Friction stir lap welding thin aluminum alloy sheets

1 Introduction

As a solid-state joining method, friction stir welding (FSW) was invented in 1991 [1,2,3]. The peak temperature during FSW is always lower than the melting point of the base metal (BM) [4,5]. Thus, FSW serves as a promising method, avoiding fusion defects [6,7,8,9]. Al alloys are widely used in modern industries because of their advantages of low densities, high strengths, good corrosion resistances and fatigue properties [10,11]. FSW has been widely used to join various kinds of Al alloys and the microstructure and mechanical properties of the welded joints have been studied [12,13,14,15,16,17].

In actual engineering applications, Al alloys of various thicknesses are required [18]. Researchers studied the joining of the alloys [19,20,21]. In case of thick sheet FSW joint, the joint consists of various microstructure and mechanical properties [19,20,21,22,23]. Martinez et al. [21] studied the microstructure and mechanical properties of thick 7449 Al alloy FSW joint and found that the joint bottom had higher hardness compared to that of the BM. The heat gradient along the thickness resulted in different microstructures in the joint. Upadhyay and Reynolds [22] investigated the effect of the backing plates on the microstructure and mechanical properties of a 25.4-mm-thick AA6061 FSW joint and reported that the back plates had significant effect on temperature at the joint root. Buchibabu et al. [23] welded a thick Al–Zn–Mg alloy and investigated the microstructure and mechanical properties of the joints. They reported that the optimum mechanical properties were achieved with a low rotating speed of 350 rpm.

For thin Al alloys sheets, low heat input is always needed during welding. However, joining thin sheets has more problems such as thickness reduction and sheet warping. Therefore, joining thin Al alloy sheets also attracted a plenty of attention [24,25,26,27]. Ahmed and Saha [24] developed a new fixture for FSW of thin Al alloy sheets. Huang et al. [25] used two tools to join 0.5-mm-thick 6061 Al alloy and reported that thickness reduction of joints was lower than 2% under rotational velocities higher than 1,500 rpm. For FSW, another joint type was lap joint. Friction stir lap welding (FSLW) is formed by two or more overlapped sheets. Similarly, lap joint is formed at solid state and therefore owns high properties.

To the author’s knowledge, no study has focused on lap joint of thin sheets. Therefore, in this work, 2024-T4 thin Al alloy sheets were lap welded. Different sleeve plunge depths were used, and the microstructure and mechanical properties of the lap joints were studied.

2 Experimental

The 2024-T4 Al alloys of 0.8 mm thick were chosen as the BM. The dimensions of the BMs were 300 × 50 mm. Before welding, the sheets were cleaned using 500 # emery papers to wipe off the oxide films. FSW-3LM-4012 machine was used. The dimensions of the tool are shown in Figure 1a. The diameter of the shoulder was 13.5 mm. The root and tip diameters of the pin were 6 and 5 mm, respectively. The length of the pin was 0.7 mm. Two sheets were lap combined at a width of 70 mm (Figure 1b). Rotating speeds of 400, 600, 800 and 1,000 rpm and welding speed of 50, 100, 150 and 200 mm/min were used. The titling angle was 2.5°.

Figure 1

Tool used in experiment (a) and schematic of the welding (b) and lap shear specimen (c).

The metallographic samples and tensile specimens were cut by an electrical discharge cutting machine. The metallographic samples underwent a standard polish procedure and were observed on an optical microscope (OM; Olympus–GX71, Olympus Corporation) and a scanning electron microscope (SEM) equipped with an energy dispersive X-ray spectrometer after etching using Keller’s reagent. The ISO 25239 was the standard used in this work for shear testing. The width of the lap shear specimen was 20 mm (Figure 1c). Vickers hardness was measured using an HVS-1000 Vickers hardness tester by a step of 0.5 mm. The testing force of 10 g was applied and the dwell time was 10 s. Two lines across the joint were tested. The first line was located at the center of the upper sheet and the second line at the lower sheet, and its distance from the bonding ligament was 0.2 mm. Lap shear tests were performed on an Instron 8801 testing machine at a speed of 3 mm/min under room temperature. After the lap shear tests, the fracture morphologies were observed using SEM.

3 Result and discussion

Figure 2 shows the joint cross section using the plunge depth of 0 mm. The shoulder surface slightly contacted the sheet during welding, resulting in weak friction between the shoulder and the sheet. Rather weak material flow behavior was induced. Only very small flash was observed. Besides, the heat input under this condition was not enough to guarantee sufficient material flow behavior. Thus, voids were formed in the stir zone (SZ; Figure 2b and c).

Figure 2

Cross section of the joint using shoulder plunge depth of 0 mm (a) and the voids (b) and (c).

Figure 3 shows the joint cross sections welded at 0.1 mm plunge depth and at different rotating speeds. The cross section presented a typical basin-like morphology similar to other typical FSW joints [28,29]. No defects were observed, illustrating that sound joints could be attained at a relatively board range. Relatively rough joint surface was obtained at 400 rpm (Figure 3a). The upper sheet bent upward at the advancing side (AS); Thus, resulting in sound surface formation (Figure 3b and c) at increased rotating speed. The alclad layers at the BM surfaces were not broken. A bonding ligament was observed at the SZ. The bonding ligament was of weak strength [30]. The sizes of the hook and cold lap were very small on the joints. The SZ widths at the bonding ligament were, respectively, 5.2, 5.4 and 5.8 mm as shown in Figure 3. The cross section was divided intoBM, heat affected zone (HAZ), thermal–mechanically affected zone (TMAZ), and SZ. Figure 4 shows the joint cross sections at different welding speeds. The cross sections presented little difference with increase in welding speed. Heat input was reduced at high welding speed, leading to narrower HAZ and TMAZ. At 150 and 200 mm/min, the upper sheets showed a little distortion (Figure 4b and c). Sheet warping is a common problem especially when welding thin sheets [31]. The warping problem was more serious at low heat input. We speculate warping is connected with the forward tool movement. When the heat input is low, the material showed bad plasticity, providing large resistance for tool movement, easily causing sheet warping.

Figure 3

Cross section of the joints using shoulder plunge depth of 0.1 mm at (a) 400 rpm, (b) 600 rpm and (c) 800 rpm.

Figure 4

Cross section of the joints using shoulder plunge depth of 0.1 mm at: (a) 100 mm/min, (b) 150 mm/min and (c) 200 mm/min.

The results of Figures 2–4 show that sound joints can be obtained when a plunge depth of 0.1 mm is used. The joint formation of FSW is closely related to the material flow behavior during welding. Thus, the schematic of the material flow behavior in this work is shown in Figure 5. Figure 5a shows the material flow behavior using the plunge depth of 0.1 mm. During welding, the plastic material of the BM flows under the stirring action of the tool. A small part of the material flows upward due to the plunge of the tool shoulder, forming flash. A large part of the material flows toward the joint center due to the collection effect of the shoulder, which is marked using red arrows in Figure 5a. At the same time, the material flows along the direction of tool rotation and also flows downward along the threads during welding [32,33]. When the material collection effect is strong enough at the shoulder, plastic material at the AS of the joint is enough not to cause void, as shown in Figure 5a. However, when the plunge depth of 0 mm is used, the material collection effect of the shoulder goes down to 0, which is not enough to compensate the material loss caused by the horizontal flow due to tool rotation and the downward flow caused by the thread. Therefore, void appears at the AS of the joint, as shown in Figure 5b. The result in Figure 5 shows that sufficient tool plunge depth should be guaranteed to avoid the void.

Figure 5

Schematic of the material flow using 0.1 mm (a) and 0 mm (b) shoulder plunge depths.

Figure 6 shows the microstructures at different regions of the joint. Figure 6a shows the microstructure of the BM. Some big grains with irregular sizes were observed. Inside the grains and at the grain boundaries, some secondary phases were observed. The secondary phases appeared black under OM. Under SEM, the secondary phases appeared white. The microstructure of HAZ is shown in Figure 6b. The grains were a little larger than those of the BM. Quantity of the secondary phase was smaller than that of the BM, as shown in the OM and SEM images. This was because the HAZ only underwent heat input but not mechanical stirring during welding. Grains with irregular sizes were observed at TMAZ (Figure 6c). The highly deformed grains in TMAZ were because of both the stirring and the heat input during welding. Due to complete dynamic recrystallization, fine and small grains were observed in SZ. Similarly, the size of the secondary phase particles was small (Figure 6d).

Figure 6

Microstructures at (a) BM, (b) HAZ, (c) TMAZ and (d) SZ.

Figure 7 shows the element distributions at the SZ. Some secondary phases with different sizes were observed in the SZ (Figure 7a). Al 2024 alloy possesses a precipitation strengthening property. The main secondary phase is Al₂Cu. The element distribution is shown in Figure 7b. Rather high Cu content was observed at the phases, which corresponded well with the distribution of Al and Cu elements in Figure 7c and b.

Figure 7

Secondary phases: (a) SEM image, (b) element distribution, (c) Al and (d) Cu.

The microstructure of the bonding ligament is shown in Figure 8. No crack was observed inside the bonding ligament (Figure 8a). As introduced above, the pin used in this work had a length of 0.7 mm and the shoulder plunge depth was 0.1 mm. Therefore, the pin tip touched the lap interface during welding. The alclad layers at both the upper and lower sheets were stirred. Figure 8b and c show the element distribution of the bonding ligament. Much high Al alloy content was observed because the bonding ligament is composed of pure Al. Little Cu element was observed at the bonding ligament (Figure 8c).

Figure 8

Morphology of bonding ligament (a), element distribution of Al (b) and Cu (c).

Figure 9 shows the microstructure of the SZs welded at different welding speeds. As shown in Figure 6, the SZ underwent complete dynamic recrystallization, so fine and exquisite grains were observed. The size of the grains was sensitive to heat input. Higher heat input was produced at 50 mm/min and, therefore, the grains were large (Figure 9a). With increase in welding speed, the heat input decreased, i.e., the grains had short time to grow. Therefore, the size of the grains became smaller (Figure 9b and c). Inside the SZ, some black lines were observed (Figure 9a and b). Figure 9d shows its SEM image. The line scan result in Figure 9d shows that the black lines were rich in Cu. Figure 9e and f show the element distributions of this area. The results show that the white spots were the secondary phases that precipitated inside the SZ.

Figure 9

Microstructure of the SZ using (a) 50 mm/min, (b) 100 mm/min and (c) 150 mm/min, (d) SEM image, (e) element distribution and (f) Cu.

Figure 10 shows the hardness curves of the joints at different rotating speeds. The hardness curve presented a typical “W” morphology at the upper SZ. The BM of 2024-T4 Al alloy had a hardness of approximately 140.3 HV. The hardness showed an obvious decrease at HAZ and TMAZ. The minimum hardness of approximately 117.6 HV was obtained at HAZ using 800 rpm, which was attributed to large grain size and less amount of secondary phases. The hardness showed an increase (133.1–135.6 HV) at SZ because of the small grains and the precipitated secondary phases. Al 2024-T4 alloy was one precipitation-strengthened Al alloy. The hardness of SZ was affected by both grain sizes and secondary phases. Generally speaking, more second phases with even distribution result in high strength, such as the BM. On the contrary, less second phases with uneven distribution result in low strength, such as the HAZ and TMAZ. The heat input at 800 rpm was higher than that at 400 rpm. Thus, the grains were larger, and more secondary phases were dissolved at 800 rpm, resulting in lower hardness. The hardness of the lower SZ is shown in Figure 10b. The lower SZs had obvious less widths. It was seen that the hardness was similar at the upper and lower SZs and higher rotating speed resulted in lower hardness.

Figure 10

Hardness of the joints using different rotating speeds: (a) the upper SZ and (b) the lower SZ.

Figure 11 shows the lap shear failure loads of the joints at different rotating speeds. Figure 11a shows the failure loads of the joints when the AS of the joint bore the main load. The failure load first increased at 600 rpm and then decreased at 800 rpm. The maximum failure load of 2,330 N was obtained at the rotating speed of 600 rpm. The minimum failure load of 1,855 N was obtained at 400 rpm. The load-displacement curve showed that the joint made at 600 rpm possessed bigger displacement. Figure 11b shows the failure loads of the joints when the retreating side (RS) of the joint bore the main load. The failure loads were much higher than those when the AS of the joints bore the load. This was attributed to different fracture loads, which was discussed in the following section. The maximum failure load of 6,419 N was obtained at 600 rpm. Similarly, the joint using 600 rpm had bigger displacement, as shown in Figure 11b.

Figure 11

Lap shear properties: (a) joint AS bears the forces and (b) joint RS bears the forces.

Figure 12 shows joint fracture positions and morphologies. Two fracture modes were obtained. Tensile fracture mode was obtained when joint AS bore the lap shear force. Crack first initiated at the hook at AS and then propagated through the SZ, finally reaching the upper surface of the joint (Figure 12a). Shear fracture mode was obtained when the joint RS bore the lap shear force. Crack initiated at the cold lap and then propagated along the bonding ligament, finally reaching the hook at the AS. The upper and lower sheets were separated from each other (Figure 12b). Figure 11 shows that the joints have higher failure loads when the joint RS bore the lap shear force. This attributed to larger bonded width of the bonding ligament. For the tensile fracture mode, the load-bearing distance was the thickness of the upper SZ, which has a value smaller than 0.8 mm because the shoulder had a plunge depth of 0.1 mm. For shear fracture mode, the bonded width of the bonding ligament was larger than the diameter of the pin tip (5 mm). Figure 12c shows the fracture surface of the tensile fracture. Some ridges were observed, whose magnified views are shown in Figure 12d and e. Numerous dimples with different sizes were observed, indicating ductile fracture. Figure 12f and g shows the fracture morphology of the shear fracture with a plenty of dimples.

Figure 12

Fracture positions and morphologies: (a) tensile fracture using 400 rpm, (b) shear fracture using 100 mm/min, (c) fracture surface, (d) and (e) magnified view of the fracture surface in (a) and (f) and (g) magnified view of the fracture surface in (b).

4 Conclusion

1. Voids appear inside the SZ at the small plunge depth of 0 mm because the tool shoulder cannot exert a good material-collecting effect. A 0.1 mm shoulder plunge depth was beneficial to eliminate the voids because this causes tight contact between the shoulder and the material and thus results in good material-collecting effect of the shoulder.

2. The failure loads are higher when the RS of joint bares the main load. The shear failure load first increases and then decreases with increase in the rotating and welding speeds, and the maximum failure load of 6,419 N is obtained at 600 rpm and 150 mm/min.