In Situ Observations of the Strain Competition Phenomenon in Aluminum Alloy Resistance Spot Welding Joints during Lap Shear Testing

Abstract

The real-time evolution of the deformation and strain field of non-heat-treatable aluminum alloy 5754 and heat-treatable aluminum alloy 6061 resistance spot welding joints during the lap shear test was extracted using the digital image correlation (DIC) technique. The strain competition phenomenon between the nugget and its peripheral metal was quantitatively analyzed by applying 2D and 3D DIC analyses. The quantitative data show the tensile strain concentrated in the peripheral metal of the AA5754-O joint, which fractured in the pull-out mode. In comparison, a significant shear strain appears in the nugget of the AA6061-T6 joint, leading to its fracture in the interfacial failure mode during the lap shear test. The phase evolution of the nugget was analyzed using the thermodynamics database JMatPro, which was further used to calculate the local strength of the joints. The results indicate that the nugget strength of AA5754 is 223 MPa, the nugget strength of AA6061 is 178 MPa, and the heat-affected zone (HAZ) strength of AA6061 is 263 MPa. By inputting the local strength data, the calculated result of the analytical load-bearing competition model is in accordance with the experimental data of the lap shear test.

1. Introduction

The mechanics of large-scale heterogeneous materials remain a significant challenge in the field of science and engineering materials [1,2,3]. Particularly, for a fusion-welded joint, the microstructure distribution is uneven, resulting in its heterogeneous mechanical properties. Under various welding parameters, the microstructure and distribution of local mechanical properties in the joint change dramatically [4,5,6]. The overall mechanical property of a fusion-welded joint is influenced by each sub-divided zone, each possessing a different local strength. It is difficult to analyze the mechanical behavior of a heterogeneous welding joint using typical analytical methods. The digital image correlation (DIC) technique is an in situ observation method that provides 2D or 3D displacement fields of solid surfaces [7,8,9], thereby aiding in understanding the mechanical behavior of large-scale heterogeneous materials. Bardel et al. investigated the in-plane strain distribution of an aluminum alloy 6061-T6 butt-weld joint during a DIC-aided tensile test [10]. The result shows that shear strain emerges under uniaxial tension due to the strength loss at the fusion zone and the heat-affected zone (HAZ). Wang revealed local strain concentration in a Cr-Ni-Mo-V steel submerged arc-welded joint during a DIC-aided tensile test [11]. Rahmatabadi et al. employed the DIC technique to measure the variation in the growth of the crack during the R-curve test of an Mg-based composite, providing accurate crack-opening data [12]. However, no studies have focused on characterizing the mechanical behavior of a complex heterogeneous structure. Spot-welded joints exhibit different failure modes (typically interfacial failure and pull-out failure) based on the nugget size, local strength distribution, and load state. Therefore, investigating the mechanical behavior of this structure using an in situ DIC observation method is imperative.

The failure mode of a resistance spot welding (RSW) joint is a result of the competition between the nugget and its periphery sheet metal during the suffering of external loads. The load-bearing ability of the resistance spot welding joints involves both metallurgical factors (i.e., strength distribution) and structural factors (i.e., the size of the nugget) [13,14,15,16]. Kang et al. designed a shear test that coupled with DIC to reveal the constitutive behavior of the nugget and the HAZ of the aluminum alloy RSW joints [17], which is a tremendous advance in characterizing the mechanical behavior of a kind of large-scale heterogeneous materials. Using this method, the structural factors are excluded from the test. Consequently, the local mechanical property of a heterogeneous joint could be extracted. Park et al. exert micro-material testing machines (equipped with a DIC system to measure strain) to test the tensile properties of the DP980 steel RSW nugget zone under different strain rates [18]. They pointed out that the yield stress shows a dramatic rise under the high strain rate condition. However, there is still a pressing need to develop a method that unveils the on-service property of the RSW joints (both two factors are involved), which is critical in evaluating the safety of the spot-jointed structure (a common structure in automotives).

Aluminum alloy was characterized by its low density, corrosion proof, and high impact energy absorptivity, which give rise to its extensive application in the automotive and aircraft industries. However, due to its high heat and electrical conductivity, the resistance spot weldability of aluminum alloy is poor. To produce qualified nugget sizes to guarantee the safety of the aluminum alloy spot-welded structure, a high-welding-current parameter is required, which involves energy consumption. Additionally, due to the high cooling rate, the non-equilibrium solidification process within the nugget raises the challenge of analyzing the phase evolution process in resistance spot welding. It is necessary to achieve a deep understanding of the load-bearing ability of aluminum alloy RSW joints, to help select the optimal welding parameters.

In this study, RSW joints of non-heat-treatable aluminum alloy 5754 and heat-treatable aluminum alloy 6061 are observed by DIC during lap shear tests. A novel lap shear test sample that could reveal the strain competition phenomenon at the cross-section of the welds is designed. Quantitative deformation and strain field data during the lap shear test of the resistance spot welding joints were extracted. Moreover, the local strength of the aluminum alloy RSW joints was calculated using the thermodynamics database JMatPro to analyze the failure mode transition of the joints. The main finding of this study yields a comprehensive understanding of the mechanical behaviors of the aluminum alloy resistance spot welded joints.

2. Materials and Methods

2.1. Deformation and Strain Distribution Test

Here, 1 mm thick 5754 aluminum alloy sheets, in an annealed state (AA5754-O), and 1 mm thick 6061 aluminum alloy sheets, in a fully artificially aged state (AA6061-T6), were used. The mechanical properties of AA5754-O and AA 6061-T6 are shown in the Table S1. A specially designed half-cut sample was observed using the 2D DIC technique during lap shear testing to reveal the in-plane strain distribution at its cross-section. The dimensions of the samples in the lap shear test are shown in Figure 1a. A 3D DIC test was exerted on typical RSW lap shear samples during testing (Figure 1b), which proved the feasibility of the 2D DIC method to show the representative strain competition of the RSW joints.

The setup of the 2D and 3D DIC tests is shown in Figure 2. The macroscopic speckle was painted on the surface of the joints in the 3D test. The samples were prepared using the electrical discharge machining technique to expose the nugget at the cross-section, which in the 2D test was where the finer speckle was painted. The lap shear tests were carried out using a CSS-44100 material test system under a tensile speed of 0.5 mm·min⁻¹. The peak load and energy absorption (the area under the load-displacement curve, up until the joint reaches its peak load) is recorded. A VIC-3D digital image correlation system (Correlated Solutions, Inc., Irmo, SC, USA) was used to record the surface distortion of the typical RSW joints during lap shear testing. Two long-lens CCD cameras (500 mm) were fixed on a rigid support, which was positioned symmetrically around the specimen. For the 2D test (Figure 2a), one short lens CCD camera (70 mm) was put perpendicular to the plane of the cross-section. A VIC-2D digital image correlation system was used to record the in-plane displacement of the RSW joints’ cross-section during testing. Then, the in-plane strain was calculated. The collection frequency of both the 2D and 3D tests was fixed at 5 frames per second.

In this work, a 1 mm thick base metal was used, which was welded using a 220 kW medium-frequency direct-current resistance spot welding machine(Company, City, State, Country). Firstly, the mechanical properties of the AA5754-O and AA6061-T6 RSW joints were compared under different processing parameters. The welding parameters were designed according to a mixed level of orthogonal trials (L₁₈(6 × 3⁶)), which involve the welding current (I), the welding time (t), and the electrodes’ pressing force (F) (Table S2). The peak load and energy absorption (integral area of the load-displacement curve from the initial point to the peak load point) of each sample were recorded. The AA5754-O and AA6061-T6 joints possess comparable nugget sizes but are fractured in different modes, meaning different modes were selected to perform the DIC observations.

2.2. Microstructure Characterization

The Vickers hardness of the AA5754-O and AA6061-T6 RSW joints at their cross-sections were tested using an HV-1000A micro-hardness hard-meter with a 200 g load for the 15 s holding time. The profile of each nugget was observed by an Olympus SZX12 stereomicroscope and an Olympus GX51 metallographic. The samples were etched with Keller’s reagent.

2.3. Local Strength Calculation

In order to analyze the failure mode transition of the two kinds of joints, the local strength of the nugget was calculated using the thermodynamics JmatPro database [19,20,21]. The phase evolution within each nugget was calculated by the Scheil–Gulliver model [19,20]. During the solidification process, the subject can be divided into a solid portion and a liquid portion. The Scheil–Gulliver model assumes the cooling rate of the melt is infinitely high. Thus, the convection within the liquid portion is complete, while there is no diffusion within the solid portion. This model is proven to be suitable for depicting the solidification process in RSW [22].

After RSW, the microstructure of the nugget exhibits a dendritic feature. The solute within the nugget is segregated at the inter-dendritic zone. During the non-equilibrium solidification process, the solute segregation causes the formation of additional intermetallic compounds (IMCs) and eutectic compared to the equilibrium state alloy. Each portion of the α-Al solid solution, IMCs, and eutectic can be calculated. Then, the local strength of the nugget can be gained by Equation (1).

$${\sigma }_S=f_{Al}{\sigma }_{Al}+\sum f_i{\sigma }_i+{\sigma }_{ppt}$$

(1)

where σ_S is the yield strength and f_Al and σ_Al are the fraction and strength of α-Al, respectively. f_i and σ_i are the fraction and strength of each type of eutectics, respectively. σ_ppt is the strength component caused by the precipitation strengthening effect. In the JMatPro aluminum alloy database [21], σ_S is calculated by the composition input and cooling condition. In this study, the composition of AA5754-O is 96.3 Al, 3.2 Mg, and 0.5 Mn, while the composition of AA6061-T6 is 97.55 Al, 1.2 Mg, 0.8 Si, 0.15 Mn, and 0.3 Cu (wt. %). Furthermore, the local strength of the heat-affected zone can be gained by calculating the over-aged state strength of AA6061. Impurity elements, such as Fe and Cr, are neglected in this study.

3. Results and Discussion

3.1. Microstructure of the Welds

Figure 3 shows the microstructure of the AA5754-O and AA6061-T6 RSW joints at their cross-sections. It could be observed that the nugget consists of α-Al and inter-dendritic IMCs and eutectic. The nugget can be further divided into a peripheral columnar dendritic zone and a centric equiaxed dendritic zone. Columnar to equiaxed dendritic transition is a typical feature of aluminum alloy RSW nuggets.

Figure 4 shows the temperature vs. phase fraction curves for the AA5754-O and AA6061-T6 nuggets, which are calculated by JMatPro using the Scheil–Gulliver model. This data shows that during the solidification process, the α-Al solid solution is the primary crystalline phase, which forms the stem of the dendrites. The IMCs and eutectic generated at the ending solidification stage form the inter-dendritic phases.

The microstructure of the nugget and the phase evolution data show that the solute of the aluminum alloy is segregated at the inter-dendritic zone, to form the eutectic and IMCs. This leads to the heterogeneous local mechanical property distribution of the joint. Especially, for the heat-treatable AA6061, the segregated dendritic nugget exhibits lower local strength compared to the base metal (which was approved by the hardness distribution in Section 3.2).

3.2. Lap Shear Testing of the Joints

The output data of the orthogonal trials of AA5754-O and AA6061-T6 RSW joints are listed in Supplementary Materials Tables S3 and S4 respectively, which were rearranged by nugget size. Figure 5 shows the peak load vs. nugget size and energy absorption vs. nugget size data for the AA5754-O and AA6061-T6 RSW joints, respectively, after lap shear testing. Under similar nugget sizes, the peak loads of the AA5754-O and AA6061-T6 RSW joints are comparable to each other. However, the energy absorption of the AA5754-O RSW joints is much higher than the AA6061-T6 joints. That is because most of the AA5754-O joints are fractured in the pull-out failure mode (PO), which involves a dramatic plastic deformation of the material adjacent to the nugget.

The micro-hardness distributions of the two kinds of joints are shown in Figure 6. The nugget in the AA6061-T6 RSW joint has a significantly lower hardness than the base metal. The Vickers hardness of the AA6061-T6 nugget is 59.3 ± 3.5 HV (hardness of the AA6061-T6 base metal is 103.4 ± 1.4 HV), while for the AA5754-O nugget it is 57.1 ± 2.8 HV (hardness of the AA5754-O base metal is 59.2 ± 1.6 HV). According to the typical theory on the competition of the two failure modes [22,23], the critical diameter for the transition from interfacial failure (IF) to pull-out failure (PO), d_c, is proportional to the ratio between the pull-out failure location’s strength and that of the nugget (under lap shear test). In this study, both the AA5754-O and AA6061-T6 RSW joints have similar nugget diameters under similar heat inputs. Due to the differences in the hardness distributions, the AA6061-T6 RSW joint fractured in the IF mode, whereas the AA5754 RSW joint fractured in the PO mode. The critical failure mode transition nugget diameter d_c for the AA5754-O RSW joints is 5.25 mm, while its counterpart for the AA6061-T6 RSW joints is 6.75 mm. The heat-affected zone (HAZ) region is enlarged when the heat input is increased. That triggers the crack to propagate in the weak region, which results in the PO failure mode of the AA6061-T6 joint under a high heat input. In fact, the long welding time promotes the PO failure mode by weakening the AA6061-T6 joint through over-aging [24,25]. Therefore, it is not recommended to make an AA6061-T6 RSW joint since it fractured in the PO mode using this method.

Figure 7 shows the typical load-displacement curve for the AA5754-O and AA6061-T6 joints in the lap shear test. Note that since the strength ratio between the nugget and base metal is higher in the AA5754-O joint than the AA6061-T6 joint (inferred by the hardness distribution), the AA5754-O joint fractured in the PO mode, while the AA6061-T6 joint fractured in the IF mode. Although the peak load of the two joints is comparable to each other, the energy absorption of the AA5754-O joint is much higher than that of the AA6061-T6 joint.

3.3. D Surface Distortion Test

Figure 8 shows the quantitative displacement data extracted from the nugget center of the AA5754-O and AA6061-T6 RSW joints. The AA5754-O RSW joint fractured in the PO mode. It could be observed that the nugget distorted dramatically in the y and z directions, while that in the x direction nearly remained constant. The AA6061-T6 RSW joint fractured in the IF mode. Compared to the AA5754-O RSW joint, the total z direction displacement decreased owing to the crack propagating in the nugget (IF mode), which caused less nugget rotation compared to its counterpart in the PO mode. Each of the two samples barely possesses any x direction displacement (there is no warping in y–z plane), meaning the half-cut sample in this study is of practical use and can reveal the in-plane strain on the RSW joints during lap shear testing.

Figure 9 shows the comparison of the surface distortion of the two kinds of joints when necking of the initial fracturing site occurs (the load-displacement curve nearly reaches its peak value). It could be observed that a massive area in the AA5754-O RSW joint distorts when the joint reaches its peak load during the test. The void in the contour in Figure 9a,c,e refers to the fracturing of the workpiece. As a comparison, the total deformation is limited for the AA6061-T6 RSW joint.

3.4. D in-Plane Strain Competition Test

Figure 10 shows the quantitative strain data of the nugget and its peripheral zone during 2D DIC testing. During the lap-shear test, due to the rotation of the nugget, mixed strain (which infers a mixed stress state) shows up at the final failure location, where tensile strain is predominant. That result is in accordance with the previous analytical and numerical simulation studies [26,27,28]. The AA5754-O RSW joint fractured in the PO mode, while the AA6061-T6 RSW joint fractured in the IF mode. The quantitative data for the necking zone in the AA5754-O RSW joint (peripheral metal) shows that when it reaches its peak load during the lap shear test, the tensile strain is 0.4314, while the shear strain is 0.0673. At this moment, the tensile strain is 0.0147 in the AA5754-O nugget. The total degree of the strain is less in the AA6061-T6 joint compared to the AA5754-O joint. Only a tensile strain of 0.2442 and a shear strain of 0.044 were observed in the AA6061-T6 nugget, while a tensile strain of 0.0017 and a shear strain of 0.0453 were observed in the peripheral metal.

Figure 11 shows the strain distribution at the cross-section of the AA5754-O RSW joint during the lap shear test. It could be observed that in the AA5754 RSW joint, which fractured in the PO failure mode, the strain is concentrated at the region near the nugget, at the drawn and fixed leg of the joint. By contrast, the strain in the nugget is fairly minimal.

Figure 12 shows the strain distribution at the cross-section of the AA6061-T6 RSW joint during the lap shear test. Since this joint fractured in the IF mode, a main shear strain zone, which runs through the nugget, could be observed. This is different from the nugget strain feature in the AA5754-O joint.

Note that the size and hardness of the nuggets are similar to each other for the two kinds of joints. It is the high strength of the periphery materials in the AA6061-T6 joint that causes more strain to occur at the nugget. Instead of the failure occurring at the necking of the sheet metal, adjacent to the nugget, in the AA5754-O joint, the IF failure of the AA6061-T6 joint was caused by plastic deformation in the nugget, where the shear strain is predominant.

3.5. Failure Mode Transition Analysis

The failure of a resistance spot welded joint is described as a competition between the necking in the sheet metal near the nugget and the shear deformation of the nugget [29,30,31,32,33]. This phenomenon has been observed using the DIC technique, which is shown in Figure 11 and Figure 12. For aluminum alloys, which possess good ductility, the peak load of a lap shear test sample, which is fractured in the IF mode, can be estimated by Equation (2).

$$P_{IF}=\frac{\pi }{4}\cdot d^2\cdot {\tau }_{nugget}$$

(2)

where d is the nugget diameter and τ_nugget is the shear strength of the nugget. According to the Tresca failure criterion, τ_nugget = 0.5σ_nugget.

The PO failure process involves massive deformation of the RSW joints, which yields the complexity of raising the analytical model. M. Pouranvari et al. [29,30] proposed an evenly distributed stress model to calculate the peak load of a PO lap shear sample, which has:

$$P_{PO}=\pi \cdot d\cdot t\cdot {\sigma }_{FL}$$

(3)

where t is the thickness of the workpiece and σ_FL is the tensile strength of the failure location. Obviously, the strain and the stress field are not evenly distributed during the lap shear test, which has been proven by the data from Figure 8 to Figure 12. Thus, the peak load calculated by this model is higher than the experimental data.

Chao [34] proposed a harmonically distributed stress model to calculate the peak load of a PO lap shear sample, which has:

$$P_{PO}={\displaystyle {\int }_{-\frac{\pi }{2}}^{\frac{\pi }{2}}{\sigma }_{FL}}\cdot \frac{d}{2}\cdot t\cdot \cos \theta d\theta =\frac{\pi }{4}\cdot d\cdot t\cdot {\sigma }_{FL}$$

(4)

however, the effect of the massive yield zone in RSW is neglected in Chao’s model. The stress is excessively concentrated in this model. Thus, the peak load calculated by this model is lower than the experimental data.

In this study, the P_PO is calculated using the average value from M. Pouranvari’s and Chao’s models [29,30,34]. When P_IF ≥ P_PO, by increasing the nugget diameter, the failure mode transition occurs. The local strength of the nugget and HAZ is calculated by JMatPro, which shows the nugget strength in AA5754 is 223 MPa, the nugget strength in AA6061 is 178 MPa, and the HAZ strength in AA6061 is 263 MPa. The calculated P_IF and P_PO curves are shown in Figure 13, which is in accordance with the experimental data in the lap shear test.

The 2D DIC data (Figure 10 to Figure 12) show that the degree of shear strain is higher in the nugget of AA6061-T6 compared to that in AA5754-O. The sign of IF fracturing is two shear strain zones at the notch merging into one. Although the nugget size of the two joints is at the same level, the high strength of the heat-treatable aluminum alloy base metal causes it to be more prone to shear deformation in the nugget, which results in the IF mode of the AA6061-T6 joint during testing (which is regarded as an unsafe weld in the automotive industry).

4. Conclusions

In this study, a novel lap shear sample that could reveal the strain competition is designed. The real-time evolution of the deformation and strain field of the resistance spot welding joints during the lap shear test were extracted. Quantitative data of the in situ strain distribution of two types of aluminum alloy RSW joints were gained using a digital image correlation technique. The following conclusions can be drawn:

(1)

Due to the strength decrease in the nugget during the resistance spot welding process, the AA6061-T6 joint is prone to fracturing in the interfacial failure (IF) mode. Although the base metal yield strength is higher than that of AA5754-O, the mechanical property of the AA6061-T6 resistance spot welding joint does not possess superiority compared to the AA5754-O joint.

(2)

The 3D surface distortion of the two kinds of aluminum alloy resistance spot welding joint during the lap shear test shows the main deformation of the joints occurs along the y axis (tensile direction) and along the z axis (caused by nugget rotation). Since almost no deformation occurred in the x direction, it makes it practical to use the half-cut sample to reveal the in-plane strain map for the resistance spot welding joint during the tensile shear test.

(3)

The quantitative data on the in-plane strain shows that there is fairly little strain at the nugget for the pull-out failure mode resistance spot welding joint. However, a mixed tensile/shear strain occurs at the periphery of the nugget in both the interfacial failure and pull-out failure mode joints. If a joint is fractured in the interfacial failure mode, the degree of this mixed strain is less than that for the pull-out failure mode joint.