Local strain measurement in tensile test of Al/steel dissimilar friction stir weld using digital image correlation

A6061-T6 aluminum alloy plate was butt welded to type 304 stainless steel plate by a friction stir welding (FSW) technique to fabricate a dissimilar weld. Subsequently, digital image correlation (DIC) analysis was performed on the dissimilar weld to investigate the local deformation behavior near the joint interface under a tensile loading condition. The normal strain in the loading direction concentrated near the interface on both top and bottom surfaces. The positions of the strain-concentrated areas were different on both sides; around 9 mm away from the interface on the top surface and 6 mm away on the bottom surface, which could be attributed to the local softening developed during the FSW process.


Introduction
Multi-material design concept is becoming common especially in the automotive industries, whilst joining of dissimilar materials is essential to achieve that design concept. Consequently, joining of dissimilar alloys have been widely tried. The solid state joining process, friction stir welding (FSW), is expected to be a joining method between aluminium (Al) alloy and steel plates, which are two major materials consisting of car body components. Therefore, there have been many trials for joining Al alloys and steels by FSW [1][2][3][4][5][6][7] and friction stir spot welding (FSSW) [8][9][10] techniques. The optimization of welding parameters and the microstructure analysis at the interface between Al and steel have been widely investigate, focusing on the development of intermetallic compounds (IMC) along the interface [11][12][13].
That is because IMC layer plays an important role for achieving good joint strengths. Recently, microsized specimens were sampled near the interface, and the interface joint strengths were directly measured [14][15][16], which revealed that the high performance of interface strength between A6061-T6 and type 304 stainless steel was achieved when the thickness of IMC layer was less than 0.4 μm [14].
In our previous study [17], residual stresses in Al/steel dissimilar friction stir weld were measured and the effect of post heat treatment was discussed. In Al/steel dissimilar welds, local deformation behaviour near the interface might be complex because elastic moduli are largely different between Al and steel. Therefore, in this study, we focused on a digital image correlation (DIC) method, which can measure the displacement using two digital images before and after loading. DIC has been used for the local displacement measurement of weldments [18], friction stir welds [19][20][21] or spot welds [22] and friction stir dissimilar welds of different Al alloys [23]. For example, the hardness distribution in the weld zone is asymmetric between top and bottom sides, and retreating and advancing sides in FSW joints. DIC measurement during tensile test revealed that the local strain concentration at the locally-softened area resulted in the nal tensile failure [21]. Although Al/steel dissimilar welds have more signi cant discontinuity at the interface compared with Al/Al similar or Al/Al dissimilar welds, the DIC analysis of Al/steel dissimilar weld has not been caried out. Consequently, Al alloy and steel plates were joined by an FSW technique, and the local strain near the interface was measured by a DIC method on both top and bottom surfaces to investigate the effect of discontinuity of microstructures and elastic moduli on the local deformation behaviour.

Materials and welding conditions
The materials used are Al alloy, A6061-T6, and stainless steel, type 304, plates with the width of 75 mm and length of 150 mm. The thicknesses of A6061 and type 304 plates are 6 and 5 mm, respectively. The tool with concave shoulder, with a diameter of 14 mm, was used for the joining. The tool has a threaded probe with the diameter of 6 mm and length of 4.7 mm. The tool rotational and traveling speeds were xed at 700 rpm and 100 mm/min, respectively. The tool plunge depth and tool-to-workpiece angle were 0.8 mm and 3° from the vertical axis, respectively. The detailed joining procedures and microstructures in the stir zone (SZ) are described in our previous studies [6,17].

Experimental Procedures
The depth of 0.1 mm was removed from the top and bottom surfaces of the as-welded plate by milling as schematically shown in Fig. 1. Subsequently, the tensile specimens were sampled from the welded plates by electrical discharge machining. Tensile tests were conducted using the specimens in accordance with JIS (Japanese Industrial Standard) Z 2241 13B, where the test piece has the gauge length of 50 mm and the width of 12.5 mm as shown in Fig. 2. The tool centre corresponds to the centre of the gauge length of the tensile specimen. Tensile tests were conducted using an electro-hydraulic testing machine (SHIMADZU: EHF-LV-020k1-020).
A digital camera, Nikon D7500, with the CMOS sensor with the size of 23.5 × 15.7 mm was used for the DIC procedures, in which images of specimen surfaces were sampled. The lens of the digital camera was Nikon AF-S VR Micro-Nikkor 105 mm f/2.8G IF-FD. Random dotted patterns on the specimen surfaces were formed by white and black colour sprays. In the DIC procedure, deformation is measured by the matching of luminance distribution in the subset areas before and after loading. The example of random dotted pattern and the subset area are shown in Fig. 3.

Hardness distribution
The detailed microstructures and mechanical properties are described elsewhere [6,17], thus the hardness distribution is brie y shown in this section. Figure 4 shows the hardness pro les measured on the top and bottom surfaces of the tensile specimen along the dotted line in Fig. 2. Softening occurred in the SZ due to the heat input during FSW, and the minimum hardness appeared at the thermo-mechanically affected zone (TMAZ). It should be noted that the minimum hardness appears at about 9 mm away from the interface on the top surface and about 6.5 mm on the bottom surface. In addition, the hardness near the interface (0~1 mm) on the bottom surface was nearly comparable to the minimum hardness at 6.5 mm away from the interface. The hardness mapping measured on the side surface of Al side is shown in Fig.  5 [6]. It should be noted that the area of lower hardness (blue area) in the TMAZ has vertical asymmetricity, where lower hardness appears further from the interface on the top side. Furthermore, minimum hardness appears near the interface (0~3 mm) only on the bottom side.

DIC analysis: Top and bottom surfaces
The ultimate tensile failure load of the joint was 201 MPa, where tensile failure occurred in Al side along TMAZ as shown in Fig. 6, indicating that the interface strength was higher than the TMAZ with lower hardness [6]. Firstly, DIC analysis was performed on the top surface of the dissimilar weld of the tensile specimen. Figure 7 shows the analytical area on the top surface including the interface between Al and steel, where y axis corresponds to the loading direction. The analytical conditions are summarized in Table 1. The image before loading was used as a reference image, and the images at 4.0 kN (86 MPa), 5.0 kN (107 MPa) and 6.0 kN (128 MPa) are used for the DIC analysis. The DIC analytical results showing y-axis displacement and strain are revealed in Fig. 8. Figure 9 shows the y-axis strain along the dotted line in Fig. 8(b) at 6.0 kN. It should be noted that the maximum local strain concentration occurred at about 9 mm away from the interface, which corresponds to the locally-softened area of TMAZ as shown in Fig. 4. The local strain distributions at 0, 1, 9, 12 and 14 mm away from the interface are shown in Fig. 10. The areas in the gure were squares with the size of 1.85 mm × 1.85 mm. At the beginning of the loading (4.0 kN and 5.0 kN), strain concentration occurred at about 1 mm away from the interface in very narrow area.
However, the wide strain concentration occurred at 9 mm from the interface at 6.0 kN, which corresponds to the softened TMAZ (Fig. 4) and the actual fracture location (Fig. 6).
Subsequently, DIC analysis was performed on the bottom surface, where the analytical conditions are summarized in Table 1. y-axis displacement and strain are shown in Fig 11. Furthermore, the local strain distributions at 0, 3, 5.6, 11.5 and 19 mm away from the interface are shown in Fig. 12. Local strain concentration occurred just beside the interface from 4.0 to 6.0 kN, while large strain occurred at about 5.6 mm away from the interface at 6.0 kN. That location corresponds to the locally-softened area (Fig. 4) and the actual fracture location (Fig. 6). Consequently, the local strain concentration occurred at the locally-softened area in TMAZ, and just beside the interface on the bottom surface.
The DIC analytical result on the side surface is shown in Fig. 13 as the strain distribution at 6.0 kN. It should be noted that the strain asymmetricity between top and bottom sides is observed on the side surface along TMAZ. In addition, local strain concentration near the interface occurred only on the bottom side as shown by the arrow in the gure, and it is consistent with Figs.10 and 12. In our previous study, the strain distribution was measured by DIC using friction stir welded Al-Mg-Sc plates. In that case, local strain concentration occurred only at TMAZ where minimum hardness appeared, and static failure occurred at the maximum straining area. In the present case, local straining occurred at TMAZ and just beside the interface on the bottom side. However, local straining area was much narrower near the interface due to the constraint by steel with higher elastic modulus. Therefore, it is assumed that the nal failure predominantly occurs at wider local-straining area in TMAZ.

Local stress-strain curves
The local straining gures shown in Figs. 10 and 12 were obtained at every 0.5 kN on the top surface, and every 1 kN on the bottom surface to draw local stress-strain curves. The local strain was defended as the average value of each square shown in Figs. 10 and 12. Figures 14a and 14b are the local stress-strain curves at SZ, TMAZ, HAZ and base metal (BM) on the top and bottom surfaces, respectively. The black solid line is a stress-strain curve of the bulk sample of the base metal, where elastic modulus of 68.3 GPa was measured by DIC in the present study and 0.2% proof stress of 287 MPa was given by a material supplier. On the top surface (Fig. 14a), the stress-strain relationships in SZ, HAZ and BM are linear, where the local elastic moduli are 65.4, 53.5 and 54.0 GPa, respectively. On the other hand, local yielding occurred only at TMAZ, where the hardness was minimum. The elastic modulus at TMAZ is 60.8 GPa. The local 0.2% proof stress at TMAZ is estimated about 120 MPa, which is 58 % lower than the base metal. The hardness drop at TMAZ was about 48 % as shown in Fig. 4, therefore the drop of local yield stress was larger than the hardness. However, the value of the local yield stress still seems reasonable because the deference in the reduction rates is small. Similar to the top surface, the local yielding is prominent only at TMAZ on the bottom surface (Fig. 14b), where the local proof stress is also about 120 MPa. The strain value at the maximum load of 6.0 kN is also slightly larger than the elastic line at SZ, but it seems a measurement error because the strain values at the maximum load on the bottom side are larger than the top side at all locations including base metal. The local elastic moduli at SZ, TMAZ, HAZ and BM on the bottom side are 63.8, 50.7, 52.9 and 60.6, respectively. It should be noted that the elastic modulus at the SZ was the largest on both top and bottom surfaces. It was con rmed that many steel fragments were dispersed in the SZ by microstructural observations [6,17]. Therefore, higher local elastic modulus at the SZ could be attributed to such steel fragments with higher elastic modulus than Al.

Conclusions
Aluminium alloy, A6061-T6, plate was welded to type 304 stainless steel plate by a friction stir welding technique. The tensile tests were performed and the local straining behaviour near the interface between Al and steel was analysed by a digital image correlation (DIC) method. The main results obtained are as follows.
1. The hardness measurement revealed that local softening occurred in the SZ, TMAZ and HAZ. The asymmetricity of hardness distribution was recognized in the thickness direction. The minimum hardness appeared at TMAZ.
2. The maximum local strain concentration occurred at 9 mm and 6 mm away from the interface on top and bottom surfaces, respectively. The locations of this local strain concentration corresponded to those where minimum hardness appeared, namely TMAZ.
3. The maximum local strain concentration also appeared just beside the interface on the bottom surface, whilst the width of local strain concentration was narrower compared with that at TMAZ.
Consequently, nal tensile failure occurred along TMAZ, where local strain concentration area was much wider.   Tensile specimen con guration.      Strain distribution along the dotted line in Fig. 8(b) at 6.0 kN.