Effect of Plate Thickness on Tensile Property of Ti–6Al–4V Alloy Joint Friction Stir Welded Below β-Transus Temperature

Joint formation

During the welding process, argon-shielding gas is used to prevent the Ti–6Al–4V alloy joints from oxidizing. The temperature in and near WZ is closely related with process parameters, decreasing with increasing the welding speed or reducing rotational speed. Figure 1 shows the surface appearances of 2 mm thick Ti–6Al–4V alloy FSW joints under different parameters. Extruded material and flashes are due to the higher peak temperature resulting from bigger rotational speed (Figure 1(a)).Therefore, severe groove defect appears because too much material is squeezed out of joint. Compared with joint under smaller rotational speed (Figure 1(b)), higher peak temperature makes the protection more difficult, leading to the serious oxidation (Figure 1(a)). In fact, high temperature also results in coarse microstructure inside joint, which is detrimental to mechanical properties [11, 12, 15].

Figure 1:

Surface appearances of 2 mm thick joints: (a) 475 rpm/50 mm/min and (b) 120 rpm/30 mm/min.

Su et al. [11] and Zhang et al. [12] showed that cracks easily generated inside the joints at higher rotational speed, while the defect-free joints were produced at lower rotational speed. The similar experimental results were also attained in this study, as shown in Figures 2 and 3. It is interesting that the defect morphology in Figure 2(a) is different from void defect in aluminum alloy FSW joints [17, 18]. In Figure 2(a), the defect is prolate, discontinuous and throughout the whole WZ. It is also parallel to welding surface. The defect is called tearing defect in this study. Welding tensile stress at the cooling stage increases with increasing peak temperature in WZ. For FSW joints of Ti–6Al–4V alloy, bigger rotational speed means higher welding peak temperature, which produces bigger welding tensile stress. In fact, besides peak temperature, temperature gradient along plate thickness has an important effect on welding tensile stress. Generally speaking, the bigger temperature gradient easily produces higher welding tensile stress in weld. Due to low thermal conductivity, Ti–6Al–4V FSW joint exists a large temperature gradient along thickness [1, 4, 8, 10], which is bad for decreasing the welding tensile stress. The tearing defect occurs when welding tensile stress exceeds tensile strength limit of material. As is well known, low peak temperature causes insufficient material flow, leading to the appearance of cavity or tunnel defect in FSW joint [19, 20]. From Figure 2, it is concluded that the tearing defect mainly results from welding tensile stress instead of material flow behavior.

Figure 2:

Cross-section morphologies of 2 mm thick joints: (a) 375 rpm/50 mm/min and (b) 120 rpm/30 mm/min.

Figure 3:

Cross-section morphologies: (a) 1.8 mm/120 rpm/30 mm/min and (b) 2.5 mm/100 rpm/30 mm/min.

In terms of FSW characteristic, the heat mostly generates from friction between shoulder and BM [8, 11]. The friction between tool pin and material brings a small number of heat. In the paper, the WZ consists of shoulder-affected zone (SAZ) and pin-affected zone. Microstructure of SAZ is used to analyze the peak temperature under different welding parameters, as shown in Figure 4. For the two-phase Ti–6Al–4V alloy, when the peak temperature at the heating stage is above transus temperature and heating time is long enough to ensure the thorough transformation of α → β, the β → α transformation usually happens at the cooling stage. Therefore, lamellar microstructures appear in SAZ near top surface of welding joint at room temperature (Figure 4(c)). Moreover, SAZ in Figure 4(b) is composed of small equiaxed microstructure, verifying that the peak temperature is lower than β-transus temperature. No lamellar microstructure is observed in Figure 4(b). It is similar to the experimental results reported by Kitamura et al. [13]. Although the material is not fused during the welding process, the welding temperature exceeds recrystallization temperature of Ti–6Al–4V alloy (700°C). During FSW process, the material experiences large strain and strain rate. Therefore, dynamic recrystallization (DRX) happens in WZ [21, 22]. As a result, grains in SAZ characterized by full equiaxed microstructure (Figure 4(b)) are smaller than BM composed of primary equiaxed α and elongated β (Figure 4(a)).

Figure 4:

Microstructures of joint under different welding parameters: (a) BM, (b) SAZ at 120 rpm/50 mm/min and (c) SAZ at 375 rpm/50 mm/min.

Mechanical property

Figures 5 and 6 present results of tensile strength limit and failure location under different thicknesses. Under reasonable welding parameters, defect-free joints exhibit greater tensile strength limit than BM. With regard to the joints of 1.8 mm, 2 mm and 2.5 mm thicknesses, the average values of tensile strength limit achieve 1,147 MPa, 1,075 MPa and 1,045 MPa, respectively. The values are higher than BM referred to Table 1. From the reported researches, obtaining the elongation higher than 50 % of BM is very difficult for FSW joints of Ti–6Al–4V alloy [23, 24]. In the present study, the similar conclusion of lower elongation is attained for FSW joints of 1.8 mm and 2 mm thicknesses. The elongations of FSW joint reach 3.01 % and 1.25 % for 1.8 mm and 2 mm thick joints, respectively. No constriction is obviously observed in these joints after tensile test (Figure 6(a) and (b)). However, the 2.5-mm thick FSW joints own bigger elongation of 6.52 %, which is 52.16 % of BM. In conclusion, FSW joints with different tensile strength limit present diverse elongation under the same thickness reduction.

Figure 5:

Tensile properties of different thick joints with surface indentation.

Figure 6:

Tensile fracture specimens with surface indentation: (a) 1.8 mm (b) 2 mm and (c) 2.5 mm; without surface indentation: (d) 2 mm.

In fact, elongation of FSW joints is closely related to gauge. Gauge length is always bigger than diameter of tool shoulder. In order to investigate the elongation differences of the welded joints, the theoretical analysis is used from the viewpoints of loading capacity.

where N is the load, σb is the tensile strength limit, l is the specimen length, h is the specimen thickness and S is the area of cross section. Subscripts w, b, t represent welding joint, BM and thickness reduction, respectively.

When BM yields,

When BM does not yield,

In the paper, thickness reduction of all the joints is 0.15 mm or so. For 2 mm thick FSW joints, ht/hb=7.5%>(σbw−σsb)/σbw=6.3%. For FSW Ti–6Al–4V alloy joints with the thickness of 1.8 mm, ht/hb=8.3%<(σbw−σsb)/σbw=8.7%. In terms of FSW joints with the thickness of 2.5 mm, ht/hb=6%<(σbw−σsb)/σbw=7.2%.

According to eq. (1) or eq. (2), BM of tensile specimens with 2 mm thickness doesn’t yield and the elongation is just 1.25 %. With regard to 2.5 mm thick tensile samples, BM yields and elongation is more than 50 % of BM. However, as for 1.8 mm thick joints, 8.3 % and 8.7 % are approximately equal so that BM or heat affected zone (HAZ) of tensile specimens may yield. Thus, the elongation of 1.8 mm thick joints is higher than 2 mm thick joints. In a word, whether elongation is high depends on tensile strength limit of joints and yield strength limit of BM.

The conclusions about elongation of titanium alloy FSW joints with surface indentation are also suitable for aluminum alloys. From the tensile strength limit of aluminum alloy FSW joints under different welding process parameters [25], the change of elongation can be explained by eqs. (1) and (2), which can indirectly verify the validation of analysis in this study.

Therefore, it is necessary for titanium alloy FSW joints to eliminate surface indentation by reasonable process and then evaluate the mechanical properties. This is why some researchers investigated the mechanical properties of Ti–6Al–4V alloy FSW/ friction stir processing (FSP) joints by welding core sampling [11, 12]. Figure 6(d) shows the 2-mm thick FSW specimen without surface indentation after tensile test. It can be found that fracture position locates at BM and obvious plastic deformation happens, which verifies that the tensile strength limit of WZ is higher than that of BM. Figure 7 presents the tensile test results of FSW joints without surface indentation. For the joints without surface indentation, that is to say, ht/hb=0, eq. (1) follows so long as the tensile strength limit is higher than yield strength limit of BM. Therefore, BM yields during the tensile test and the FSW joints without surface indentation easily own higher elongation. For the 2-mm thick FSW joints, the tensile strength limit is nearly equivalent to that of BM and elongation reaches 58.5 % of BM.

Figure 7:

Tensile properties of 2 mm thick joints without surface indentation at 120 rpm/30 mm/min.

Certainly, for titanium alloy plates thinner than 2 mm, smaller shoulder plunge depth or higher joint tensile strength limit of WZ may result in higher elongation, which is not further discussed in the present study.

Fracture position and morphology

In order to further investigate tensile characteristic of specimens, Figures 8 and 9 indicate the fracture position and fracture surface morphology of cross section of welding joints with surface indentation, respectively. During the welding process, because of the plunge of shoulder, a slight surface indentation appears in the WZ, where stress concentration occurs. Moreover, compared with other regions of joint, SAZ owns bigger welding residual tensile stress owing to higher peak temperature during the welding process. Therefore, SAZ is a weak zone of welding joint, where the fracture position locates (Figure 8(a) and (b)).

Figure 8:

Fracture location of different thick joints: (a) 2 mm and (b) 1.8 mm.

Figure 9:

Fracture surface morphology of different thick joints: (a) 2 mm and (b) 1.8 mm.

For the 2.0-mm and 1.8-mm thick FSW joints with surface indentation, there exists the difference of fracture position. In the present study, with the increase of tensile load, tensile stress of WZ in 2 mm thick FSW specimens first reaches the tensile strength limit of materials and then crack appears, while BM undergoes the tensile stress lower than yield limit. With the propagation of crack, the effective thickness of WZ decreases, leading to the increase of tensile stress in the region without crack and then uninterruptedly deteriorating the mechanical property of WZ. Therefore, fracture position locates at the WZ of 2 mm thick FSW joints. Compared with 2 mm thick specimen, the fracture position of 1.8 mm thick specimen locates near the edge rather than middle of SAZ, resulting from the difference of stress concentration. In fact, HAZ is made up of coarser grains and then owns lower tensile strength limit compared with BM. Therefore, if the yield strength limit of BM is replaced by that of HAZ, eq. (1) may come into existence. In other words, when the tensile strength limit of welding joints with the thickness of 1.8 mm reaches 1,147 MPa, yield phenomenon may happen in HAZ, where the effective width decreases, worsening the loading capacity. From the experimental results mentioned above, WZ happens DRX and consists of fine and equiaxed grains, resulting in the better mechanical property compared to BM. Therefore, the fracture position along the thickness direction of 1.8 mm thick welding joint locates at SAZ, HAZ and then BM, which is different from 2 mm thick welding joints. Moreover, the experimental results in Figure 8 indirectly represent that there are no hole, groove or lack of penetration defects in WZ. From Figure 9, it is seen that fracture morphology of cross section of Ti–6Al–4V alloy FSW joints composes of dimples with different size and depth, which presents the typical ductile fracture. Generally speaking, bigger and deeper dimples indicate higher elongation of welding joints. Compared to 2 mm thick specimens (Figure 9(a)), the dimples of fracture morphology of 1.8 mm thick FSW specimens are larger and deeper (Figure 9(b)), which agrees with the experimental results of elongation in the present study (Figure 5).