclusions can be drawn: 1. There was no particular trend in the microstructure and mechanical positioning when looking at the specimen positioning in all the passes. The mean grain sizes were refined from 15.12 to 5.42 µm. The minimum mean grain sizes were refined from 3.54 to 1.49 µm and the standard deviation from 5.43 to 1.87 µm.
2. The ultimate tensile strength was improved from 84.96 to 94.77 MPa. The four-pass SFSPed specimens were found to have more ductile properties compared to the one-pass SFSPed specimen. The percentage elongation of the joints was improved from 18.13% to 24.13%. The fracture surface morphologies revealed a ductile failure mode.

Conclusions
The effects of multiple passes of the submerged friction stir-processed friction stir-welded AA6082-AA8011 dissimilar joints on the microstructure and mechanical properties were successfully investigated. Based on the results obtained, the following conclusions can be drawn:

1.
There was no particular trend in the microstructure and mechanical positioning when looking at the specimen positioning in all the passes. The mean grain sizes were refined from 15.12 to 5.42 µm. The minimum mean grain sizes were refined from 3.54 to 1.49 µm and the standard deviation from 5.43 to 1.87 µm.

2.
The ultimate tensile strength was improved from 84.96 to 94.77 MPa. The four-pass SFSPed specimens were found to have more ductile properties compared to the one-pass SFSPed specimen. The percentage elongation of the joints was improved from 18.13% to 24.13%. The fracture surface morphologies revealed a ductile failure mode.

3.
The hardness of the stir zones in all the passes was found to be higher compared to the AA8011 base material but lower than the AA6082 one. The maximum stir zone hardness of 75 HV was observed on the one-pass SFSP joints.