Study on mechanical properties of 2219 aluminum alloy bobbin tool friction stir welding

The Bobbin tool friction stir welding (BT-FSW) process experiment was performed on 2219 aluminum alloy with different butt clearances and misalignments. The maximum allowable clearance and misalignment amount for BT-FSW were obtained, and the weld formation, mechanical properties and microstructure under different parameters were compared and analysed. The results show that the face and back sides of the weld joint surface are perfect and there are no defect inside with the butt clearance under 2 mm and the misalignment under 1.5 mm. During the increase of butt clearance and misalignment, the tensile strength of the weld will be decreased gradually. In comparison, the impact of misalignment on weld strength is lower than that of the butt clearance. The tensile strength from both of them will reach more than 60% of the base metal strength. The lowest hardness value occurs in the thermal-mechanical influence zone, which is 79.1HV, about 70% of the base metal hardness value. The microstructure characteristics are similar to those of conventional friction stir welding. The macroscopic morphology of the welded joint is typical ‘dumbbell type’, and there are many ‘dimples’ in the tensile fracture, which is a typical ductile fracture. A large number of dispersed phase particles (Al2Cu) were found in the dimples, which is the main reason for the decrease in mechanical strength and hardness compared with the substrate.


Introduction
Friction stir welding technology has been widely used in the connection and manufacturing of aluminum alloy materials since it was invented by the British Welding Institute in 1991 [1]. It has a wide range of applications in aerospace, vessels, automobiles, and rail trains [2]. 2219 aluminum alloy is an Al-Cu-Mn ternary alloy with copper as the main alloying element. The precipitation sequence is as follows: α→α+GP zone→α+θ″ →α+θ′→α+θ [3,4]. The aluminum solid solution is represented by α, metastable phases are represented by θ′ and θ″, and the stable phase is represented by θ phase (Al 2 Cu) [5]. As a typical lightweight high-strength aluminum alloy, it has good mechanical properties and stress corrosion resistance at high and low temperatures, so it is widely used in the aerospace field. The welding characteristics of 2219 aluminum alloy friction stir welding have also been widely concerned by scholars from countries all over the world: Weifeng Xu et al [6,7]. studied the microstructure, second-phase particles and mechanical properties of the joint during friction stir welding of 2219-T6, and they studied on the effect of cooling conditions with thick plate 2219-T6 FSW. Kanwer S Arora et al [8] studied the influence of welding parameters and welding force on the weld microstructure; Gabriel et al [9] studied the failure of friction stir welded aluminum 2219.
As the first applied friction stir welding technology -conventional single shoulder friction stir welding (C-FSW) has been widely used, its inherent disadvantage during welding have also gradually exposed. These disadvantages have not been well improved or solved for a long time. For example, the root weak bonding problem cannot be fundamentally solved, and complex tooling is required for welding large structural parts. Therefore, in recent years, with the further development of friction stir welding technology, such as Bobbin tool friction stir welding [10,11], Static shoulder friction stir welding [12,13], the problems in C-FSW have been solved to a certain extent, and gradually researched by experts and scholars in the field.
Bobbin tool friction stir welding, as a new type of solid phase connection technology, has the advantage that, it does not need the internal rigid support of conventional friction stir welding, which not only simplifies the tooling structure, but also reduces the manufacturing cost for its internal rigid support. At the same time, it can obtain the welding joint with better performance [14], which has received more and more attention in recent years.
NASA Aerospace Agency of the United States took the lead in the research and application of this aspect, confirmed the unique advantages of BT-FSW in controlling welding deformation, circumferential weld structure formation, and workpiece thinning. Besides they applied this technology to the circumferential seam welding for the rocket tank, which belongs to the new generation of 'Orion' manned spacecraft [15]. Zhao Yanhua et al [16] tested 2219-T87 aluminum alloy with a thickness of 6 mm using Archimedean spiral shoulder and bobbin tool shoulder stirring tool with a two-way reverse thread structure and finally obtained a wellformed and defect-free welded joint; Q Chu et al [17] studied the defect formation and microstructure evolution of BT-FSW in the welding process; Z Liu et al [18] studied the relationship between welding force and welding defect formation during BT-FSW welding; Zhao Yunqiang et al [19] carried out BT-FSW test on 6063-T4 aluminum alloy with a thickness of 3 mm and found that the weld surface was well formed, and there was no obvious delamination of the microhardness value of the weld section. With the continuous exploration and optimization of BT-FSW, great breakthroughs have been made in BT-FSW in recent years with the unremitting efforts of researchers from universities and enterprises around the world [20,21]. Figure 1 is the working schematic diagram of BT-FSW. According to this figure we can see when BT-FSW works, the back side of the workpiece does not need rigid support, and it is better to manufacture some structures without rigid support on the back, such as cylinders. The upper and lower shoulders friction against the workpieces at the same time, and the indentation and weld bead lines are produced on the upper and back surfaces, which provide symmetrical welding heat input, making it more uniform in thickness direction than C-FSW, the performance difference in the thickness direction of the plate and the weakening of the joint root in the case of thick plate FSW are avoided [22].
Because of the excellent advantages of BT-FSW in the manufacture of cylindrical workpieces, which cannot be effectively rigidly supported on the back due to structural (diameter, length) constraints, it is widely used in launch vehicle tanks and missile shells. However, for the manufacture of high-strength materials, affected by the strength of the material itself, it is difficult to form the cylinder during the rolling process, and the size is difficult to guarantee. When BT-FSW welding is performed on it in engineering applications, different degrees of distortion (butt clearances and misalignments) would be occurred from time to time. Therefore, it has important engineering application significance to study the influence of different butt clearance and misalignment on the performance of BT-FSW welded joints. At present, there are relatively few research reports on this aspect. In this paper, 2219 aluminum alloy is taken as the research object, and two parameters of butt clearance and misalignment are taken as the research object. The weld formation and mechanical properties are compared and analyzed, and their changed rules are obtained, thus certain reference and data can be provided in the future research for supporting BT-FSW research.

Test materials
The material used in the test is 2219 aluminum alloy, and the material state is T651(Cooling after solution heat treatment to achieve higher strength, then cold stretching by stretching machine to eliminate residual internal stress after heat treatment), the aluminum alloy is obtained after solution treatment and assisted by cold processing process. The size of the test plate is 400 × 150 × 8 mm. The main chemical composition and mechanical property parameters are shown in tables 1 and 2. Figure 13(a) shows the metallographic structure of base metal, figure 8(BM1) and figure 9(BM2) show the Stress-strain curve of base metal.

Welding method
The stirring tool used in the test is a self-designed 7.5 mm bobbin tool. The diameter of the pin is 10 mm, and the upper and lower shoulders are 22 mm. Tool material is H13 hot work tool steel under ASTM A681 standard, which chemical composition and properties are listed in table 3. The upper and lower shoulders have spiral shape structure, and the pin has special thread structure [1,23]. During welding, the inclination angle is 0°, and the dimensions of the upper and lower shoulders pressed into the 0.2 mm test plates. Figure 2 shows the working schematic diagram of BT-FSW tool [24,25].
Pretreatment procedure of the test: wipe the oil and impurities on the surface of the test plate with alcohol and remove the oxide film on its surface with a wire brush. Then fix the test plate with tooling. By taking the butt clearance and misalignment variable as the control variable, this experiment explores the reasonable range of the two parameters of the allowable butt clearance and misalignment amount of the BT-FSW. Besides it studies the effect of different welding process parameters on the mechanical properties of the welded joint. The specific test parameters are shown in table 4. Figure 3 shows the experimental process with butt clearance, and figure 4 shows the experimental process with misalignment.

Metallographic and mechanical property test
When the welding finished, carry out the tensile test on a universal electronic tensile testing machine. Three tensile samples were taken from the welded test plate of each parameter, then use the average tensile strength values as the judging basis. Then observe the welded joints by using OLYMPUS-BX 53 and MH-500 micro-Vickers hardness testers. For the metallographic structure and hardness distribution of the joint, the scanning electron microscope ZEISS GeminiSEM 450 was used to analyze the tensile fracture morphology.    figure 6), the front and back sides of the welded area are basically like 'fish scale' with relatively uniform small weld deformation. the surface forming of 1# joint of front and back sides (No clearance, No misalignment) with BT-FSW are shown in figure 5.
The macroscopic metallographic (figure 7) observation of the above 8 groups of BT-FSW welded joints showed that the welded joints were all dumbbell-shaped, and no welding defects were found in the macroscopic metallographic examination of the 8 groups of test participants.

Tensile property analysis of the weld joints
The tensile properties of 2219 aluminum alloy welded joints with different butt clearances and misalignments are shown in table 5. The tensile strength under both parameters decrease with the increase of butt clearances   and misalignments, but the degree of decline is different. When the butt clearance is 0-0.5 mm, its tensile strength changes a little, however when the butt clearance reaches to 0.5-2 mm, its tensile strength decreases significantly. If the butt clearance is 2 mm, its tensile strength can still reach more than 60% of the base metal strength. During the change of misalignment between 0 to1.5 mm, the tensile strength also showed a trend of gradual reduction, but the degree of strength reduction was smaller than the butt clearance. Figure 8 is the stressstrain curves of tensile test in the contrast of different butt clearances, and figure 9 is the stress-strain curves of tensile test in the contrast of different misalignments.
By observing table 5 and figure 8, we can conclude that the tensile strength of the BT-FSW welded joint decreases with the increase of the butt clearances of the welded workpieces. A large butt clearance will result in   incompletely filled groove, poor connectivity in the welded area, and the butt clearance will cause restricted material flow which will affect the tensile strength of the welded joint [26,27]. The heat and mechanical force from FSW are generated by friction stir in order to soften and mix the metal. With the butt clearance increase, the friction stir effect cannot be fully exerted, which will affect the structure and strength of the weld. By the increase of the butt clearance, it will lead to insufficient heat input. The temperature of the welding area cannot reach the ideal welding temperature, which affects the softening ,fluidity of the metal and the reduction of the bond welding strength [28,29].
By observing table 5 and figure 9, we can conclude that the tensile strength of the BT-FSW welded joint decreases with the increase of the misalignments of the welded workpieces. When the amount of misalignment gradually increases, it will affect the frictional heat generation between the shoulder and the workpiece. At the same time, the heat input of the materials on both sides and the temperature in the welding area of the materials will be changed on both sides. Because of that, it may lead to the non-uniformity and Phase change, especially the  mechanical properties of the weld [30,31]. The difference between BT-FSW and C-FSW is that there are shoulders on the upper and lower sides. With their frictions against the workpiece, the heat will be generated, and the gradually increasing misalignment will lead to different contact depths between shoulder and both surfaces of the workpieces. It will affect the heat input during the welding [13,32]. The Frictional heat production is different, the upper surface of the material has high heat input and lower surface has low heat input. The opposite side is symmetrical to it. This will not only cause a complex stress field due to the temperature difference in the thickness of the materials on both sides, but also cause an uneven distribution of shear force during the FSW process. Uneven shear force will lead to uneven strain distribution in the welding area, which will affect the structure and strength of the weld. At the same time, large assembly misalignment will affect the metal structure of the welding area. Besides the deformation and dislocations may cause distortion of the grains and abnormal grain boundary morphology, which will reduce the strength and ductility of the weld [18,33].
By observing the surface of the tensile fracture sample in figure 10, it can be found that there is nearly no necking phenomenon when the welded joint of BT-FSW breaks. The reason for this phenomenon is, because of the high temperature and mechanical stirring during the welding process, the grain structure in the weld area is refined and homogenized, which improves the strength and plasticity of the weld area. The refined and homogenized grain structure makes the welded area of aluminum alloy get higher toughness and tensile strength, so there is no obvious necking in the welded seam area [34]. The high plasticity and good ductility of aluminum alloys are also one of the reasons why there is no obvious necking after BT-FSW [35]. The aluminum alloy can undergo large plastic deformation when it is stressed, which makes the weld area have good ductility, which can absorb stress and extend during the tensile test, and is not prone to necking [36].
4. Microstructure analysis of the weld joints 4.1. Macroscopic morphology of the weld joint Figure 11 shows the macroscopic joint morphology of BT-FSW, which is mainly composed of the base metal zone (BM), heat affected zone (HAZ), thermo-mechanical affected zone (TMAZ), and nugget zone (NZ). The weld seam is a 'dumbbell' shape with a wide top and bottom and narrow middle. It can be seen that the welded  joint is well-formed and has no internal defects, indicating that the process parameters in this study can achieve relatively high-quality welding.
Affected by the upper and lower shoulders and the reverse double helix structure stirring tool, the weld nugget area has obvious 'onion ring' characteristics. According to the understanding of the plastic metal flow theory (shown on figure 12), the formation of this feature is closely related to the plastic metal flow behaviour of the material. During welding, the metal on both sides of the base metal moves axially to the centre of the weld section under the action of the stirring tool [37]. Driven by the plastic material, the front metal moves towards the weld edge and finally moves towards the front and back of the weld respectively, thus forming a ring.

Microstructure and morphology of the weld joint
The microstructure of the welded joint is analyzed. After the sample is grounded and polished, the surface of the sample s corroded by Keller reagent (1 ml HF +1.5 ml HCl + 2.5 ml HNO 3 + 95 ml H 2 O), and the surface of the sample is observed by the OLYMPUS-BX 53 metallographic microscope. The metallographic structure is shown in figure 13. Figure 13(a) shows the microstructure of the base metal zone, which is mainly composed of a series of relatively regular lath-like grains, which is related to the rolling process of 2219 aluminum alloy. Figure 13(b) shows the structure of the heat-affected zone (HAZ). Its microstructure is similar to traditional fusion welding. Because it is located on the edge of the pin, it is weakly affected by mechanical stirring. Under the action of the welding heat cycle, the grains appear and grow. However, due to the low heat input of BT-FSW, the change of grain orientation is small [38,39]. Figure 13(c) shows the structure of the TMAZ. Its grain shape is significantly different from that of the nugget zone and the heat-affected zone. Its two sides are close to the periphery of the nugget zone and are squeezed and disturbed by the flowing metal in the nugget zone deformed by elongation [40]. In addition, under the action of the thermal cycle below the nugget zone, the incompletely plasticized metal in the thermal engine affected zone is affected by the flow direction of the plastic metal, resulting in a certain directionality in the grain deformation; Figure 13(d) shows the microstructure of the weld nugget Zone (NZ), which is mainly composed of fine equiaxed grains with relatively uniform size distribution. The main reason for its formation is that the base metal in the nugget area is subjected to the heat generation effect of the upper and lower shoulders and the sufficient stirring effect of the stirring tool [41]. The slab-like grains are completely crushed, and complete dynamic recrystallization occurs. At the same time, the heat input of the BT-FSW is low, which makes the cooling speed of the nugget area faster, and finally nucleation grows to form a small equiaxed crystal structure [42].

Microhardness of the weld joint
When measuring the microhardness of the joint, take points for measurement on the upper, middle and lower parts of the joint to obtain the characteristics of the hardness in the thickness direction after BT-FSW. Take 31  points for each layer for measurement, and the point spacing is 1mm, The loading pressure is 200g, and the pressure is maintained for 10s. Figures 14 and 15 are schematic diagrams of joint hardness dots and hardness distribution diagrams of 8mm thick 2219 aluminum alloy BT-FSW joints with different thicknesses: It can be seen from figure 14 that the hardness distribution law of BT-FSW is similar to that of C-FSW, and the hardness trend of different thicknesses is 'W' type. Among them, the hardness of the base metal is the highest. The hardness value gradually decreases when it reaches the HAZ. After that the hardness reaches the lowest value in the TMAZ, which hardness is 79.1 HV and it is about 70% of the hardness value of the base metal. The hardness value increases gradually from TMAZ to NZ.
The main reason for this phenomenon is, the heat input during friction stir welding can cause the precipitation of undesirable phases or changes in the microstructure, which can lead to a decrease in hardness. The 2219 Aluminum alloy contain the strengthening phase particles θ (Al 2 Cu) contribute to their mechanical properties [43,44]. The high temperatures experienced during FSW can disrupt the distribution of these precipitates, which lead to a coarser and less uniform microstructure. During friction stir welding, the metal in the weld zone undergoes deformation at high temperatures and high strain rates. This deformation leads to a reduction of the metal grain size, which is also resulting in a smaller grain size in the area of the welded joint than in the parent metal. A smaller grain size not only increases the strength of the material, but also leads to a decrease in hardness [45]. This is the main reason why the hardness value of the welded joint is only 70% of the base metal.
The difference between BT-FSW conventional friction stir welding is that the hardness variation trend of the upper, middle and lower layer of the BT-FSW joint is almost the same, and the difference in the thickness direction is small. The upper and lower shoulders of BT-FSW interact with the workpiece at the same time to generate frictional heat. Then the simultaneous heat input from the upper and lower sides will make the heating more uniform in the thickness of the material [46]. Compared with C-FSW [47], the material flow effect and frictional heat input caused by the friction between the shoulder and the upper surface of the plate are difficult to act on the lower surface of the plate away from the shoulder, which will lead to an insufficient material flow and the heat input on the back of the weld [48]. Therefore, there are obvious performance differences in the thickness direction of the welded joints of C-FSW thick plates. The microhardness test simply and clearly reflects this difference. From the results in figure 14, we can see that the microhardness values of the upper and lower regions of the welded joint of BT-FSW effectively solves the problem of the performance difference of C-FSW in the material thickness direction. Figure 16 shows the fracture surface of tensile breakage sample. From the figure, we can clearly see that there are a lot of dimples and tearing edges at the fracture. The dimples are small and deep, the tearing edge is full of micropores, cracked second-phase particles are observed at the bottom of the dimples.

Tensile fracture SEM microscopy analysis
Through EDS detection and analysis (To see the figure 17), we determined that the precipitated second phase particles are θ phase (Al 2 Cu), which also have been confirmed by some other articles [49,50]. From the SEM images, we can see that a large number of precipitates accumulate at the root of the dimples, which indicates that a large part of the crack initiation started from the precipitated phase. Because of the weak bond between the particles and the matrix, cracks can easily nucleate and coalesce. Basically, the precipitation of its strengthening phase leads to the degradation of material properties and the formation of weak points, which eventually leads to the fracture of welded joints. According to the comprehensive analysis of figure 16, 2219 aluminum allow BT-FSW to have good plasticity. Its fracture type is typical ductile fracture, and the the typical microvoid coalescence mechanism is reflected in the fracture surface. Comparative analysis found that when the butt clearance and misalignment increased, the dimple tended to decrease. At the same time the toughness of the welded joint will be decreased, which was consistent with the results of our tensile test.

Conclusions
(1) Bobbin tool friction stir welding technology was used to weld 8 mm thick 2219 aluminum alloy. The face and back sides of the weld joint surface are beautiful 'fish scale' features and there are no defects inside when the butt clearance is within 2 mm and the misalignment is within 1.5 mm; (2) The macroscopic appearance of the weld is 'dumbbell-shaped', with wide upper and lower surfaces and narrow middle. This phenomenon is mainly related to the flow behaviour of plastic metal during welding, and there is an obvious 'onion ring' feature in the weld nugget area. The spacing decreases with increasing spacing from the nugget area.
(3) With the increase of butt clearance and misalignment, the tensile properties of welded joints gradually decrease. For misalignment on joint tensile, properties are slightly lower than that of butt clearance to joint performance. The tensile strength under all parameters of the test reached more than 60% of the base metal strength.
(4) Judging from the tensile fracture morphology, there are a large number of small and deep dimples inside. The macroscopic performance shows that the plasticity of the welded joint is good, and the fracture mode is a typical ductile fracture.
(5) The precipitation and coarsening of the Al 2 Cu strengthening phase seriously weakens the strength of the welded joint and the precipitated second phase particles. Because the bonding between the particles and the matrix is weak, the cracks are easy to nucleate and coalesce. Combined with the HAZ grain coarsening caused by the welding thermal cycle, the grain coarsening and complex thermo-mechanical effects of TMAZ, fractures are mostly generated from TMAZ and HAZ.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).