Introduction

Aluminum alloy AA2024 has a high strength-to-weight ratio, and is mainly used for manufacturing aircraft skeleton parts, skins, and other high-load parts (Ref 1). AA7075 has a high specific strength and good weldability; moreover, it is widely used in the upper skin of aircraft wings, in fuselage frames and ribs, and in ship structures, and high-speed trains (Ref 2). In practical applications and production, different materials are selected for various parts according to the performance requirements of the actual working conditions, manufacturing process, and cost, and the welding of 2024 and 7075 aluminum alloys is inevitable. Conventional fusion welding of aluminum and its alloys results in defects (porosity and hot cracking) in the fusion zone, which significantly affect the mechanical properties of the joints (Ref 3, 4). Therefore, it is necessary to adopt novel methods to reduce or eliminate these problems.

Friction stir welding (FSW) is a solid-state welding technique invented in 1991 at The Welding Institute (TWI), UK (Ref 5). Owing to the low heat input during welding, FSW can effectively avoid hot cracking, porosity, and other defects caused by fusion welding (Ref 6, 7). Additionally, the tensile strength of FSW-welded specimens is greater than that of fusion-welded joints (Ref 8). FSW is a promising welding technique for joining lightweight alloys, and is widely used in the welding of aluminum alloys (Ref 9). Although FSW of metals and alloys is relatively easy to implement, using different welding parameters to weld dissimilar materials changes the microstructure, affecting the mechanical properties of the joint. Therefore, it is crucial to understand the microstructure development in FSW and how it affects the mechanical properties of the joint.

Owing to the large temperature gradients and plastic flow of FSW, the microstructure of the welded joint changes. Generally, the microstructure of a joint can be divided into four areas: stir zone (SZ), thermo-mechanical affected zone (TMAZ), heat-affected zone (HAZ), and base metal (BM) (Ref 10). Various studies have focused on FSW joints. Radisavljevic et al. (Ref 11) studied FSW of AA2024, and established that the particles in the SZ were the finest. Studies have shown that the weakest regions of FSW joints of 7075 alloy are generally in the TMAZ and HAZ, where the hardness values are low (Ref 12, 13). This is mainly because of the high weld heat input in these areas, leading to the coarsening of the precipitates and grains. Some precipitates can be dissolved in the SZ of 7075-T6 after FSW. This indicates that the welding temperature is higher than the solution temperature of the hardening precipitate, but lower than the melting temperature of the alloy, which is between 450 and 480 °C (Ref 14). Zhang et al. (Ref 15) established that, owing to the thermal cycle generated in AA2024 / AA7075 FSW, the precipitates in the TMAZ were coarsened, and the grains in the TMAZ were more twisted and bent than those in the BM. Research on the FSW of 2024 and 6061 aluminum alloys established that the materials in the dynamic recrystallization (DRX) region were exposed to a lower temperature than materials in the SZ far away from the TMAZ. The grain size in the SZ adjacent to the TMAZ (DRX region) is finer than that in areas far away from the TMAZ (DRX and continuous grain growth region). In contrast, the lower temperature of the DRX region results in smaller CuAl2 precipitates (Ref 16). The high temperature and plastic deformation in the SZ will lead to the refinement of crystal grains and dissolution or re-precipitation of the precipitated phase (Ref 17). Extensive precipitates are re-precipitated in the SZ at a high heat input (Ref 18).

In summary, the microstructure of the FSW affects the mechanical properties of the joint, and the heat input is crucial to the microstructural changes of the joint. Although the different microstructures of FSW joints have been studied to a certain extent, the effect of heat input on the microstructures of FSW joints of AA2024 and AA7075 dissimilar alloys is unclear. Therefore, this study evaluates the microstructural characteristics of AA2024/AA7075 dissimilar aluminum alloy FSW. Additionally, the precipitate distributions of the welded joints were analyzed through scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), and x-ray diffraction (XRD), and these observations were then related to the mechanical properties of the welds.

Experimental Procedure

In this study, the base materials were AA2024-T3 and AA7075-T6 aluminum alloys. The dimensions of the welded plates were 200 × 100 × 3.2 mm. The chemical compositions of the alloys are listed in Table 1. The mechanical properties are listed in Table 2.

Table 1 Chemical composition (Wt%)
Full size table
Table 2 Mechanical properties
Full size table

A rotary tool, with a 14 mm diameter concave shoulder and a 2.7 mm long tapered thread pin, was used. The top diameter of the pin was 3 mm, and the inclination angle was 14° (see Fig. 1a). A gantry numerical control FSW machine was used for welding, as shown in Fig. 1(b). Prior to welding, dirt and oxide layers were removed from the plates. The plates were rigidly clamped. Welding was performed at various rotational speeds of 600–1000 rpm and welding speeds of 300 and 400 mm/min. Assuming an ideal condition where the total work is converted into heat, the average heat input per unit time is obtained using an empirical equation (Ref 19).

$$ Q = \frac{{4\pi ^{2} \cdot \mu \cdot P \cdot \omega \cdot R^{3} }}{{3V}} $$
(1)

where Q is the amount of heat generated, \(\mu\) is the friction coefficient, P is the pressure, R is the shoulder radius, and ω and V are the tool rotation rate and welding speed, respectively. According to Eq. 1, the order of amount of heat input during welding is: Q600-400 <Q600-300 <Q800-400 <Q800-300 <Q1000-400 <Q1000-300. The welding parameters are listed in Table 3. Six experiments were conducted to analyze the microstructure and mechanical properties.

Fig. 1
figure 1

(a) Tool (unit: mm), (b) FSW machine, (c) Dimension of the specimen for tension test (unit: mm)

Full size image
Table 3 Welding parameters
Full size table

After welding, microstructural studies were performed using a metallurgical microscope. The specimens were cut along the weld cross section, and then prepared by wet-grinding with water using different grades of emery paper, followed by polishing with diamond paste. The surfaces of the specimens were etched using Keller’s solution (HNO3: 5 mL, HCl: 3 mL, HF: 2 mL, H2O: 190 mL). Microstructural examination of butt-welded joints was carried out via SEM and EDS.

The tensile tests were performed using a universal testing machine. The testing specimens were cut to the shape shown in Fig. 1(c) using wire cut electrical discharge machining (WEDM). Three specimens were obtained from each welded joint, and the average of the three specimens was taken as the test result. The fracture morphology was observed via SEM. Microhardness measurements were taken from the polished cross sections of the welds using a Vickers indenter with a load of 200 gf for 15 s.

Results and Discussions

Surface Morphology

The surface morphologies of the AA2024 and AA7075 FSW joints are shown in Fig. 2. The advancing side (AS) and retreating side (RS) are on the right and left of the photographs, respectively. There are no defects or cracks evident on the surfaces of any of the specimens. This indicates that FSW successfully connected the two different BMs when a suitable speed and feed rate were used. The interface features depend on the total heat input and temperature during welding (Ref 19). The welds in Fig. 2(a), (d), and (e) are smooth, whereas those in Fig. 2(b), (c), and (f) are rough. Arbegast (Ref 20) used the equation of maximum temperature of FSW (Eq. 2) to demonstrate the effects of the rotational and welding speeds.

$$ \frac{T}{{T_{m} }} = K\left( {\frac{{\omega ^{2} }}{{V \times 10^{4} }}} \right)^{\alpha } $$
(2)

where T is the FSW temperature (°C); Tm is the melting temperature of the sheet material (°C); 0.04 <\(~\alpha\) < 0.06 and 0.65 <K <0.75 are two defined constants; and \(\omega\) and V are the tool rotational and welding speeds, respectively. Based on Eq. 2, the FSW temperature was related to the \(\omega ^{2} /V\) ratio. According to Eq. 2, the temperatures corresponding to Fig. 2(b), (c), and (f) are higher than those for Fig. 2(a), (d), and (e). Because of the increase in heat input, abnormal sticking of the pasty aluminum, and shearing under the shoulder can occur, making the surface of the weld less bright, and resulting in burrs and peeling (Ref 21).

Fig. 2
figure 2

Surface appearances of joints (a) 600-300, (b) 800-300, (c) 1000-300, (d) 600-400, (e) 800-400, (f) 1000-400, (unit: rpm-mm/min)

Full size image

Microstructural Observations

The stirring of the FSW tool causes a complex material to flow in the SZ. Figure 3(a) shows that AS (AA2024 BM) flows to RS (AA7075 BM). AA2024 accounts for most of the material in the SZ. The microstructure of the mixed interface of the two materials is relatively complex; the cross section of the joint is divided into four areas: SZ, TMAZ, HAZ, and BM. According to the results presented in Fig. 3(e) and (i), the SZ experienced high temperature and plastic deformation, so dynamic recrystallization occurred, and equiaxed and very fine grain structures could be observed. The material in the TMAZ was severely deformed, and had thin, twisted grains owing to deformation and high extrusion. Moreover, as shown in Fig. 3(d) and (h), sharper demarcation between the weld nugget and TMAZ was observed on the AS, and a more gradual transition from the weld nugget to the TMAZ was observed on the RS. Grains were observed on the AS oblique upward at 45°, but not on the RS. This is because of the plastic deformation in FSW. The linear velocity of the material at the RS is equal to the difference between the rotational speed and feed rate, and the linear velocity of the material at the RS is significantly lower than that at the AS. Therefore, the boundary on the RS is not as clear as that on the AS. However, the HAZ is only affected by the heat input; the grains are similar to those of the BM, but larger. Figure 3 shows that the grains of the BM AA2024 and AA7075 are oblong shaped. The quantities of second-phase particles in the BM and HAZ are significantly more than in the SZ and TMAZ, because the SZ and TMAZ have experienced the influence of high thermal input during the welding process, which can dissolve the precipitates.

Fig. 3
figure 3

Macro-image and microstructures of specimen 5: (a) macrostructure, (b–e) 7075 on RS, (f–i) 2024 on AS, (b, f) BM, (c, g) HAZ, (d, h) TMAZ and (e, i) SZ

Full size image

Figure 4 shows the SEM images of the SZ of the various welded joints. Figure 4(a), (c), and (e) shows that precipitates are distributed within the joint. Moreover, the precipitate distribution on the 2024 side is significantly denser than that on the 7075 side. Figure 4(b), (d), and (f) shows higher-magnification SEM images. Clear pores can be observed along the joint surfaces in Fig. 4(b) and (f), and the pores in Fig. 4 (b) are more significant than those in Fig. 4(f). These pores are a result of the high heat input and abnormal agitation of the material at high rotational speeds. Furthermore, owing to the high heat input, the cooling rate is faster, which can cause hole defects.

Fig. 4
figure 4

SEM images of the SZ in the as-welded joints, (a, b) 1000 rpm-300 mm/min, (c, d) 800 rpm-400 mm/min, (e, f) 1000rpm-400 mm/min

Full size image

The SEM, EDS, and XRD analyses of the joints are shown in Fig. 5, Table 4, and Fig. 6, respectively. More Al2Cu, AlCuMg, Al2CuMg, and Cu2FeAl7 precipitates could be observed at the 2024 aluminum alloy side, whereas Cu2FeAl7 and MgZn2 precipitates could be observed at the 7075 aluminum alloy side, which is consistent with the results of (Ref 22). Therefore, more precipitates form on the AA2024 side than on the AA7075 side, as shown in Fig. 4. There were AA2024 is richer in copper, while AA7075 contains more zinc. Evidently, more precipitate phases containing copper were produced in the AS, whereas those containing zinc were produced in the RS. This is mainly determined by the corresponding BMs in the AS and RS, because of the high temperature and small plastic deformation during FSW. When there is less Fe, it could dissolve in the aluminum matrix; when there is more Fe, Cu2FeAl7 could be formed (Ref 23). Cu2FeAl7 is formed at high temperatures (and is also known as the β (Cu-Fe) or N phase), and more Cu2FeAl7 is formed on the AA7075 side. Particle analyses at point 4 show that these particles can be determined to be S(Al2CuMg). In addition, the particles in the SZ have a composition close to that of MgZn2. As previously reported (Ref 24), there are two different types of precipitates in welded joints: S(Al2CuMg) and η (MgZn2). Because of the high heat input, most of the precipitates (Cu2FeAl7 and MgZn2) are dissolved in the SZ. Moreover, this is why there is more precipitation on the 2024 side than on the 7075 side.

Fig. 5
figure 5

SEM enlargement of (a) 2024 aluminum alloy BM, (b) SZ at 800 rpm-400 mm/min, (c) 7075 aluminum alloy BM

Full size image
Table 4 EDS scan element distributions at the points shown in Fig. 5 Atomic percent (%)
Full size table
Fig. 6
figure 6

XRD patterns of the joints

Full size image

Mechanical Characterization

To analyze the effect of the heat input on the tensile strength, the welded joints were subjected to tensile tests, and the fracture parameters—ultimate tensile strength (UTS) and elongation to break—were compared. The UTS values depend on the welding conditions, as shown in Fig. 7. The temperatures obtained under different welding parameters were calculated using Eq. 2, and were in the following order: T600-400 <T600-300 <T800-400 <T800-300 <T1000-400 <T1000-300. When the welding parameter was 800 rpm–400 mm/min, the UTS of the joint was the highest (346 MPa), and the elongation was 3%. The minimum UTS of value obtained was 257 MPa, when the welding parameters were 600 rpm-400 mm/min and 1000 rpm–300 mm/min. This is because the 800 rpm-400 mm/min combination has a suitable heat input, which leads to enhanced precipitation in the SZ, resulting in a considerable amount of Al2CuMg, MgZn2, and other precipitates which improve tensile strength. A significantly higher or lower temperature results in lower tensile strength and elongation. This is because the heat input is too low to deform the material completely, and the fluidity of the material from the two BMs is poor during welding. In addition, increasing the temperature excessively causes a higher heat input. This could cause microstructural changes, such as precipitate dissolution, the grains becoming coarsened or thicker, and fewer grain boundaries (Ref 25). In addition, when the heat input is higher during welding, the cooling rate is faster, which could form hole defects on the welding interface and significantly affect the quality of the joint.

Fig. 7
figure 7

The tensile properties of welded specimens

Full size image

Figure 8(a) shows the fracture position of the tensile test specimen that was welded at a rotational speed of 800 rpm and a feed rate of 400 mm/min. The fracture locations of all samples were in the SZ, and no significant necking was observed. From the cross section of the fracture, it can be observed that fracture occurred at the bonding surface at the bottom of the weld, whereas the upper part of the weld was not on the bonding surface. This was also confirmed by the fracture scan shown in Fig. 8(b). The fracture location was similar to that reported in (Ref 26). During welding, owing to the difference in the heat gradient between the upper and lower parts of the weld, and the lower part being close to the bottom plate, there is insufficient heat input to the lower part of the weld. Stress concentration occurs at the bottom of the weld, and this could be the reason for the fracture being located in the SZ.

Fig. 8
figure 8

(a) Tensile test fracture position of the sample at 800 rpm and 400 mm/min feed speed. SEM images at different magnifications of the fracture surfaces of tensile tests at (b, c) 1000 rpm and 300 mm/min; (d, e) 800 rpm and 400 mm/min

Full size image

Fracture surface images of selected tensile specimens are shown in Fig. 8(b)–(e). All the fracture patterns are similar, as shown in the enlarged view. All fracture locations have dimples, which are associated with ductile fracture. The dimples on the sample welded with parameters of 800 rpm-400 mm/min are smaller and denser than on the sample welded with parameters of 1000 rpm-300 mm/min. The dimples are caused by the nucleation, growth, and coalescence of voids. These voids are formed because of the inhomogeneous rates of deformation in the material under the applied load (Ref 27).

The hardness was measured at 0.5 mm spacing across the welds, and two representative examples are shown in Fig. 9. According to Fig. 9, the hardness of AA7075 is higher than that of AA2024. Their values were approximately 161 and 143 HV, respectively. The microhardness chart of the joint shows that the low-hardness areas of the joint are in the HAZ on the AA7075 side and at the interface between the SZ and TMAZ. Minimum hardness occurred at the junction of the SZ and TMAZ. This could be because the heat input is significantly higher, such that the hardening precipitates dissolve. The low hardness of the HAZ on the AA7075 side is attributed to the coarsening of the crystal grains and precipitates. This is consistent with the results of previous studies (Ref 28, 29). However, the hardness of SZ is slightly higher because of the fine equiaxed recrystallized grains in the SZ. As shown in the figure, the SZ becomes wider as the rotation speed increases.

Fig. 9
figure 9

Microhardness distribution of dissimilar FSW cross sections

Full size image

Conclusions

  1. 1.

    Al2Cu, Al2CuMg, AlCuMg, and Cu2FeAl7 particles were detected on the AA2024 side in the SZ, whereas Cu2FeAl7 and MgZn2 particles were present on the AA7075 side.

  2. 2.

    The 800 rpm-400 mm/min sample experienced the optimum heat input, and had the best tensile performance; its UTS reached 346 MPa, and elongation was 3%. Reaching a significantly high or low temperature during welding results decreases tensile strength and elongation.

  3. 3.

    The TMAZ exhibited the lowest hardness values in the joints. The thermal cycle generated via FSW leads to the coarsening of crystal grains and precipitated phases.

  4. 4.

    Because the thermal gradient is extremely large, the heat input at the bottom of the weld is insufficient; thus, the fracture position is in the SZ. Dimples were observed on the surface of the fracture, indicating that the fracture mode was ductile fracture.